IoT practical case using the people counter sensing traffic density build using Oracle IOT cloud
•Download as PPTX, PDF•
1 like•509 views
iot solution using the yolo people counter and the Oracle IoT cloud based on machine learning. by Gertjan van het Hof and ; Kjettil Hennis of AMIS
Recommended
Recommended
More Related Content
What's hot
What's hot (20)
Similar to IoT practical case using the people counter sensing traffic density build using Oracle IOT cloud
Similar to IoT practical case using the people counter sensing traffic density build using Oracle IOT cloud (20)
More from Getting value from IoT, Integration and Data Analytics
More from Getting value from IoT, Integration and Data Analytics (20)
Recently uploaded
Recently uploaded (20)
IoT practical case using the people counter sensing traffic density build using Oracle IOT cloud
- 2. 2 WHO AM I? Gertjan van het Hof IoT Solution Architect IoT Enthusiast Linkedin: linkedin.com/in/gertjanvanhethof Twitter: @GertjanvhHof Email: gertjan.van.het.hof@amis.nl Phone: +31 (0)6 29540619 nlOUG Tech Experience 2018: People Counter
- 3. 3 WHO AM I? Kjettil Hennis IoT Developer IoT Enthusiast Linkedin: linkedin.com/in/kjettil-hennis-059910138 Email: kjettil.hennis@amis.nl Phone: +31 (0)6 23494454 nlOUG Tech Experience 2018: People Counter
- 4. AGENDA • Introduction • Use Cases • Architecture & Design Decisions • People Counter Hardware • First approach • Image recognition • Final solution • Oracle IoT Cloud • Backend integration 4nlOUG Tech Experience 2018: People Counter
- 5. AMIS AS IOT IMPLEMENTATION PARTNER • References: • Smart Industry (Industry IoT) • Condition monitoring • Asset management • Predictive maintenance • Smart lubrication solution • Smart Office Cleaning nlOUG Tech Experience 2018: People Counter 5
- 6. USE CASES 6nlOUG Tech Experience 2018: People Counter
- 7. USE CASES • (Waiting row) counting • Festival • Ticket desk • Deposit desk • Entrance • Transportation • Buildings • Shops • Wild detection • Cattle density • Traffic density 7
- 8. ARCHITECTURE & DESIGN DECISIONS 8nlOUG Tech Experience 2018: People Counter
- 9. GOAL • Learn more about Oracle IoT Cloud. • Learn more about Image detection. • MVP: • Counting people in front of AMIS Conference Stand and send this to the Oracle IoT Cloud. • Back-end integration to drive something in the physical world. Titel van de presentatie 9
- 10. AMIS TECH EXPERIENCE CONFERENCE STAND 10nlOUG Tech Experience 2018: People Counter
- 11. PEOPLE COUNTER DESIGN DECISIONS • Oracle IoT Cloud (because of this event) • Security and privacy by Design • Pictures are removed after object count is determined • Only the count is send to the cloud • Edge Computing • Local image processing • Cheap hardware (drawback is slow image detection) • Programming Language: Fit for Purpose • Image recognition library (Python) • Oracle IoT Cloud SDK (JavaScript) 11nlOUG Tech Experience 2018: People Counter
- 12. PEOPLE COUNTER ARCHITECTURE 12 Oracle IoT CloudPeople Counter nlOUG Tech Experience 2018: People Counter
- 13. HARDWARE 13nlOUG Tech Experience 2018: People Counter
- 14. RASPBERRY PI 3B+ 14 • Broadcom BCM2837B0, Cortex-A53 (ARMv8) 64-bit SoC @ 1.4GHz • 1GB LPDDR2 SDRAM • 2.4GHz and 5GHz IEEE 802.11.b/g/n/ac wireless LAN, Bluetooth 4.2, BLE • Gigabit Ethernet over USB 2.0 (maximum throughput 300 Mbps) • Extended 40-pin GPIO header • Full-size HDMI • 4 USB 2.0 ports • CSI camera port for connecting a Raspberry Pi camera • DSI display port for connecting a Raspberry Pi touchscreen display • 4-pole stereo output and composite video port • Micro SD port for loading your operating system and storing data • 5V/2.5A DC power input • Power-over-Ethernet (PoE) support (requires separate PoE HAT) nlOUG Tech Experience 2018: People Counter
- 15. Titel van de presentatie 15
- 16. PEOPLE COUNTER • Take a picture • Analyze the picture • Send the result of analysis to cloud • If above a certain threshold something happens 16nlOUG Tech Experience 2018: People Counter
- 17. FIRST VERSION • Processing • OpenCV Library • Performance issues 17nlOUG Tech Experience 2018: People Counter
- 18. 18nlOUG Tech Experience 2018: People Counter
- 19. 19nlOUG Tech Experience 2018: People Counter
- 20. PEOPLE COUNTER SOLUTION 20 People Counter Oracle IoT CloudOracle IoT SDK nlOUG Tech Experience 2018: People Counter
- 21. YOLO: YOU ONLY LOOK ONCE • Object detection system • Divides images into regions • Predicts bounding boxes and probabilities • Different versions of the library (normal and tiny) • https://www.youtube.com/watch?v=MPU2HistivI4 21nlOUG Tech Experience 2018: People Counter
- 22. Titel van de presentatie 22
- 23. SCRIPTS • Create a script with multiple commands according to idea • To run more fluent: Python script. 23nlOUG Tech Experience 2018: People Counter
- 24. PEOPLE COUNTER TECHNICAL DETAILS 24 People Counter Oracle IoT CloudOracle IoT SDK nlOUG Tech Experience 2018: People Counter
- 25. NODE.JS • Based on Javascript Oracle IoT Hub Client Library (SDK) • Customized for People Counter • Reads file with count information and sends this to Oracle IoT Cloud • Put IoT Hub Provisioning file in program root folder to communicate with Oracle IoT Cloud • Activation in IoT Cloud is automatically done when program is started 25nlOUG Tech Experience 2018: People Counter
- 26. PEOPLE COUNTER TECHNICAL DETAILS 26 People Counter Oracle IoT CloudOracle IoT SDK nlOUG Tech Experience 2018: People Counter
- 27. ORACLE IOT CLOUD 27nlOUG Tech Experience 2018: People Counter
- 28. PEOPLE COUNTER ARCHITECTURE 28 Oracle IoT CloudPeople Counter nlOUG Tech Experience 2018: People Counter
- 29. ORACLE IOT CLOUD ARCHITECTURE 29nlOUG Tech Experience 2018: People Counter
- 30. ORACLE IOT CLOUD PROVISIONING • Provisioning with Oracle Cloud Stack • Provisioning is template based • Choose for Oracle IoT Enterprise Template • This results into: • Oracle IoT Cloud • Content Storage • Backup Storage • Java Cloud Service • Database Cloud Service 30nlOUG Tech Experience 2018: People Counter
- 31. ORACLE IOT CLOUD AFTER PROVISIONING 31nlOUG Tech Experience 2018: People Counter
- 32. PEOPLE COUNTER CLOUD SERVICES DETAILS 32 Start & Stop nlOUG Tech Experience 2018: People Counter
- 33. ORACLE IOT CLOUD 33 Device Management IoT Application nlOUG Tech Experience 2018: People Counter
- 34. APPLICATION MANAGEMENT 34 Application Device Model Exploration 1..* 1..* Integration 1..* • Interface that lets any device communicate with Oracle IoT Cloud Service regardless of its manufacturer or operating system. • The message specification. • A container for configuring an IoT Application. • The data stream of a specific Device Model for the application. • Configuration of aggregations and alerts. • Integration to back office or other Oracle Cloud services. nlOUG Tech Experience 2018: People Counter
- 35. DEMO Oracle IoT Cloud 35nlOUG Tech Experience 2018: People Counter
- 36. Device Model (1/2) Titel van de presentatie 36
- 37. Device Model (2/2) Titel van de presentatie 37
- 38. Application Titel van de presentatie 38
- 39. Applications Titel van de presentatie 39
- 40. Explorations (1/2) Titel van de presentatie 40
- 41. Explorations (2/2) Titel van de presentatie 41
- 42. Data Titel van de presentatie 42
- 43. Exploration Data ON Titel van de presentatie 43
- 44. Integrations Titel van de presentatie 44
- 45. BACKEND INTEGRATION 45nlOUG Tech Experience 2018: People Counter
- 46. PEOPLE COUNTER ARCHITECTURE 46 Oracle IoT CloudPeople Counter nlOUG Tech Experience 2018: People Counter
- 47. IFTTT 47nlOUG Tech Experience 2018: People Counter
- 48. Questions? Titel van de presentatie 48 QUESTIONS?
- 49. REFERENCES 49 • Oracle IoT Cloud: https://cloud.oracle.com/en_US/iot • Oracle IoT Client Library: http://www.oracle.com/technetwork/indexes/downloads/iot-client-libraries-2705514.html • Yolo: https://pjreddie.com/darknet/yolo/ • Yolo Python library (optimized for RPi): https://github.com/digitalbrain79/darknet-nnpack • IFTTT https://ifttt.com/ nlOUG Tech Experience 2018: People Counter
Editor's Notes
- Buying hardware and other necessities: To do analysis on: Use of Raspberry 3B+ best specs of all Pi’s Most important specs are the CPU and RAM memory Which has a camera interface to connect a pi camera module
- Bought a case to protect and encase the Pi and its camera module Included a wall mount
- To show affinity and practice with IoT AMIS decided to something with IoT. Mention idea: Picture Analyze picture Send result Something happens
- Processing is an IDE to create computational visualizations; Java Based (Contains) OpenCV Library: library used for computer visioning Background removal Movement detection Blob calculation (first idea to use) Early on perfomance issues Show Photos of possibilities
- Right to you: Background Subtraction Left: Face Detection More possibilities
- Own test of implementation of the opencv library Resolution was very low to run more smoothly
- Our Solution Yolo Python Save NodeJS
- Normal accurate, but slow Tiny less accurate but fast More Pixels more time Used a Pi optimized version.
- This is an example. Yolo library can identify many kind of objects (like people, dog, truck, aeroplane, etc.) In our solution we count the number of occurrences of people.
- Pipe Output of analysis through reader which stores number occurences of certain object Only run analysis and print number of certain object, it runs smoothly Kind of loop, which also takes a photo and send input Python We also have created a local web page showing the foto with boxes. Technical Details: Come to the AMIS stand
- Provisioning file is used for activation of the device in Oracle IoT Cloud. Provisioning file is for securing communication between Oracle IoT Cloud and the device.
- Summary of the hardware solution.