Magnetic suspension track obstacle detection method based on point cloud

Abstract

The invention relates to a magnetic suspension track obstacle detection method based on point cloud. The magnetic suspension track obstacle detection method comprises the following steps: 1) buildinga magnetic suspension track vehicle-mounted mobile scanning system; 2) obtaining original reference point cloud of the magnetic suspension rail surface and to-be-detected point cloud data through mobile scanning; 3) establishing K-D tree index and matching the K-D tree index with to-be-detected point cloud data to obtain obstacle point cloud; 4) performing Euclidean clustering processing on the obstacle point cloud to obtain an obstacle point cloud clustering set; And 5) finally outputting the outline boundary and position of each independent obstacle to realize automatic detection of the obstacle on the magnetic suspension track surface, compared with the prior art, the method has the advantages that the environmental perception capability is stronger, and the information support effect can be realized on the operation and maintenance safety of the magnetic suspension.

Description

technical field [0001] The invention relates to a method for detecting obstacles on a magnetic levitation track, in particular to a method for detecting obstacles on a magnetic levitation track based on point clouds. Background technique [0002] Maglev transportation relies on electromagnetic force to make the train float on the track and "fly close to the ground". It has the advantages of fast speed, low power consumption, no friction, low noise, and long life. As a new cutting-edge rail transit technology, it has been increasingly used attention of many countries. With the improvement of the design speed requirements for maglev, it is necessary to detect obstacles on the maglev track surface and eliminate them in time, so as to avoid the high-voltage current causing obstacles to be energized, thereby causing certain damage to the track surface; The danger caused by collision with obstacles during driving ensures the safety of driving routes. [0003] At present, there a...

AI Technical Summary

Problems solved by technology

(1) Obstacle detection based on binocular stereo vision is currently the mainstream detection method for obstacles. After obtaining the left and right images of obstacles through binocular cameras, threshold segmentation and other methods are used to obtain obstacle target areas, and then feature point detection, Extraction, stereo matching, etc., to obtain obstacle contours, positions, and depth information, which has the advantages of strong flexibility and low cost, but it is greatly affected by weather, such as light, rain and snow and other complex environments; (2) based on laser radar The detection method uses the laser transmitter to emit the laser light, then receives the reflected light after the laser hits the surface of the object, and measures the distance by measuring the time difference between emission and reception, and can obtain the emission characteristics, distance and speed of the target, etc. information, and finally detect obstacles through obstacle and road segmentation, clustering, etc., which have the characteristics of simple principle, high precision, and strong anti-interference; (3) detection methods based on multi-sensor fusion, such as using visual sensors to obtain obstacle Image, after obtaining the approximate position of the obstacle by extracting the edge and features of the obstacle, using a laser sensor, infrared sensor or ultrasonic sensor, etc., using the combination of the measured signal and the visual sensor to describe the environment, which is less affected by environmental changes. It has strong robustness, reliable perceptual information and high accuracy, and overcomes the shortcoming of insufficient information of a single sensor, but the perfect fusion algorithm of different sensors is difficult and costly; (4) based on the optical flow method, a single camera is used in different Obtain the surrounding environment information at all times, estimate the optical flow through a sequence of images, and segment the optical flow field by using the feature that the obstacle in the optical flow is a mutation, so as to realize obstacle detection, but the optical flow method has a large amount of calculation and high processing time.

Images