Research On Measurement Method Of Irregular Object Volume Based On Binocular Stereo Vision

In large bulk cargo bases such as coal yards,mines,granaries,garbage treatment plants and earthworks,it is a difficult problem to count the volume of irregular production materials such as coal and grain.At present,manual measurement and instrumental measurement are mainly adopted,which is difficult,time-consuming,not real-time,and can not be measured manually in some cases.The cost of using special instruments such as coal pan,total station and portable laser measuring instrument is high,and the human influence is great.In order to solve this kind of volume measurement problem,this paper proposes a method of using binocular stereo vision measurement method to solve the volume measurement problem of irregular objects.This method uses stereo vision principle,image acquisition,image processing,stereo matching,three-dimensional reconstruction and other technologies.It is a_fast non-contact volume measurement method.The principle of measurement is to use binocular camera to collect the image of the object to be measured on the stacking surface,acquire the disparity information of the left and right images through stereo matching,reconstruct the three-dimensional information of all points in the camera's field of vision by di'sparity information,divide the object to be measured into several parts,calculate the volume of each part by using the idea of integration,and finally obtain the volume of the obj ect to be measured.Experiments show that the volume measurement method based on binocular stereo vision can solve the problem of irregular object volume measurement.The main research contents are as follows:(1)The measurement principle of binocular stereo vision system and the calibration principle of binocular stereo system are studied.The binocular system is calibrated by Zhang Zhengyou calibration method.The final calibration result has a reprojection error of 0.17 pixels,which can be used in calibration,stereo matching and volume measurement to meet the measurement requirements.(2)A volume measurement method based on semi-global matching algorithm(SGM)is proposed.The semi-global matching algorithm(SGM)and the global matching algorithm have the same matching strategy,which is divided into four parts:cost calculation,cost aggregation,disparity calculation and disparity optimization.This paper uses OpenCV to implement semi-global matching algorithm and obtains the dense image to be tested.The parallax map reconstructs the three-dimensional coordinate information of the image to be measured by the principle of triangulation,and uses the coordinate information to solve the volume of the object to be tested.(3)A volume measurement method based on graph theory minimum cutting algorithm(Graph Cuts)is proposed.The stereo matching method based on graph theory minimum cut is to convert the stereo matching into multi-label problem by constructing pixel labels,overcome the problem of occlusion in the left and right cameras,improve the matching precision of the algorithm,and obtain the object to be measured in the dense parallax map.The contour is more complete and smooth,and the three-dimensional coordinate information of the image to be tested is reconstructed by the principle of triangulation.The experimental results show that the accuracy of the measurement is higher.(4)Based on dynamic objects on the conveyor belt,a dynamic measurement scheme is proposed.In the case of ensuring that the adjacent images have a certain overlap,the images of the objects to be measured on the conveyor belt are collected at equal intervals,and the adjacent images collected by the left and right cameras are respectively spliced by the image splicing technique,and the spliced according to a certain size,etc.According to the continuous processing,each part of the body is accumulated and acquired to obtain the volume transported by the conveyor belt in the whole time period,and the splicing of adjacent images is achieved,and the effect is good.