Font Size: a A A

Research On Automatic Crushed Stone Guidance Method Based On 3D Point Cloud Model

Posted on:2021-02-05Degree:MasterType:Thesis
Country:ChinaCandidate:B DengFull Text:PDF
GTID:2381330647961449Subject:Electrical engineering
Abstract/Summary:PDF Full Text Request
With the development of computer and information processing technology,our mining industry has achieved remarkable results in the direction of intelligent construction.However,compared with the mining development of industrialized countries,my country's mining industry still has deficiencies in the application of computer technology.How to use computers and information processing technology to mine intelligent construction has become a research hotspot.Therefore,based on the broken hammer automatic rock guidance project,this paper uses 3D point cloud modeling technology,3D point cloud filtering and area growth segmentation method to obtain the 3D coordinate information of the stone to realize the automatic hammer guidance.The main contents of this article are as follows:Since the point cloud image collected in the actual underground environment scene is composed of a grid screen and stones,when the plane model is used for filtering,the algorithm will fit the point cloud corresponding to the stones into the plane during the process of fitting the plane The problem of overfitting will now filter out the point cloud of the stone together with the plane.In response to this problem,this paper proposes a method based on background model filtering,using the background invariance of the broken area to extend the two-dimensional plane background subtraction method to the three-dimensional point cloud.Model the background point cloud,binarize the current point cloud after subtracting the background point cloud.Since there is no correlation between the stone point cloud and the background point cloud,after the method of subtraction,the stone point cloud can be perfectly filtered out,and the sparse point cloud can be filtered out with the help of statistical filtering.The point cloud data below is the required point cloud.Under the Windows10 system,combined with open CV vision library and PCL point cloud library to complete the algorithm process development,through experimental verification,it is concluded that the background model filtering algorithm can retain more stone features while filtering the grid screen plane than the planar model filtering.The calculated centroid coordinates are more accurate.The automatic gravel guidance method has a better effect in the visual processing part.It has a good performance in practical applications.In view of the dim underground environment of the mine and the high height of the mine hole in the broken area,this paper studies to determine the basic parameters and hardware layout of the visual system according to the underground environment of the mine,and selects a depth camera based on time of flight(TOF)to complete the visual acquisition part.The choice of hardware,because the working principle of the TOF camera can still run under poor lighting conditions,and its working distance can reach more than 10 meters depending on the device.It can meet the environment of mine underground.Theoretical analysis and actual collected data verify the rationality of the design and determine the design and installation of the vision system platform.Obtain point cloud data through the construction of on-site hardware facilities to verify the effect of the entire process under actual conditions.The final calculated coordinates of the broken point meet the requirements.The realization of the entire process can not only improve the level of mine automation,but also enable the crushing hammer to achieve automatic stone crushing and ensure the safety of personnel operating large industrial and mining machinery.
Keywords/Search Tags:intelligent mine, TOF, 3D point cloud, point cloud filtering, regional growth
PDF Full Text Request
Related items