A question on the usage of drivable space detection module

[ry4nzhu]

Hi @hangqiu,

Thanks for open-sourcing your great work. Recently I am interested in the road surface detection and drivable space detection part of the code and want to ask a few questions on how to use them.

Based on my investigation of the codebase, OccupancyGrid::ConvertPCAndSetOccupancyGrid_ManualPlaneModel(int frameSeq, cv::Mat &pc, sl::Mat &slpc, float A, float B, float C, float D) is used to determine whether grids are occupied. However, I am not quite sure how to set the function arguments (and their meanings) for float A, float B, float C, float D. From your code, I see that you are using -0.1, -0.924138, 0.2, 1.1 in mPathPlanner->PlanPath_AStar_ManualRoadModel(currFrame.frameSeq, tmp,slpc,-1,4,-2,30,-0.1, -0.924138, 0.2, 1.1,HorizonZMax,HorizonZMin);/// FOR PSA ROOF FOOTAGE, proves to be working on BMW roof mount, pretty robust and consistent, which further passes the four floats to the OccupancyGrid::ConvertPCAndSetOccupancyGrid_ManualPlaneModel() function. Can you elaborate on how I should set these parameters if I want to use it on other point cloud data?

Besides, from the paper, it seems that the evaluation is performed on point clouds collected from stereo cameras. Do you think the same code (drivable space detection) can also work with LiDAR point clouds?
Thanks in advance.

Best,
Ruiyang