Research On The Construction Machinery Travel Speed Detection Based On Machine Vision

The real-time detection of the construction machinery travel speed and slip plays a key role in the speed adaptive control, constant speed control, rational use of power, improving its operating efficiency and operating quality. It is difficult to detect the actual traveling speed of the construction machinery by the conventional contact speed detection method because of the effected of the road conditions and other factors. An indirect speed detection method based on the machine vision which can detect the actual traveling speed in time was proposed in this paper.According the principle of video velocity measurement, the camera was mounted on the traveling construction machinery and an indirect speed detection model based on machine vision was established. The internal and external parameters of the camera were acquired by camera calibration. The road image collected in real time in the actual traveling process of construction machinery was analyzed. According to the influence of the camera vibration on image quality, the motion compensation and the purpose of image stabilization were achieved by using the gray projection method. Several common target detection technologies were introduced and the advantage of optical flow calculation method based on the Lucas-Kanade in this project was analyzed. The optical flow calculation between the images of the successive frames was achieved by using the pyramid layered method during the experiment, and the main direction principle of the optical flow was put forward to determine the actual traveling direction of construction machinery. Finally, the average optical flow speed was solved by using the statistical method, and according to the relationship between the image coordinate and world coordinate, the average optical flow speed was converted to the actual traveling speed of construction machinery. The testing speed which using intelligent car to simulate construction machinery has1.56%error with the actual speed. The experimental results show that this method can correctly reflect the actual traveling speed of construction machinery.