The Application Of Photogrammetry And Computer Vision In Mining Similar Material Simulation Experiment

Deformation monitoring of mining similar material simulation experiment is a necessary content for simulation experiment of large mining engineering. The content mainly provides safety guide for actual mining. In order to realize the deformation monitoring, we need artificially layout a lot of monitoring sites on the model. By measuring each monitoring site’s object space coordinates in every working face. We can get deformation characteristic of overlying strata in mining. From the movement characteristic of rock, we can determine activity cycle of the roof’s periodic pressure.In view of the boundedness that the present measurement methods exist, the article puts forward an arithmetic based on photogrammetry and computer vision to realize three-dimensional reconstruction and deformation measurement of similar material simulation experiment. This arithmetic has three steps:First, camera calibration. The main purpose of camera calibration is to calculate parameters describing geometric model between the object points and the corresponding image points in 2-D images. By result can get projection relationship in object space and image space. Meanwhile, it can solve deformation caused by the ordinary camera in the edge of images- realize image preprocessing. By comparing and analyzing the merit and demerit of many kinds of common calibration algorithms. We choice Zhang Zhengyou plane camera calibration to finish this work finally.Second, Image coordinates measurement of high precision. The efficiency and accuracy of image coordinates measurement arithmetic is a key in three-dimensional reconstruction arithmetic this paper puts forward. Monitoring points and encoded points are made of back light reflective materials. Monitoring points are circular target and encoded points are coded marking. Using Canny operator and ellipse fitting arithmetic realizes Image coordinates measurement of high precision.Third, Calculating space coordinates of monitoring points. By comparing and analyzing the insufficient of traditional three-dimensional reconstruction arithmetic, combining with the study object plane model in this experiment, finally realize three-dimensional reconstruction based on homographic matrix. This arithmetic finishes based on single image. It can avoid the complex process of homonymy points matching.After confirming the theoretical model of three-dimensional reconstruction arithmetic, we choice a proper computer language to realize. This paper uses MATLAB and C++ language mainly.As the research object of the physical model Dongxia coal established, we can get one set of result by above algorithm. If the difference value meets tolerance demand, we can use the result to draw deformation curve of every working face. First, we compare this curve with the result of high precision measuring means. According to a preliminary inspection, we can ensure the algorithm the article puts forward reasonable. Then, if the two kinds of curves are same, we should judge that the curve is same with the actual deformation of physical model or not, and ensure the curve agrees with actual situation.We can prove this arithmetic’s feasibility and calculation speed and accuracy by real example. The example results indicate: the arithmetic has realized to calculate space coordinate of monitoring points automatically. Its speed is fast and accuracy is high. The result can meet similar material simulation experiment.