Track Altitude Difference Detection System And Its Laser Spot Image Position Recognition Method

Cranes play an important role in today's industrial economy.In order to ensure the safe operation of cranes,it is necessary to inspect crane tracks regularly.At present,track detection mostly adopts manual mode,using total station,tower ruler and other equipments.The detection efficiency is low and the accuracy is poor.There are great potential safety hazards for the inspectors during high-altitude operation.At the same time,the number of cranes is huge in China that the task of detection is very heavy.It is urgent to develop a crane track detection system to replace manual detection methods.In this paper,the detection method and system of crane track altitude difference and the method of laser spot image position recognition are studied by using laser projection and vision measurement methods.The main research contents and conclusions are as follows:Firstly,the automatic detection method and system of orbit altitude difference of "one station + two robots" are established.By collecting the spot images in different positions of track to recognize the position change of the spot image and to detect the track altitude difference;The track altitude difference detection system includes imaging robot,launching robot and PC;Robots cooperate to collect and transmit images.PC monitors the motion state of the robot and recognizes spot image;The image acquisition device uses OV2640 camera and STM32 as the main control chip of the robot;The main control chip read the image and transmits it to PC by ESP8266 WIFI module for image position recognition and height difference calculation.Secondly,the propagation characteristics of laser beams are analysised.Based on the energy distribution law of the beam propagation process,the mathematical models of different positions of the beam are established.The propagation characteristics of the beam in the atmosphere and the energy distribution of the spot after imaging were analyzed.The results are as follows: The spot formed by Gauss beam is regularcircular;Spot radius increases due to divergence during beam propagation;The atmospheric inhomogeneity results in the aberration of the beam field,while the strong noise around the spot is caused by atmospheric scattering,characterized by high brightness and large area continuity;The use of thin convex lens can improve the focus of the beam without changing the shape of the spot.Then,the position recognition method of laser spot image is studied.The camera calibration is used to eliminate the image distortion and determine the corresponding relationship between the pixel scale and the actual scale.Aiming at the problem that the scattering noise can not be effectively suppressed in the traditional spot image recognition processing,a gray-level comprehensive correction method of spot image based on the spot centroid and region is established by enhancing the details of the spot edge in the image frequency domain.By searching the center of mass of the spot,setting the threshold to find the spot area,and by reducing the gray value of the image outside the spot area,the difference between the image sharpness and the gray distribution of the image is enhanced,and the spot image is segmented accurately by using the improved two-dimensional Otsu algorithm.By comparing the actual effects of various edge extraction operators,Canny operator is selected to recognize the edge of the spot,and the concentric circle of the edge points is constructed by projective transformation of the circle.The least square circle fitting method based on the geometric structure edge is established.The fitting process is analyzed in depth.The center coordinate of the spot is corrected by iteration method,and the position recognition of the laser spot image is completed.Finally,the experimental study of laser spot image position recognition and track altitude difference detection was carried out.When the beam propagation distance is30 m,the experimental analysis of the recognition method of spot proposed in this paper shows that The maximum error does not exceed ±0.4mm.Compared with the traditional brightness threshold spot recognition method,the accuracy is improved by0.13 mm,and the spot position recognition results are more stable.The test resultsshow that the detection error of single track height difference is-1.0mm~1.7mm,and that of double track height difference is-1.2mm~1.8mm.Its detection accuracy is no more than ±2mm,and it can replace manual means for automatic track detection.The research results of this paper provide technical basis and experimental basis for further research and development of crane track detection system.