SlideShare a Scribd company logo
1 of 27
Download to read offline
IMAGE SIGNAL PROCESSING


                               VERHAERTINNOVATIONDAY – OCTOBER 20th, 2006
www.mastersininnovation.com




                                                                                          IMAGE SIGNAL PROCESSING



                                               Frederik Wouters
                                               frederik.wouters@verhaert.com
                                               www.verhaert.com



Commercially confidence – This presentation contains ideas and information which are proprietary of VERHAERT, Masters in Innovation®*, it is given in confidence. You are authorized to
open and view the electronic copy of this document and to print a single copy. Otherwise, the material may not in whole or in part be copied, stored electronically or communicated to third
parties without prior agreement of VERHAERT, Masters in Innovation®*.

* VERHAERT, Masters in Innovation is a registered trade name of Verhaert Consultancies N.V.




                                                                   www.mastersininnovation.com
                                                                                                                                                               20.10.2006               Slide 1
IMAGE SIGNAL PROCESSING


                              Agenda
www.mastersininnovation.com




                                        Application   System Definition   Management
                                            &                              Challenges
                                         Market




                                         Spec’s &        Technical
                                                                             Costs
                                        constraints      solutions




                                                                                        20.10.2006   Slide 2
IMAGE SIGNAL PROCESSING

                              Product Development
www.mastersininnovation.com




                                            Integrated    Control
                                              Sensors     Systems




                                          Mechanisms &   Cabinets &
                                            Robotics      Housings




                                                                      20.10.2006   Slide 3
IMAGE SIGNAL PROCESSING

                              Application & Market
www.mastersininnovation.com




                                                        -   1m/s
                                                        -   Working temperature range 15 – 40°C
                                                            Movement tire ± 1 cm
                                                        -
                                                            Req. precision < 20 µm
                                                        -




                                                                              20.10.2006   Slide 4
IMAGE SIGNAL PROCESSING

                              Specification Definition
www.mastersininnovation.com




                                                        1.0508 ⋅ 10 −7 ⋅ x 3 − 7.8461 ⋅ 10 −5 ⋅ x 2 + 1.0339 ⋅ 10 −2 ⋅ x + 166.6760




                                                                                                         20.10.2006          Slide 5
IMAGE SIGNAL PROCESSING

                              Specification Definition
www.mastersininnovation.com




                                            Range     Precision    Repeatability (2 sigma)



                                    Toe     +/- 10°     +/- 0,5’           +/-0,1’



                                  Camber    +/- 10°      +/- 1’            +/-0,1’




                                   Caster   0°-10°       +/- 3’            +/- 1’



                                    SAI     0°-20°      +/- 0,3°          +/- 0,2°




                                                                                        20.10.2006   Slide 6
IMAGE SIGNAL PROCESSING

                                  Specification Definition
                              -    Pricing                                              Pricing
www.mastersininnovation.com




                                                                8.000




                                                                6.000

                                                                                                             Processing
                                                         Euro


                                                                                                             Camera
                                                                4.000
                                                                                                             Total



                                                                2.000

                                  Target pricing
                                                                   0
                                                                        Giga Ethernet             Firewire




                              -    Options
                                         -   2 x DSP with IEEE 1394                                                       -> custom design
                                         -   1 x Floating point DSP + 1 FPGA with IEEE 1394                               -> COTS
                                         -   1 x Fixed point DSP + 1 FPGA with IEEE 1394                                  -> custom design
                                         -   ...




                                                                                                                             20.10.2006      Slide 7
IMAGE SIGNAL PROCESSING

                                  Thermal Challenges
                              -   Thermal expansion of distance CCD – light source i.e.
                                   -   Normal distance of 250 mm
www.mastersininnovation.com




                                   -   Working temperature range 15 – 40°C
                                   -   I.e. 3 micron error introduced




                                                                                          20.10.2006   Slide 8
IMAGE SIGNAL PROCESSING


                                  Optical Challenges
                              -   Optical errors
www.mastersininnovation.com




                                   -   Speckle
                                   -   Optical elements
                                         - Aberration
                                         - Non linearity




                                                            20.10.2006   Slide 9
IMAGE SIGNAL PROCESSING

                                      Processing Challenges
                                  -       Hard Realtime vs. Soft Realtime
www.mastersininnovation.com




                              Carnegie Mellon University                               RTOS Real-time Performance
                              18-849b Dependable Embedded Systems                         by John A. Carbone, VP,
                              Kanaka Juvva                                                              Marketing
                              http://www.ece.cmu.edu/~koopman/des_s99/real_time/                Express Logic, Inc




                                                                                   20.10.2006         Slide 10
IMAGE SIGNAL PROCESSING

                                   System Definition

            -                 Hard Realtime vs. Soft Realtime
www.mastersininnovation.com




                               -    guaranteed worst-case response times

                               -    1m/s                              30 fps (1image 10 Mb per 33,3 ms)

                               -    1280*1024 , 8 bit, 30 fps         300 Mbps              1.0508 ⋅ 10 −7 ⋅ x 3 − 7.8461 ⋅ 10 −5 ⋅ x 2 + 1.0339 ⋅ 10 −2 ⋅ x + 166.6760


                               -    300 Mbps                          5 x fixed point
                                                                      32 bit integer




                                                                                                                              20.10.2006             Slide 11
IMAGE SIGNAL PROCESSING


                              Algorithms
www.mastersininnovation.com




                                                        20.10.2006   Slide 12
IMAGE SIGNAL PROCESSING

                                 Algorithms
-                             Noise filtering                                                              - Smoothing
www.mastersininnovation.com




-                             Binning                                                                      - Denoising
                                                 50
                                                 50

                                                                                                           - Compensate for saturation
                                                100
                                                100

                                                150
                                                150
-                             Erode                                                                        - Decrease noise
                                                200
                                                200
                                                                                                            (Shrinking object)
                                                250
                                                250

-                             Dilate                                                                       - Increase signal
                                                300
                                                300

                                                                                                           (Object growing)
                                                350

                                                400
                                                400
-                             Multiply                                                                     - AND operation
                                                450
                                                450
                                                      50   100   150   200   250   300   350   400   450
                                                      50   100   150   200   250   300   350   400   450
-                             Classification                                                               - Which points belong together ?

-                             Fitting                                                                      - Polynominal fit




                                                                                                                     20.10.2006   Slide 13
IMAGE SIGNAL PROCESSING

                                  Algorithms
                              -   Dilate & Erode
www.mastersininnovation.com




                                                            20.10.2006   Slide 14
IMAGE SIGNAL PROCESSING

                              Algorithms
                              - Noise filtering   - fixed, non recursive, line based
www.mastersininnovation.com




                              - Binning           - fixed point or integer, no recursive, line based

                              - Erode             - fixed point or integer, no recursive, line based

                              - Dilate            - fixed point or integer, no recursive, line based

                              - Multiply          - binary operation, pixel based

                              - Classification    - recursive search algorithm, pixel based

                              - Fitting           - floating point



                                                                                              20.10.2006   Slide 15
IMAGE SIGNAL PROCESSING


                              Processing
                                     Matlab - CPU Power
www.mastersininnovation.com




                                                          Noise Filtering
                                                          Bin by thresholding
                                                          Bin by local maxima                         Comparison of Processing speed
                                                          Erode & Dilate
                                                                                          1000.00
                                                          Multiplication
                                                          Classify
                                                          Polynomial fit
                                                                                           100.00




                                                                                Seconds
                                                                                  Log
                                                                                            10.00




                                                                                             1.00
                                                                                                    MATLAB -    PC     Single DSP Dual DSP    DSP &
                                                                                                      PC                                      FPGA




                                                                                                                          20.10.2006         Slide 16
IMAGE SIGNAL PROCESSING

                              Processing Requirements
                              -   MMACS – Million Multiply and Accumulate operations per second
                                  (example)
www.mastersininnovation.com




                                   -   Noise filtering convolution with 16 x 1 Byte window filter
                                   -   1280 * 1024 @ 30 fps
                                   -   I.e. 1280 * 1024 * 30 * 16 = 630 MMAC’s per second.


                              -   Floating Point Multipliers in hardware
                                   -   PC & ARM: 1 multiplier
                                   -   DSP: upto 6 multipliers
                                   -   FPGA: built to requirement


                              -   Test of noise filtering on PC P4 2.8GHZ, 1Gbyte RAM
                                   -   Processing time = [13.22 – 18.78] ms for a 492x460 image
                                   -   I.e. for a 1280*1024 image will this be about 100ms.




                                                                                                    20.10.2006   Slide 17
IMAGE SIGNAL PROCESSING

                              PowerPC
                                                        -   4000 - 8000 MFLOPS
                                                        -   1 Floating point multiplier
www.mastersininnovation.com




                                                        -   1 Fixed point multiplier
                                                        -   RAM access upto 128 bit @ 200 Mhz




                                                                                20.10.2006   Slide 18
IMAGE SIGNAL PROCESSING

                                  Floating Point DSP

                              -    1800 – 3600 MFLOPS
www.mastersininnovation.com




                              -    Upto 6 Floating point multipliers @ 32 bit
                              -    2 Fixed point multipliers
                              -    RAM access 32 bit @ 100 Mhz
                              -    2 MB (TI) upto 3 MB (Analog) memory onboard




                                                                                 20.10.2006   Slide 19
IMAGE SIGNAL PROCESSING


                              Fixed Point DSP
                               -   3200 – 8000 MIPS
                               -   4000 MMACs @ 32 bit
www.mastersininnovation.com




                               -   2 MB (TI) upto 3 MB memory onboard (Analog)
                               -   RAM access 64 bit @ 133 Mhz (TI)




                                                                                 20.10.2006   Slide 20
IMAGE SIGNAL PROCESSING


                              Memory requirements
                              -   Bandwidth to External Memory - Theoretical limits:
www.mastersininnovation.com




                                   -   ARM interface to external RAM is 32 bit @ 400 Mhz

                                   -   Floating Point DSP interface to external RAM is 32 bit standard @ 100MHz:
                                         - 3200Mbps: 320 images/sec can be read or written to the external memory. At 30fps:
                                           (320/30)/ 2 (read/write) = 5 image handlings per cycle.


                                   -   Fixed Point DSP interface to external RAM is 64 bit standard @ 133MHz:
                                         - 8512 Mbps: 851 images/sec can be read or written to the external memory. At 30fps
                                           (851/30) / 2 (read/write) = 14 image handlings per cycle.


                                   -   FPGA interface external RAM is DDR2 32 bit @ 400 Mhz
                                         - 12800Mbps: 1280 images/sec can be read or written to the external memory. At 30fps
                                            (12800/30)/ 2 (read/write) = 21 image handlings per cycle.




                                                                                                            20.10.2006   Slide 21
IMAGE SIGNAL PROCESSING

                              Processing selection
                              -   DSP for floating operations
www.mastersininnovation.com




                              -   Algorithm steps for FPGA parallel implementation :
                                   -   Denoising with filter, binning steps and erode & dilate image processing
                                   -   AND functions of 2 or more images
                                   -   (optionally) classification     1000.0


                                                                       100.0

                                                                                                                                                           MATLAB
                                                                        10.0
                                                                                                                                                           PC
                                                                                                                                                           DSP
                                                                          1.0
                                  Gain of 10 ms due to fixed                                                                                               DSP & FPGA
                                                                                                                                          l
                                                                                                                                 it
                                                                                    g     g              e                y
                                                                                                               n                        ta
                                                                                                 a
                                                                                                                       sif ial f
                                                                                                      lat atio
                                                                                r in ldin xim                                         To
                                                                                                   Di               as
                                                                             lte       o                     c
                                                                                             a
                                                                          Fi                                     Cl         m
                                                                                                           li
                                                                                     sh                                   no
                                                                                          l m d e & ltip
                                             op’s in FPGA             ise th re oca                                   oly
                                                                                                      u
                                                                                               o
                                                                    No b y                   Er    M                P
                                                                                        l
                                                                                     by
                                                                        in         n
                                                                                Bi
                                                                      B




                                                                                                                                              20.10.2006    Slide 22
IMAGE SIGNAL PROCESSING

                              Management
                              -   Trade off Performance – Cost – Development Risk
                              -   Algorithm vs. Processing hardware
www.mastersininnovation.com




                                      - Camera’s: 1 Mpixels, 2 Mpixels
                                      - More precision in algorithms & calibration
                                      - More subpixel resolution 1/25    1/50
                                        less smoothing, better peak detectors, better fittings
                                        more floating point operations, more MMACs required, ....




                                                                                                    20.10.2006   Slide 23
IMAGE SIGNAL PROCESSING

                              Management
                              -   Trade off Performance – Cost – Development Risk
                              -   Frame rate vs. CPU power
www.mastersininnovation.com




                                      - Camera’s: 15 fps       30 fps     60 fps
                                      - Doubling required processing power, ....
                                      - More statistical algorithms that improve accuracy




                                                                                            20.10.2006   Slide 24
IMAGE SIGNAL PROCESSING

                              Management of Development
www.mastersininnovation.com




                                -   Multi core DSP vs. DSP and FPGA

                                -   COTS image processing board vs. Custom image processing board

                                -   TI, meanwhile, has a roadmap in place to quickly move the 'C6x floating platform
                                    to 3,000 MMAC, said Rick Rienhart, 'C6000 product line manager at TI in
                                    Houston.

                                -   TI expects to produce devices achieving speeds of 3 trillion instructions per
                                    second by 2010.

                                -   In the future processing power will as good as for for free (More’s Law)




                                                                                                       20.10.2006   Slide 25
IMAGE SIGNAL PROCESSING

                              Conclusions
www.mastersininnovation.com




                                -   Choice of DSP + FPGA on COTS board

                                -   Waiting for DSP implementations with GMAC processing

                                -   Algorithms is mathematics

                                -   Manage development risks by early stage breadboarding.

                                -   System cost impacts earning capabilities in our prior application and its transfer
                                    opportunties to other market applications, thus interact with business
                                    development from the start of the program.




                                                                                                       20.10.2006   Slide 26
IMAGE SIGNAL PROCESSING
www.mastersininnovation.com




                                                                   Verhaert New Products & Services nv
                                                                   Hogenakkerhoekstraat 21
                                                                   9150 Kruibeke
                                                                   Belgium
                                                                   Tel +32 (0)3 250 19 00
                                                                   Fax +32 (0)3 254 10 08
                                                                   www.verhaert.com
                                                                   info@verhaert.com




                                               www.mastersininnovation.com
                                                                                             20.10.2006   Slide 27

More Related Content

Similar to Image Signal Processing: Technical Solutions and Costs

Packet shaper datasheet 81
Packet shaper datasheet 81Packet shaper datasheet 81
Packet shaper datasheet 81Zalli13
 
Packet shaper datasheet 81
Packet shaper datasheet 81Packet shaper datasheet 81
Packet shaper datasheet 81Zalli13
 
Magento Imagine eCommerce, Day 2, Yoav Kutner CTO
Magento Imagine eCommerce, Day 2, Yoav Kutner CTOMagento Imagine eCommerce, Day 2, Yoav Kutner CTO
Magento Imagine eCommerce, Day 2, Yoav Kutner CTOvarien
 
Track g when did test - da integrated
Track g   when did test - da integratedTrack g   when did test - da integrated
Track g when did test - da integratedchiportal
 
Cryptographic Data Splitting and Cloud Computing
Cryptographic Data Splitting and Cloud ComputingCryptographic Data Splitting and Cloud Computing
Cryptographic Data Splitting and Cloud ComputingGovCloud Network
 
Glassfish Overview for Sogeti 20090225
Glassfish Overview for Sogeti 20090225Glassfish Overview for Sogeti 20090225
Glassfish Overview for Sogeti 20090225Eugene Bogaart
 
Visure Solutions Requirements Definition and Management
Visure Solutions Requirements Definition and ManagementVisure Solutions Requirements Definition and Management
Visure Solutions Requirements Definition and ManagementVisure Solutions
 
Move SAP to Cloud in 3 Easy Steps
Move SAP to Cloud in 3 Easy StepsMove SAP to Cloud in 3 Easy Steps
Move SAP to Cloud in 3 Easy StepsAppZero
 
2008 brokerage 04 smart vision system [compatibility mode]
2008 brokerage 04 smart vision system [compatibility mode]2008 brokerage 04 smart vision system [compatibility mode]
2008 brokerage 04 smart vision system [compatibility mode]imec.archive
 
2008 brokerage 04 smart vision system [compatibility mode]
2008 brokerage 04 smart vision system [compatibility mode]2008 brokerage 04 smart vision system [compatibility mode]
2008 brokerage 04 smart vision system [compatibility mode]imec.archive
 
Visure Solutions Requirements Engineering_The word in a nutshell - Ulf Sandberg
Visure Solutions Requirements Engineering_The word in a nutshell - Ulf SandbergVisure Solutions Requirements Engineering_The word in a nutshell - Ulf Sandberg
Visure Solutions Requirements Engineering_The word in a nutshell - Ulf SandbergVisure Solutions
 
Putnam f01
Putnam f01Putnam f01
Putnam f01anissa18
 

Similar to Image Signal Processing: Technical Solutions and Costs (20)

Abrige catlogue
Abrige catlogueAbrige catlogue
Abrige catlogue
 
Packet shaper datasheet 81
Packet shaper datasheet 81Packet shaper datasheet 81
Packet shaper datasheet 81
 
Packet shaper datasheet 81
Packet shaper datasheet 81Packet shaper datasheet 81
Packet shaper datasheet 81
 
Magento Imagine eCommerce, Day 2, Yoav Kutner CTO
Magento Imagine eCommerce, Day 2, Yoav Kutner CTOMagento Imagine eCommerce, Day 2, Yoav Kutner CTO
Magento Imagine eCommerce, Day 2, Yoav Kutner CTO
 
DST group
DST groupDST group
DST group
 
Innovative Space Mechatronics Verhaert Space
Innovative Space Mechatronics Verhaert SpaceInnovative Space Mechatronics Verhaert Space
Innovative Space Mechatronics Verhaert Space
 
Lte asia 2011 s niri
Lte asia 2011 s niriLte asia 2011 s niri
Lte asia 2011 s niri
 
Track g when did test - da integrated
Track g   when did test - da integratedTrack g   when did test - da integrated
Track g when did test - da integrated
 
Virtual Box Aquarium May09
Virtual Box Aquarium May09Virtual Box Aquarium May09
Virtual Box Aquarium May09
 
Cryptographic Data Splitting and Cloud Computing
Cryptographic Data Splitting and Cloud ComputingCryptographic Data Splitting and Cloud Computing
Cryptographic Data Splitting and Cloud Computing
 
Glassfish Overview for Sogeti 20090225
Glassfish Overview for Sogeti 20090225Glassfish Overview for Sogeti 20090225
Glassfish Overview for Sogeti 20090225
 
Improvement e13 link
Improvement e13 linkImprovement e13 link
Improvement e13 link
 
Visure Solutions Requirements Definition and Management
Visure Solutions Requirements Definition and ManagementVisure Solutions Requirements Definition and Management
Visure Solutions Requirements Definition and Management
 
ApresentaçãO Improvement 11 E Link
ApresentaçãO Improvement 11 E LinkApresentaçãO Improvement 11 E Link
ApresentaçãO Improvement 11 E Link
 
Move SAP to Cloud in 3 Easy Steps
Move SAP to Cloud in 3 Easy StepsMove SAP to Cloud in 3 Easy Steps
Move SAP to Cloud in 3 Easy Steps
 
3D Scanner Case Study Verhaert
3D Scanner Case Study Verhaert3D Scanner Case Study Verhaert
3D Scanner Case Study Verhaert
 
2008 brokerage 04 smart vision system [compatibility mode]
2008 brokerage 04 smart vision system [compatibility mode]2008 brokerage 04 smart vision system [compatibility mode]
2008 brokerage 04 smart vision system [compatibility mode]
 
2008 brokerage 04 smart vision system [compatibility mode]
2008 brokerage 04 smart vision system [compatibility mode]2008 brokerage 04 smart vision system [compatibility mode]
2008 brokerage 04 smart vision system [compatibility mode]
 
Visure Solutions Requirements Engineering_The word in a nutshell - Ulf Sandberg
Visure Solutions Requirements Engineering_The word in a nutshell - Ulf SandbergVisure Solutions Requirements Engineering_The word in a nutshell - Ulf Sandberg
Visure Solutions Requirements Engineering_The word in a nutshell - Ulf Sandberg
 
Putnam f01
Putnam f01Putnam f01
Putnam f01
 

More from Verhaert Masters in Innovation

Software language over the last 50 years, what will be next (by Pieter Zulian...
Software language over the last 50 years, what will be next (by Pieter Zulian...Software language over the last 50 years, what will be next (by Pieter Zulian...
Software language over the last 50 years, what will be next (by Pieter Zulian...Verhaert Masters in Innovation
 
Geospatial technologies, the evolution and impact on our daily life (by Nicol...
Geospatial technologies, the evolution and impact on our daily life (by Nicol...Geospatial technologies, the evolution and impact on our daily life (by Nicol...
Geospatial technologies, the evolution and impact on our daily life (by Nicol...Verhaert Masters in Innovation
 
Advanced human interfaces, the underestimated enabler for innovation (by Bert...
Advanced human interfaces, the underestimated enabler for innovation (by Bert...Advanced human interfaces, the underestimated enabler for innovation (by Bert...
Advanced human interfaces, the underestimated enabler for innovation (by Bert...Verhaert Masters in Innovation
 
The first humanoid robot, wabot 1 (by Robrecht Van Velthoven)
The first humanoid robot, wabot 1 (by Robrecht Van Velthoven)The first humanoid robot, wabot 1 (by Robrecht Van Velthoven)
The first humanoid robot, wabot 1 (by Robrecht Van Velthoven)Verhaert Masters in Innovation
 
The government as launching customer, a great opportunity for companies (by R...
The government as launching customer, a great opportunity for companies (by R...The government as launching customer, a great opportunity for companies (by R...
The government as launching customer, a great opportunity for companies (by R...Verhaert Masters in Innovation
 
Landing on the moon, the impact and future opportunities (by Sam Waes)
Landing on the moon, the impact and future opportunities (by Sam Waes)Landing on the moon, the impact and future opportunities (by Sam Waes)
Landing on the moon, the impact and future opportunities (by Sam Waes)Verhaert Masters in Innovation
 
Building an innovation culture, steering individual and team behavior (by Möb...
Building an innovation culture, steering individual and team behavior (by Möb...Building an innovation culture, steering individual and team behavior (by Möb...
Building an innovation culture, steering individual and team behavior (by Möb...Verhaert Masters in Innovation
 
Is the start-up way of working really different than the corporate one (by Fr...
Is the start-up way of working really different than the corporate one (by Fr...Is the start-up way of working really different than the corporate one (by Fr...
Is the start-up way of working really different than the corporate one (by Fr...Verhaert Masters in Innovation
 
Is the house of quality still a valid model to manage innovation (by Dany Rob...
Is the house of quality still a valid model to manage innovation (by Dany Rob...Is the house of quality still a valid model to manage innovation (by Dany Rob...
Is the house of quality still a valid model to manage innovation (by Dany Rob...Verhaert Masters in Innovation
 
How to shape your innovation ecosystem to create impact in your organization ...
How to shape your innovation ecosystem to create impact in your organization ...How to shape your innovation ecosystem to create impact in your organization ...
How to shape your innovation ecosystem to create impact in your organization ...Verhaert Masters in Innovation
 
The evolution of the bicycle industry 50 years after eddy merckx' victory (by...
The evolution of the bicycle industry 50 years after eddy merckx' victory (by...The evolution of the bicycle industry 50 years after eddy merckx' victory (by...
The evolution of the bicycle industry 50 years after eddy merckx' victory (by...Verhaert Masters in Innovation
 
The acceleration of Artificial Intelligence (by Jochem Grietens)
The acceleration of Artificial Intelligence (by Jochem Grietens)The acceleration of Artificial Intelligence (by Jochem Grietens)
The acceleration of Artificial Intelligence (by Jochem Grietens)Verhaert Masters in Innovation
 
The drivers of value creation, 50 years of research (by Dany Robberecht)
The drivers of value creation, 50 years of research (by Dany Robberecht)The drivers of value creation, 50 years of research (by Dany Robberecht)
The drivers of value creation, 50 years of research (by Dany Robberecht)Verhaert Masters in Innovation
 
Multi-sided business models in smart cities (IoT Convention 2019)
Multi-sided business models in smart cities (IoT Convention 2019)Multi-sided business models in smart cities (IoT Convention 2019)
Multi-sided business models in smart cities (IoT Convention 2019)Verhaert Masters in Innovation
 
Dany Robberecht - The benefits of cross industry innovation
Dany Robberecht - The benefits of cross industry innovationDany Robberecht - The benefits of cross industry innovation
Dany Robberecht - The benefits of cross industry innovationVerhaert Masters in Innovation
 
Space 4.0 and the Belgian start-up ecosystem by Omar Mohout
Space 4.0 and the Belgian start-up ecosystem by Omar MohoutSpace 4.0 and the Belgian start-up ecosystem by Omar Mohout
Space 4.0 and the Belgian start-up ecosystem by Omar MohoutVerhaert Masters in Innovation
 

More from Verhaert Masters in Innovation (20)

Technology watch - AI in chemical industry
Technology watch - AI in chemical industryTechnology watch - AI in chemical industry
Technology watch - AI in chemical industry
 
Software language over the last 50 years, what will be next (by Pieter Zulian...
Software language over the last 50 years, what will be next (by Pieter Zulian...Software language over the last 50 years, what will be next (by Pieter Zulian...
Software language over the last 50 years, what will be next (by Pieter Zulian...
 
Geospatial technologies, the evolution and impact on our daily life (by Nicol...
Geospatial technologies, the evolution and impact on our daily life (by Nicol...Geospatial technologies, the evolution and impact on our daily life (by Nicol...
Geospatial technologies, the evolution and impact on our daily life (by Nicol...
 
Advanced human interfaces, the underestimated enabler for innovation (by Bert...
Advanced human interfaces, the underestimated enabler for innovation (by Bert...Advanced human interfaces, the underestimated enabler for innovation (by Bert...
Advanced human interfaces, the underestimated enabler for innovation (by Bert...
 
The first humanoid robot, wabot 1 (by Robrecht Van Velthoven)
The first humanoid robot, wabot 1 (by Robrecht Van Velthoven)The first humanoid robot, wabot 1 (by Robrecht Van Velthoven)
The first humanoid robot, wabot 1 (by Robrecht Van Velthoven)
 
The government as launching customer, a great opportunity for companies (by R...
The government as launching customer, a great opportunity for companies (by R...The government as launching customer, a great opportunity for companies (by R...
The government as launching customer, a great opportunity for companies (by R...
 
Landing on the moon, the impact and future opportunities (by Sam Waes)
Landing on the moon, the impact and future opportunities (by Sam Waes)Landing on the moon, the impact and future opportunities (by Sam Waes)
Landing on the moon, the impact and future opportunities (by Sam Waes)
 
Building an innovation culture, steering individual and team behavior (by Möb...
Building an innovation culture, steering individual and team behavior (by Möb...Building an innovation culture, steering individual and team behavior (by Möb...
Building an innovation culture, steering individual and team behavior (by Möb...
 
The era of pretotyping has arrived (by Kevin Douven)
The era of pretotyping has arrived (by Kevin Douven)The era of pretotyping has arrived (by Kevin Douven)
The era of pretotyping has arrived (by Kevin Douven)
 
Is the start-up way of working really different than the corporate one (by Fr...
Is the start-up way of working really different than the corporate one (by Fr...Is the start-up way of working really different than the corporate one (by Fr...
Is the start-up way of working really different than the corporate one (by Fr...
 
Behind the waterfall methodology (by Jan Buytaert)
Behind the waterfall methodology (by Jan Buytaert)Behind the waterfall methodology (by Jan Buytaert)
Behind the waterfall methodology (by Jan Buytaert)
 
Is the house of quality still a valid model to manage innovation (by Dany Rob...
Is the house of quality still a valid model to manage innovation (by Dany Rob...Is the house of quality still a valid model to manage innovation (by Dany Rob...
Is the house of quality still a valid model to manage innovation (by Dany Rob...
 
How to shape your innovation ecosystem to create impact in your organization ...
How to shape your innovation ecosystem to create impact in your organization ...How to shape your innovation ecosystem to create impact in your organization ...
How to shape your innovation ecosystem to create impact in your organization ...
 
The evolution of the bicycle industry 50 years after eddy merckx' victory (by...
The evolution of the bicycle industry 50 years after eddy merckx' victory (by...The evolution of the bicycle industry 50 years after eddy merckx' victory (by...
The evolution of the bicycle industry 50 years after eddy merckx' victory (by...
 
The acceleration of Artificial Intelligence (by Jochem Grietens)
The acceleration of Artificial Intelligence (by Jochem Grietens)The acceleration of Artificial Intelligence (by Jochem Grietens)
The acceleration of Artificial Intelligence (by Jochem Grietens)
 
The drivers of value creation, 50 years of research (by Dany Robberecht)
The drivers of value creation, 50 years of research (by Dany Robberecht)The drivers of value creation, 50 years of research (by Dany Robberecht)
The drivers of value creation, 50 years of research (by Dany Robberecht)
 
Multi-sided business models in smart cities (IoT Convention 2019)
Multi-sided business models in smart cities (IoT Convention 2019)Multi-sided business models in smart cities (IoT Convention 2019)
Multi-sided business models in smart cities (IoT Convention 2019)
 
Space for Artificial Intelligence
Space for Artificial IntelligenceSpace for Artificial Intelligence
Space for Artificial Intelligence
 
Dany Robberecht - The benefits of cross industry innovation
Dany Robberecht - The benefits of cross industry innovationDany Robberecht - The benefits of cross industry innovation
Dany Robberecht - The benefits of cross industry innovation
 
Space 4.0 and the Belgian start-up ecosystem by Omar Mohout
Space 4.0 and the Belgian start-up ecosystem by Omar MohoutSpace 4.0 and the Belgian start-up ecosystem by Omar Mohout
Space 4.0 and the Belgian start-up ecosystem by Omar Mohout
 

Recently uploaded

Unblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesUnblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesSinan KOZAK
 
CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):comworks
 
Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationSlibray Presentation
 
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024BookNet Canada
 
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr LapshynFwdays
 
Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Enterprise Knowledge
 
Install Stable Diffusion in windows machine
Install Stable Diffusion in windows machineInstall Stable Diffusion in windows machine
Install Stable Diffusion in windows machinePadma Pradeep
 
Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024BookNet Canada
 
Unlocking the Potential of the Cloud for IBM Power Systems
Unlocking the Potential of the Cloud for IBM Power SystemsUnlocking the Potential of the Cloud for IBM Power Systems
Unlocking the Potential of the Cloud for IBM Power SystemsPrecisely
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsAndrey Dotsenko
 
APIForce Zurich 5 April Automation LPDG
APIForce Zurich 5 April  Automation LPDGAPIForce Zurich 5 April  Automation LPDG
APIForce Zurich 5 April Automation LPDGMarianaLemus7
 
Understanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitectureUnderstanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitecturePixlogix Infotech
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Scott Keck-Warren
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024Scott Keck-Warren
 
Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Mattias Andersson
 
New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024BookNet Canada
 
Bluetooth Controlled Car with Arduino.pdf
Bluetooth Controlled Car with Arduino.pdfBluetooth Controlled Car with Arduino.pdf
Bluetooth Controlled Car with Arduino.pdfngoud9212
 

Recently uploaded (20)

DMCC Future of Trade Web3 - Special Edition
DMCC Future of Trade Web3 - Special EditionDMCC Future of Trade Web3 - Special Edition
DMCC Future of Trade Web3 - Special Edition
 
Unblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesUnblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen Frames
 
CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):CloudStudio User manual (basic edition):
CloudStudio User manual (basic edition):
 
Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck Presentation
 
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
 
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
"Federated learning: out of reach no matter how close",Oleksandr Lapshyn
 
Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024
 
Install Stable Diffusion in windows machine
Install Stable Diffusion in windows machineInstall Stable Diffusion in windows machine
Install Stable Diffusion in windows machine
 
Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
 
Unlocking the Potential of the Cloud for IBM Power Systems
Unlocking the Potential of the Cloud for IBM Power SystemsUnlocking the Potential of the Cloud for IBM Power Systems
Unlocking the Potential of the Cloud for IBM Power Systems
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
 
APIForce Zurich 5 April Automation LPDG
APIForce Zurich 5 April  Automation LPDGAPIForce Zurich 5 April  Automation LPDG
APIForce Zurich 5 April Automation LPDG
 
Understanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitectureUnderstanding the Laravel MVC Architecture
Understanding the Laravel MVC Architecture
 
Hot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort Service
Hot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort ServiceHot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort Service
Hot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort Service
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024
 
Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?
 
New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
 
Bluetooth Controlled Car with Arduino.pdf
Bluetooth Controlled Car with Arduino.pdfBluetooth Controlled Car with Arduino.pdf
Bluetooth Controlled Car with Arduino.pdf
 
Vulnerability_Management_GRC_by Sohang Sengupta.pptx
Vulnerability_Management_GRC_by Sohang Sengupta.pptxVulnerability_Management_GRC_by Sohang Sengupta.pptx
Vulnerability_Management_GRC_by Sohang Sengupta.pptx
 

Image Signal Processing: Technical Solutions and Costs

  • 1. IMAGE SIGNAL PROCESSING VERHAERTINNOVATIONDAY – OCTOBER 20th, 2006 www.mastersininnovation.com IMAGE SIGNAL PROCESSING Frederik Wouters frederik.wouters@verhaert.com www.verhaert.com Commercially confidence – This presentation contains ideas and information which are proprietary of VERHAERT, Masters in Innovation®*, it is given in confidence. You are authorized to open and view the electronic copy of this document and to print a single copy. Otherwise, the material may not in whole or in part be copied, stored electronically or communicated to third parties without prior agreement of VERHAERT, Masters in Innovation®*. * VERHAERT, Masters in Innovation is a registered trade name of Verhaert Consultancies N.V. www.mastersininnovation.com 20.10.2006 Slide 1
  • 2. IMAGE SIGNAL PROCESSING Agenda www.mastersininnovation.com Application System Definition Management & Challenges Market Spec’s & Technical Costs constraints solutions 20.10.2006 Slide 2
  • 3. IMAGE SIGNAL PROCESSING Product Development www.mastersininnovation.com Integrated Control Sensors Systems Mechanisms & Cabinets & Robotics Housings 20.10.2006 Slide 3
  • 4. IMAGE SIGNAL PROCESSING Application & Market www.mastersininnovation.com - 1m/s - Working temperature range 15 – 40°C Movement tire ± 1 cm - Req. precision < 20 µm - 20.10.2006 Slide 4
  • 5. IMAGE SIGNAL PROCESSING Specification Definition www.mastersininnovation.com 1.0508 ⋅ 10 −7 ⋅ x 3 − 7.8461 ⋅ 10 −5 ⋅ x 2 + 1.0339 ⋅ 10 −2 ⋅ x + 166.6760 20.10.2006 Slide 5
  • 6. IMAGE SIGNAL PROCESSING Specification Definition www.mastersininnovation.com Range Precision Repeatability (2 sigma) Toe +/- 10° +/- 0,5’ +/-0,1’ Camber +/- 10° +/- 1’ +/-0,1’ Caster 0°-10° +/- 3’ +/- 1’ SAI 0°-20° +/- 0,3° +/- 0,2° 20.10.2006 Slide 6
  • 7. IMAGE SIGNAL PROCESSING Specification Definition - Pricing Pricing www.mastersininnovation.com 8.000 6.000 Processing Euro Camera 4.000 Total 2.000 Target pricing 0 Giga Ethernet Firewire - Options - 2 x DSP with IEEE 1394 -> custom design - 1 x Floating point DSP + 1 FPGA with IEEE 1394 -> COTS - 1 x Fixed point DSP + 1 FPGA with IEEE 1394 -> custom design - ... 20.10.2006 Slide 7
  • 8. IMAGE SIGNAL PROCESSING Thermal Challenges - Thermal expansion of distance CCD – light source i.e. - Normal distance of 250 mm www.mastersininnovation.com - Working temperature range 15 – 40°C - I.e. 3 micron error introduced 20.10.2006 Slide 8
  • 9. IMAGE SIGNAL PROCESSING Optical Challenges - Optical errors www.mastersininnovation.com - Speckle - Optical elements - Aberration - Non linearity 20.10.2006 Slide 9
  • 10. IMAGE SIGNAL PROCESSING Processing Challenges - Hard Realtime vs. Soft Realtime www.mastersininnovation.com Carnegie Mellon University RTOS Real-time Performance 18-849b Dependable Embedded Systems by John A. Carbone, VP, Kanaka Juvva Marketing http://www.ece.cmu.edu/~koopman/des_s99/real_time/ Express Logic, Inc 20.10.2006 Slide 10
  • 11. IMAGE SIGNAL PROCESSING System Definition - Hard Realtime vs. Soft Realtime www.mastersininnovation.com - guaranteed worst-case response times - 1m/s 30 fps (1image 10 Mb per 33,3 ms) - 1280*1024 , 8 bit, 30 fps 300 Mbps 1.0508 ⋅ 10 −7 ⋅ x 3 − 7.8461 ⋅ 10 −5 ⋅ x 2 + 1.0339 ⋅ 10 −2 ⋅ x + 166.6760 - 300 Mbps 5 x fixed point 32 bit integer 20.10.2006 Slide 11
  • 12. IMAGE SIGNAL PROCESSING Algorithms www.mastersininnovation.com 20.10.2006 Slide 12
  • 13. IMAGE SIGNAL PROCESSING Algorithms - Noise filtering - Smoothing www.mastersininnovation.com - Binning - Denoising 50 50 - Compensate for saturation 100 100 150 150 - Erode - Decrease noise 200 200 (Shrinking object) 250 250 - Dilate - Increase signal 300 300 (Object growing) 350 400 400 - Multiply - AND operation 450 450 50 100 150 200 250 300 350 400 450 50 100 150 200 250 300 350 400 450 - Classification - Which points belong together ? - Fitting - Polynominal fit 20.10.2006 Slide 13
  • 14. IMAGE SIGNAL PROCESSING Algorithms - Dilate & Erode www.mastersininnovation.com 20.10.2006 Slide 14
  • 15. IMAGE SIGNAL PROCESSING Algorithms - Noise filtering - fixed, non recursive, line based www.mastersininnovation.com - Binning - fixed point or integer, no recursive, line based - Erode - fixed point or integer, no recursive, line based - Dilate - fixed point or integer, no recursive, line based - Multiply - binary operation, pixel based - Classification - recursive search algorithm, pixel based - Fitting - floating point 20.10.2006 Slide 15
  • 16. IMAGE SIGNAL PROCESSING Processing Matlab - CPU Power www.mastersininnovation.com Noise Filtering Bin by thresholding Bin by local maxima Comparison of Processing speed Erode & Dilate 1000.00 Multiplication Classify Polynomial fit 100.00 Seconds Log 10.00 1.00 MATLAB - PC Single DSP Dual DSP DSP & PC FPGA 20.10.2006 Slide 16
  • 17. IMAGE SIGNAL PROCESSING Processing Requirements - MMACS – Million Multiply and Accumulate operations per second (example) www.mastersininnovation.com - Noise filtering convolution with 16 x 1 Byte window filter - 1280 * 1024 @ 30 fps - I.e. 1280 * 1024 * 30 * 16 = 630 MMAC’s per second. - Floating Point Multipliers in hardware - PC & ARM: 1 multiplier - DSP: upto 6 multipliers - FPGA: built to requirement - Test of noise filtering on PC P4 2.8GHZ, 1Gbyte RAM - Processing time = [13.22 – 18.78] ms for a 492x460 image - I.e. for a 1280*1024 image will this be about 100ms. 20.10.2006 Slide 17
  • 18. IMAGE SIGNAL PROCESSING PowerPC - 4000 - 8000 MFLOPS - 1 Floating point multiplier www.mastersininnovation.com - 1 Fixed point multiplier - RAM access upto 128 bit @ 200 Mhz 20.10.2006 Slide 18
  • 19. IMAGE SIGNAL PROCESSING Floating Point DSP - 1800 – 3600 MFLOPS www.mastersininnovation.com - Upto 6 Floating point multipliers @ 32 bit - 2 Fixed point multipliers - RAM access 32 bit @ 100 Mhz - 2 MB (TI) upto 3 MB (Analog) memory onboard 20.10.2006 Slide 19
  • 20. IMAGE SIGNAL PROCESSING Fixed Point DSP - 3200 – 8000 MIPS - 4000 MMACs @ 32 bit www.mastersininnovation.com - 2 MB (TI) upto 3 MB memory onboard (Analog) - RAM access 64 bit @ 133 Mhz (TI) 20.10.2006 Slide 20
  • 21. IMAGE SIGNAL PROCESSING Memory requirements - Bandwidth to External Memory - Theoretical limits: www.mastersininnovation.com - ARM interface to external RAM is 32 bit @ 400 Mhz - Floating Point DSP interface to external RAM is 32 bit standard @ 100MHz: - 3200Mbps: 320 images/sec can be read or written to the external memory. At 30fps: (320/30)/ 2 (read/write) = 5 image handlings per cycle. - Fixed Point DSP interface to external RAM is 64 bit standard @ 133MHz: - 8512 Mbps: 851 images/sec can be read or written to the external memory. At 30fps (851/30) / 2 (read/write) = 14 image handlings per cycle. - FPGA interface external RAM is DDR2 32 bit @ 400 Mhz - 12800Mbps: 1280 images/sec can be read or written to the external memory. At 30fps (12800/30)/ 2 (read/write) = 21 image handlings per cycle. 20.10.2006 Slide 21
  • 22. IMAGE SIGNAL PROCESSING Processing selection - DSP for floating operations www.mastersininnovation.com - Algorithm steps for FPGA parallel implementation : - Denoising with filter, binning steps and erode & dilate image processing - AND functions of 2 or more images - (optionally) classification 1000.0 100.0 MATLAB 10.0 PC DSP 1.0 Gain of 10 ms due to fixed DSP & FPGA l it g g e y n ta a sif ial f lat atio r in ldin xim To Di as lte o c a Fi Cl m li sh no l m d e & ltip op’s in FPGA ise th re oca oly u o No b y Er M P l by in n Bi B 20.10.2006 Slide 22
  • 23. IMAGE SIGNAL PROCESSING Management - Trade off Performance – Cost – Development Risk - Algorithm vs. Processing hardware www.mastersininnovation.com - Camera’s: 1 Mpixels, 2 Mpixels - More precision in algorithms & calibration - More subpixel resolution 1/25 1/50 less smoothing, better peak detectors, better fittings more floating point operations, more MMACs required, .... 20.10.2006 Slide 23
  • 24. IMAGE SIGNAL PROCESSING Management - Trade off Performance – Cost – Development Risk - Frame rate vs. CPU power www.mastersininnovation.com - Camera’s: 15 fps 30 fps 60 fps - Doubling required processing power, .... - More statistical algorithms that improve accuracy 20.10.2006 Slide 24
  • 25. IMAGE SIGNAL PROCESSING Management of Development www.mastersininnovation.com - Multi core DSP vs. DSP and FPGA - COTS image processing board vs. Custom image processing board - TI, meanwhile, has a roadmap in place to quickly move the 'C6x floating platform to 3,000 MMAC, said Rick Rienhart, 'C6000 product line manager at TI in Houston. - TI expects to produce devices achieving speeds of 3 trillion instructions per second by 2010. - In the future processing power will as good as for for free (More’s Law) 20.10.2006 Slide 25
  • 26. IMAGE SIGNAL PROCESSING Conclusions www.mastersininnovation.com - Choice of DSP + FPGA on COTS board - Waiting for DSP implementations with GMAC processing - Algorithms is mathematics - Manage development risks by early stage breadboarding. - System cost impacts earning capabilities in our prior application and its transfer opportunties to other market applications, thus interact with business development from the start of the program. 20.10.2006 Slide 26
  • 27. IMAGE SIGNAL PROCESSING www.mastersininnovation.com Verhaert New Products & Services nv Hogenakkerhoekstraat 21 9150 Kruibeke Belgium Tel +32 (0)3 250 19 00 Fax +32 (0)3 254 10 08 www.verhaert.com info@verhaert.com www.mastersininnovation.com 20.10.2006 Slide 27