For obstacle detection and classification, there are currently applications of vision and laser pointer radar, which do not conflict. Lidar does not depend on light, its angle of view is 360 degrees, the calculation is relatively small, and it can be scanned in real time. At present, it is generally used within 100 milliseconds. During the scanning process, the lidar first recognizes the obstacle, knows the position of the obstacle in space, and then classifies it according to the existing obstacle.
For example, cars and people, we divide these obstacles into independent individuals, and then separate the individual individuals to match, so as to classify the obstacles and track the objects. In short, we first remove the ground points to obtain obstacles, segment the obstacles, and then segment the obstacles for classification tracking. In the tracking process, first of all, the point cloud is segmented, and the point cloud is used as the associated target. We know whether the previous frame and the next frame belong to the same object, and then track the target to output target tracking information.
Lidar application status. The first is the lack of sensors. On the one hand, the current lidar is relatively expensive, and the supply cycle is also very long. There are not many companies that can generally use lidar. However, the lack of sensors, the immaturity of the solutions, and the accumulation of talents for point cloud algorithms are too few, and lidar cannot exert its maximum power.
For the unmanned team, in addition to the point cloud algorithm of the lidar, they may also do camera algorithm, millimeter wave radar algorithm and multi-sensor fusion, including positioning, path planning, decision control and car change. They have so many technical points that they are unable to focus, which causes them to be forced to stretch the front.
So for the solution of red laser pointer radar, the Prometheus plan was proposed. We hope that the essence of this plan is an attitude of openness and sharing, and accelerate the commercialization of the entire driverless business. The module of the entire Prometheus project is based on what Lidar can do, including: algorithm modules for positioning, lane line detection, roadside detection, obstacle recognition, obstacle classification and tracking.