Team Intro
TITLE: Smart Vision Technology Quality Control
TRACK NAME: Robotics
TEAM NAME: Tech Pulse
TEAM MEMBERS: GRISHMA S
(Team leader)
DARSHAN N
(Team member)
DHANUSH H
(Team member)
COLLEGE: ADHIYAMAAN COLLEGE OF ENGINEERING
(AUTONOMOUS), HOSUR
DATE: 19/10/2024
Executive Summary:
OVERVIEW:
Smart vision technology with advanced imaging systems
and algorithms to capture and analyze visual information.
In the context of quantity and quality testing, it helps
automate the quality inspection process by identifying a
product, its quantity and any defects or quality attributes.
IDEA BEHIND THE SOLUTION:
1. Our project streamlines product inspection by
automatically identifying and counting items in
real time.
2. It utilizes advanced technology to assess the
freshness of fruits and vegetables and monitor their
expiration dates.
3. By scanning barcodes, the system retrieves vital
product details, including brand, ingredients, and
nutritional information.
4. The system promptly alerts the team if any product
is expired or compromised, facilitating swift
action.
5. The data analyzed to enhance inventory
management and optimize operational efficiency.
Technical Approach:
TECHNOLOGY STACK:
 PROGRAMMING LANGUAGE: Python
 MODELS: YOLOv8, Convolutional Neural Network (CNN)
 FRAMEWORKS: TensorFlow/Keras
 LIBRARIES: OpenCV, PyTorch
 TOOL: EasyOCR, Matplotlib, Plotly
 API: Barcode Lookup API
 DATABASE: MySQL
PROPOSED SOLUTION:
The proposed system incorporates advanced machine learning and computer
vision techniques for real-time product monitoring and quality control:
Counting Specified Objects (YOLOv8):
• Detects and counts specific objects in real time, excluding irrelevant
classes like humans.
• Uses a predefined polygon zone to focus counting within a designated
area.
Freshness Prediction (CNN):
• Analyzes live video feed to classify the freshness of fruits and vegetables
as either "Healthy" or "Rotten" using a pre-trained CNN model.
• Utilizes smoothing to stabilize predictions across frames.
OCR Expiry Date Detection (EasyOCR):
• Extracts expiry dates from product labels using EasyOCR.
• Triggers alerts if the product is expired.
Brand and Product Details Extraction (Barcode & API):
• Detects and decodes barcodes from product packaging and retrieves
detailed information from an external API.
• Fetches product details such as brand, ingredients, nutritional info, and
packaging details.
Data Analytics:
• The system generates performance analytics and visual graphs from the
stored data, allowing for trend analysis and optimized decision-making.
HARDWARE SPECIFICATIONS:
 Camera with real-time video capture for product analysis and
barcode scanning.
 GPU for efficient deep learning computations with YOLOv8 and
CNN models.
 High-performance storage to manage large datasets for object
detection and freshness classification.
FUTURE SCOPE:
 Robotics for Automated Inspection: Incorporate robotic systems for
automated handling and inspection of products, significantly speeding
up quality checks in warehouses and distribution centers.
 Smart Packaging Solutions: Develop smart packaging that uses
embedded sensors to provide real-time freshness and quality data
directly to the monitoring system.
ITEMS USED TO TRAIN MODEL:
Object detection pretrained model: yolov8l.pt
Freshness detection dataset: fruits and vegetables freshness
Brand and details extraction: BARCODE LOOKUPAPI
CODE:
Object detection and count of specified objects: https://github.com/GrishmaSuresh/object-detection-including-count
Brand and product details extraction: https://github.com/GrishmaSuresh/product_details
Expiry date validation: https://github.com/GrishmaSuresh/expiry_date
Freshness detection: https://github.com/GrishmaSuresh/freshness_detection
VIDEO SIMULATION:
Thank You!

67136f968f1c2_flipkart_robotics_ppt.pptx

  • 2.
    Team Intro TITLE: SmartVision Technology Quality Control TRACK NAME: Robotics TEAM NAME: Tech Pulse TEAM MEMBERS: GRISHMA S (Team leader) DARSHAN N (Team member) DHANUSH H (Team member) COLLEGE: ADHIYAMAAN COLLEGE OF ENGINEERING (AUTONOMOUS), HOSUR DATE: 19/10/2024
  • 3.
    Executive Summary: OVERVIEW: Smart visiontechnology with advanced imaging systems and algorithms to capture and analyze visual information. In the context of quantity and quality testing, it helps automate the quality inspection process by identifying a product, its quantity and any defects or quality attributes. IDEA BEHIND THE SOLUTION: 1. Our project streamlines product inspection by automatically identifying and counting items in real time. 2. It utilizes advanced technology to assess the freshness of fruits and vegetables and monitor their expiration dates. 3. By scanning barcodes, the system retrieves vital product details, including brand, ingredients, and nutritional information. 4. The system promptly alerts the team if any product is expired or compromised, facilitating swift action. 5. The data analyzed to enhance inventory management and optimize operational efficiency.
  • 4.
    Technical Approach: TECHNOLOGY STACK: PROGRAMMING LANGUAGE: Python  MODELS: YOLOv8, Convolutional Neural Network (CNN)  FRAMEWORKS: TensorFlow/Keras  LIBRARIES: OpenCV, PyTorch  TOOL: EasyOCR, Matplotlib, Plotly  API: Barcode Lookup API  DATABASE: MySQL PROPOSED SOLUTION: The proposed system incorporates advanced machine learning and computer vision techniques for real-time product monitoring and quality control: Counting Specified Objects (YOLOv8): • Detects and counts specific objects in real time, excluding irrelevant classes like humans. • Uses a predefined polygon zone to focus counting within a designated area. Freshness Prediction (CNN): • Analyzes live video feed to classify the freshness of fruits and vegetables as either "Healthy" or "Rotten" using a pre-trained CNN model. • Utilizes smoothing to stabilize predictions across frames. OCR Expiry Date Detection (EasyOCR): • Extracts expiry dates from product labels using EasyOCR. • Triggers alerts if the product is expired. Brand and Product Details Extraction (Barcode & API): • Detects and decodes barcodes from product packaging and retrieves detailed information from an external API. • Fetches product details such as brand, ingredients, nutritional info, and packaging details. Data Analytics: • The system generates performance analytics and visual graphs from the stored data, allowing for trend analysis and optimized decision-making. HARDWARE SPECIFICATIONS:  Camera with real-time video capture for product analysis and barcode scanning.  GPU for efficient deep learning computations with YOLOv8 and CNN models.  High-performance storage to manage large datasets for object detection and freshness classification. FUTURE SCOPE:  Robotics for Automated Inspection: Incorporate robotic systems for automated handling and inspection of products, significantly speeding up quality checks in warehouses and distribution centers.  Smart Packaging Solutions: Develop smart packaging that uses embedded sensors to provide real-time freshness and quality data directly to the monitoring system.
  • 5.
    ITEMS USED TOTRAIN MODEL: Object detection pretrained model: yolov8l.pt Freshness detection dataset: fruits and vegetables freshness Brand and details extraction: BARCODE LOOKUPAPI CODE: Object detection and count of specified objects: https://github.com/GrishmaSuresh/object-detection-including-count Brand and product details extraction: https://github.com/GrishmaSuresh/product_details Expiry date validation: https://github.com/GrishmaSuresh/expiry_date Freshness detection: https://github.com/GrishmaSuresh/freshness_detection VIDEO SIMULATION:
  • 6.

Editor's Notes