A Fast Spherical Coordinate LiDAR Simulation Method

Abstract

The invention relates to a rapid spherical coordinate laser radar simulation method, and discloses a rapid simulation method for a laser radar. A CPU and a GPU perform collaborative operation; the CPUperforms partial initialization operation; the GPU performs coordinate conversion on vertex coordinates of triangular surfaces in a vertex shader; in a geometric shader, each triangular surface is changed into a rectangle composed of two triangles, and the coordinates of three vertexes ABC of the original triangular surface are mounted and output; and in a fragment shader, ray detection is performed on the triangular surface, collision point coordinates are calculated, and an output of the laser radar is calculated and mapped to a space (R, G, B, A). According to the method, traditional render pipelines are used, most types of display cards are compatible, and the acceleration performance of the modern display card can be obtained; and meanwhile, the problem of equidistant sampling in laser radar simulation is solved, and the high-speed and high-precision equiangular sampling is realized.

Description

technical field [0001] The invention belongs to the field of simulation technology, in particular to a fast spherical coordinate laser radar simulation method. Background technique [0002] At present, lidar is widely used in autonomous driving, robotics, ships, aviation and other fields. However, in the current mainstream simulation system, the simulation of lidar is still at the stage of using a depth camera or using a CPU to calculate Raycast (ray detection). [0003] For example, in Microsoft's AirSim simulation system and LG's LG Simulator simulation system, depth cameras are used as lidar simulation methods. [0004] There is a serious problem in this simulation method: when the virtual camera performs sampling, it performs equidistant sampling according to the screen pixels, such as figure 1 As shown, the beam of the camera is actually emitted from the origin, and the direction is from the origin of the camera to the position of each pixel on the NearClip. Therefore...

AI Technical Summary

Problems solved by technology

[0006] In view of this, the present invention proposes a fast simulation method for lidar, which still uses the traditional rendering pipeline, is compatible with most types of graphics cards, and can obtain the acceleration performance of modern graphics cards, while solving the problem of equidistant sampling in lidar simulation. The problem of realizing high-speed and high-precision equiangular sampling

Images