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

4.1 System Flow

This application is built based on TI’s edgeai-gst-apps project that includes all required infrastructures for an end-to-end processing application. It configures the model, creates the video processing pipeline using gstreamer, makes inference calls to the compiled model, and post processes the model’s output for visualization.

GStreamer is an open-source multimedia platform that links media processing systems as a pipeline. Some of the media handling processes include video capturing, recording, and streaming. TI provides a suit of gstreamer plugins that allow offloading some of the media processing and the deep learning inference to the hardware accelerators. The edgeai-gst-apps provides required software to configure and use the gstreamer pipeline. It is used to run the out of box demos on the AM62A EVM and its source code is available for developers to use.

The defect detection demo updates the post processing application code while keeping the gstreamer wrapper and the inference call part the same. The features added to the application code include object tracker, performance evaluation, graphical dashboard, and colorful bounding boxes. Figure 4-1 shows the full system flow updated for the defect detection demo with the application code and the gstreamer. The application has been tested with two types of cameras including a USB camera and an IMX219 camera with a CSI interface.

GUID-20230630-SS0I-R28H-ZNDG-4NTBZPV1KHN2-low.svg Figure 4-1 Full System Flow Including Application Code and gstreamer Pipeline With Two input Options: a USB camera or a CSI IMX219 camera. The defect detection application supports the two input options.