3D depth map - 3D depth map - 2025.2 English

A 3D depth map represents the distance of each point in a scene from the camera. It is generated from a stereo disparity map, which measures the pixel shift (disparity) between two slightly different images captured by stereo cameras. Since closer objects have larger disparities and farther objects have smaller disparities, these disparities are mathematically converted into depth values, producing a depth map that encodes how far each part of the scene is in 3D space.

𝑍 = (𝑓⋅𝐵)/𝑑

𝑍 : depth (distance from the camera to the object)

𝑓 : focal length of the camera (in pixels)

𝐵 : baseline, i.e., distance between the two cameras

𝑑 : disparity (pixel shift between left and right images)

API Syntax

template <int SRC_T,
        int DST_T,
        int ROWS,
        int COLS,
        int NPC,
        int XFCVDEPTH_disp = _XFCVDEPTH_DEFAULT,
        int XFCVDEPTH_out = _XFCVDEPTH_DEFAULT>
void depth3D(xf::cv::Mat<SRC_T, ROWS, COLS, NPC, XFCVDEPTH_disp>& _disp_mat,
        xf::cv::Mat<DST_T, ROWS, COLS, NPC, XFCVDEPTH_out>& _depth_mat,
        float focal_len,
        float  base_dis)

The following table describes the template and the function parameters.

Resource Utilization

The following table summarizes the resource utilization in different configurations, generated using Vitis HLS 2024.2 tool for the xcu200-fsgd2104-2-e, to process a 1280x720, three channel image.

Performance Estimate

The following table summarizes the performance estimates in different configurations, generated using Vitis HLS 2024.2 tool for the xcu200-fsgd2104-2-e, to process a 1280x720, three channel image.