StereoVisionForADAS module

The StereoVisionForADAS module is responsible for managing the main processes of binocular vision, including camera initialization, disparity map generation, and target detection analysis.

Module Function

1. Initialize and manage binocular camera parameters.

2. Call StereoMatching to generate disparity map.

3. Use StixelEstimation to extract columnar pixels.

4. Use StixelSegmentation to cluster Stixels and generate bounding boxes.

---

Core Function Analysis

1. InitStereoParam(int nDatasetName)

Initialize the camera internal and external parameters according to the dataset name. Example:

objStereoCamParam.m_dBaseLine = 0.54;
objStereoCamParam.m_dFocalLength = 722;

2. Objectness(Mat& imgLeft, Mat& imgRight)

Main process:

1. Call StereoMatching to generate disparity map.

2. Use StixelEstimation to extract Stixel.

3. Use StixelSegmentation to cluster and generate bounding box.

3. Display(...)

Visualize the results, you can choose to draw the ground area or target bounding box.

---

Usage example

StereoCamParam_t param = CStereoVisionForADAS::InitStereoParam(KITTI);
CStereoVisionForADAS stereoVision(param);
stereoVision.Objectness(imgLeft, imgRight);
stereoVision.Display();

---

Notes

• Make sure the incoming camera parameters match the dataset.

• Dependent submodules include StereoMatching, StixelEstimation, and StixelSegmentation.

For more details, please refer to Main implementation details.