SPRADC9 july   2023 AM62A3 , AM62A7

 

  1.   1
  2.   Abstract
  3.   Trademarks
  4. 1Introduction
    1. 1.1 Defect Detection Demo Summary
    2. 1.2 AM62A Processor
    3. 1.3 Defect Detection Systems
    4. 1.4 Conventional Machine Vision vs Deep Learning
  5. 2Data Set Preparation
    1. 2.1 Test Samples
    2. 2.2 Data Collection
    3. 2.3 Data Annotation
    4. 2.4 Data Augmentation
  6. 3Model Selection and Training
    1. 3.1 Model Selection
    2. 3.2 Model Training and Compilation
  7. 4Application Development
    1. 4.1 System Flow
    2. 4.2 Object Tracker
    3. 4.3 Dashboard and Bounding Boxes Drawing
    4. 4.4 Physical Demo Setup
  8. 5Performance Analysis
    1. 5.1 System Accuracy
    2. 5.2 Frame Rate
    3. 5.3 Cores Utilization
    4. 5.4 Power Consumption
  9. 6Summary
  10. 7References

1.3 Defect Detection Systems

A defect detection systems is used to inspect products and detect abnormalities such as irregular shape, broken part, cracks, and so forth. It feeds information to a filtering system (for example, robotic arm) to separate units that are not accepted for packaging or for the following production processes. Depending on the types of the products and the expected defects, various types of input can be used including camera, laser, ultrasound, and so forth. This demo focuses on vision-based inputs. A typical defect detection system consists of the following components:

  • A camera with appropriate resolution and frame rate.
  • Computational unite to perform AI model inference and other processes such as logging results, perform statistics calculation, connect to server if required to log information, networking, and so forth. This demo uses AM62A SoC.
  • Conveyor system to deliver units under test.
  • Screen (monitor) to show information as needed. Live feed of camera with detection results overlaid is a common example.
  • A mechanical system (could be a robotic arm) to filter rejected units based on AI model decision.
  • Other parts such as alarm and networking solution as needed.

Using deep learning to process the vision-based input in defect detection systems has several advantages over using conventional machine vision algorithms. The next section presents a detailed comparison between the two options.