Birds Eye View - Evaluating

⇦ Getting Started/Birds Eye View and Libpanorama

Home

Getting Started/How to Get the Code ⇨

Getting Started
Birds Eye View and Libpanorama Evaluating the Product How to Get the Code Quickstart Guide
User Guide
Building and Installation Minimal Application Camera Sources Advanced Usage
Calibration Guide
Understanding the Calibration Process Preparing the Calibration Images The Calibration Tool GUI Performing the Calibration
GStreamer
GstBEV Plugin Generic Pipelines Jetson Pipelines
Performance Measurements
NVIDIA Jetson
Contact Us

This wiki will show you how to run the evaluation version of the project. The evaluation version will be fully functional, except for the following limitations:

• No source code provided.
• Limited evaluation period (the binary won't work 1 month after delivery).
• The eval is licensed under the RidgeRun's Eval License Agreement.

Requesting the Evaluation Version

Please send us an email to support@ridgerun.com with the request and specifying the following information:

• General description of the use case
• Target platform (NXP i.MX8, Qualcomm RB5, NVIDIA Jetson Orin, Xavier, etc...)
• OS Version (Jetpack version, Ubuntu version, Yocto version, etc...)
• Hardware Acceleration (only CUDA and CPU are supported at the time)
• Project constrains:
  • Number of inputs
  • Input size
  • Framerate
  • Output size
  • Lens type

We will use this information to build an evaluation version that runs on your target system within a business week.

Installing Dependencies

Once you have been delivered the evaluation binary, you're ready to proceed with the next steps.

Run the following command to make sure all the runtime dependencies are installed:

sudo apt update
sudo apt install -y \
  libjsoncpp-dev \
  libopencv-core-dev libopencv-video-dev libopencv-highgui-dev libopencv-videoio-dev \
  gstreamer1.0-plugins-bad gstreamer1.0-plugins-good gstreamer1.0-plugins-base \
  gstreamer1.0-libav gstreamer1.0-plugins-ugly \
  wget unzip

If you are using hardware accelerators (such as CUDA), make sure you follow the vendor instructions on how to install them as well.

Testing the Sample Images

Once you have dependencies installed, you can proceed to test the Birds Eye View evaluation binary.

Unpack the provided tarball:

tar xf $EVAL_TARBALL

Along with the executable, you have been provided with some sample images and a calibration file. Run the following command to test the sample calibration:

./birds_eye_view -p birds_eye_view.json

You should see the resulting image as below. Press ESC to stop the execution.

Resulting image after running the evaluation binary with the sample images.

Calibrating your Own System

The evaluation process was designed so that you can test the software with your own images.

Follow the instructions in the Calibration Guide, step by step, to generate a calibration file.

CALIBRATION GUIDE

Once you have generated the calibration file for you custom images and system, proceed as before:

./birds_eye_view -p custom_calibration.json

Sample Capture Pipelines

Once you've proven the system functionality with your calibration images, you can modify the calibration file to capture from a different source, such as videos or cameras.

Generic Video File Capture

filesrc location=$VIDEO_FILE ! decodebin ! queue ! videoconvert ! video/x-raw,format=RGBx ! appsink name=appsink drop=false sync=true qos=false max-buffers=1 max-lateness=-1

Generic V4L2 Camera Capture

v4l2src device=$DEVICE ! queue ! videoconvert ! video/x-raw,format=RGBx ! appsink name=appsink drop=false sync=true qos=false max-buffers=1 max-lateness=-1

Jetson Video File Capture

filesrc location=$VIDEO_FILE ! decodebin ! queue ! nvvidconv ! video/x-raw,format=RGBA ! appsink name=appsink drop=false sync=true qos=false max-buffers=1 max-lateness=-1

Jetson NvArgus Camera Capture

nvarguscamerasrc device-id=$DEVICE ! queue ! nvvidconv ! video/x-raw,format=RGBA ! appsink name=appsink drop=false sync=true qos=false max-buffers=1 max-lateness=-1

Jetson V4L2 Camera Capture

nvv4l2camerasrc device=$DEVICE ! queue ! nvvidconv ! video/x-raw,format=RGBA ! appsink name=appsink drop=false sync=true qos=false max-buffers=1 max-lateness=-1

Troubleshooting

If, for any reason, you can't get the evaluation binary to run, start by inspecting the output of the GStreamer debug:

GST_DEBUG=2 ./birds_eye_view -p birds_eye_view.json

If that doesn't reveal the cause, please send us an email to support@ridgerun.com describing the problem. Please make sure you include the following information:

• Environment information (Platform, OS version, accelerator, etc...).
• GStreamer version (gst-launch-1.0 --gst-version).
• Output of GStreamer debug (GST_DEBUG=2 ./birds_eye_view -p birds_eye_view.json).
• Calibration images,
• Calibration file.
• Any output printed by the binary.