A Research On Engineering Machine Speed Detect System Based On Machine Vision

When the construction machinery undercarriages generate traction, slip occurs inevitably. But the slip will cause the loss of construction machinery walking speed to form a slip power and lower the machinery’s operating and increase the machinery’s energy consumption. So it is crucial to get the construction machinery to work in top working condition by controlling and adjusting slip through the study of walking speed. But in practice, it is difficult to achieve accurate measurements due to the impact of surface adhesion properties, operating load, tire pressure and other comprehensive factors. Therefore, the high-precision measurements of actual walking speed and slip rate have become an urgent problem to be solved. Based on this, the paper analyzed the existing speed measurement technologies and designed and developed engineering machinery walking speed measurement system based on machine vision and set up a mathematical model of walking speed and image sequences and single blurred image. By its own "eyes" to identify ground image information, the global motion vector of adjacent image sequences and single motion blurred image frames are researched. Combined with the installation parameters of mechanical engineering, walking speed relative to the ground is calculated. On this basis, a measurement test platform based on machine vision is independently developed, and field tests are conducted and the accuracy and stability of the visual velocity measurement algorithm are verified. Thus the measurement of slip rate is achieved.The main research items are as following.(1) A walking speed measurement and control system is developed based on machine vision. According to the main functions of the system, the paper mainly researched the performance and type selection principle of such hardware devices as movement mechanism, industrial camera, camera lens, light source and image capture card. A modular design of software is made at the same time. A two-dimensional control of movement of the moving body and real-time image acquisition and processing are achieved. Thus the software and hardware infrastructure for the follow-up theoretical and experimental researches are provided.(2) This paper researches the effects of the calibration of machine vision systems on visual measurement. It also defines the importance of the coordinate system calibration and analyzes the difficulties of the traditional calibration methods in practical applications. Based on selection of a small distortion of the optical lens, a relative calibration method is adopted. Taking the standard gauge blocks as calibration objects, without the need for industrial camera’s internal and external parameters, industrial camera’s object plane resolution can be calibrated quickly and easily. Calibration error is 0.0046 mm.(3) According to the impact of uneven illumination on the image quality, after the study of illumination- basic reflection model, this paper presents the function fitting algorithm based on the theory of function approximation. By the background illumination variation in B-spline fitting estimate image, the original image is stretched with the division operation. Thus the degraded images generated from uneven illumination are effectively corrected and the contrasts of the images are improved. Contrasted with the traditional homomorphism filtering algorithm, post-processing of information entropy algorithm has increased by 26.6%. At the same time the details of the image are highlighted and image contrast is enhanced. Thus the matching accuracy of the sequence images’ correlation curve is greatly improved.(4) The correlation of adjacent frames in the sequence of images is studied. A velocity measurement model is established based on the sequence of images. A method for measuring the engineering machines’ walking speed is proposed based on image sequences: using the relevance of same ground scene in adjacent frame image overlapping and the time of inter-regional displacement to account the walking speed of engineering machinery. An estimation algorithm for the global motion vector of image shift is studied. Multi-window gray projection algorithm is proposed to obtain a stable global motion vectors. And experimental research is made in the clay and sandy soil- two typical pavement surfaces- at different speeds. The results show that within a given speed range, the relative error rate of is less than 7%, the adjacent frame image operation takes about 15 ms. The correctness and feasibility of the theoretical model and the algorithm are verified.(5) The effects of velocity and exposure time for motion blurred image are studied. A mathematical model of motion blur scale and velocity is established. A measuring method based on motion blurred image for the speed of engineering machinery is proposed. With the features of motion blur’s spacing and orientation of the images appearing in the spectrum of dark stripes, Radon algorithm is developed to estimate the direction and scale motion blur. The results show that within a given speed range, the relative error rate is less than 10% and the operation speed of a single frame of motion blur image is about 30 ms. The correctness and feasibility of the algorithm is verified.(6) Experiments and studies on the slip rate measurement. According to the 5 different slip phenomenon, by analyzing the measurement of two key parameters(walking speed and theory walking speed) in the measurement of slip rate, combined with the measurement of machine vision algorithms developed in this paper, a slip ratio measurement test platform based on machine vision is designed. Practical measurements are conducted through this platform. The accuracy and stability of the visual velocity measurement algorithms are verified. A measure of instantaneous slip ratio is achieved. Within a given range of ground conditions and speed, the measurement accuracy of slip rate can be controlled within 1.75%.