Part Number: TDA2HG
I have already seen the FCW routines. It first uses SFM to acquire the 3D point cloud, then uses the 3D point cloud to estimate the ground plane, and finally uses the ground plane and the position of the object in the image to get the object distance.
1. I want to know the difference between the ground information estimated by the 3D point cloud and the ground segmented by deep learning. Why not use the semantic segmentation to get the ground.
2. Before calculating the distance of the object in the image, the FCW module needs to input the position of the ground and the object in the image to determine whether the object is on the ground. I want to know how this is judged?
3. If I already know the size and position of the vehicle on the road, how should the actual distance of the vehicle be calculated?
regards, cheng
