Encoder + VPAC Application

Table of Contents

• Introduction
• Supported plaforms
• Data flow
• Steps to run the application on J721e EVM

Introduction

This application demonstrates the use of VISS node and TI's 2A (AE & AWB) algorithm in multi-channel mode, where the raw output is encoded to file. The dataflow is

1. Configure image sensors using sensor driver f/w APIs
2. Acquire RAW images from camera sensors
3. Process the images on VISS
4. Run AE (Auto Exposure) and AWB (AutoWhiteBalance) algorithms, using H3A output from VISS
5. Run LDC on the output of VISS
6. Scale LDC output using MSC and arrange in a mosaic
7. Encode the output using H.264 video encoder to filesystem

Any number of cameras from 1-8 can be selected by the user. The application assumes that all cameras are identical.

Supported plaforms

Platform	Linux x86_64	Linux+RTOS mode	QNX+RTOS mode	SoC
Support	NO	YES	YES	J721e

Data flow

Steps to run the application on J721e EVM

1. Build the application and related libraries as mentioned in Build Instructions
2. Connect four D3-IMX390 RCM sensor boards on cam0, cam1, cam2 and cam3 input ports of the Fusion1 board.
3. Use the is_interactive flag to run the demo in an interactive mode which allows the user to print performance characteristics on the UART console.
4. Run the app as shown below

   cd /ti_fs/vision_apps
   . ./vision_apps_init.sh
   ./run_app_encode.sh

5. The processed output for one of the captured channel is is encoded to file and saved as /opt/vision_apps/test_data/output/encode_output.h264
6. Copy this file to linux PC and play h.264 raw file using ffplay as shown below.

   ffplay -f h264 encode_output.h264