Key Detection Method Based On 3D Point Cloud Excessive Packaging

Excessive packaging of goods is banned repeatedly,the traditional means of the quality supervision department of the manual measurement and evaluation experience and calculation inference has been unable to meet the needs of daily inspection.On the other hand,computer vision,after years of development,has been widely applied in the three-dimensional measurement and other fields.Aiming at the difficult situation of excessive packaging detection,through the research on several key problems in computer vision,this paper designs and implements a kind of intelligent detection system of excessive packaging of goods based on stereo vision,with advanced technical means to solve the difficult problem of excessive packaging of goods inspection.such as formation,transformation and structure of cloud point,and volume calculations.The main works of this dissertation are as follows:1.Dimensional point cloud to the 3D point cloud conversion method.In the research and study computer vision,through the understanding of the translation,rotation,and projection operations and plane and spatial coordinates understanding.Compare the existing rotary axis calibration method to achieve and improve the existing rotary axis calibration method,the use of a computer can be detected on the checkerboard,effectively reduced the corner detection and identification of the process,reducing the amount of calculation,it has greatly improved the accuracy of the calibration.2.Multi-angle rotation stitching of point clouds.To put products into rotation with a rotating platform,the cloud point of the surface information of goods stored in the form of Dimensional lattice information from image captured by a binocular camera.Studying and researching space reconstruction principles,and the establishment of a suitable rotary table coordinate system.Effective and accurate to restore outline of the object in the computer,and to achieve multi-view point cloud data splicing,from the experimental results,the reconstruction3.Structured treatment of scattered points cloud.Introduced 3D point cloud structured methods that more mainstream,for example point by point insertion method based on Delaunay algorithm and Graham scan algorithm based on the convex hull,through the above-mentioned study and analysis of algorithms.Proposed and improved structured approach Graham scan algorithm.By three methods of comparison and analysis,the improved algorithm Graham scan point cloud contour has a better descriptive,reduction of data points and more time-saving and convenience.4.Volume calculation and optimization of results.This chapter explains the procedure for point cloud space volume calculated.By determining the projection plane,and the point cloud space in a direction perpendicular to the plane of projection and slice each projection.With this construction method,and then find the projection surface area of each slice by vector cross product method,the final volume of goods can be drawn.Experiments show that the volume of paper solving method compared with the conventional method,a higher accuracy and shorter program run time.5.Excessive packaging judgment.The shape of Goods in the field of view is a regular contour or irregular contour.irregular contour is divided into approximate regular volume and commuted irregular volume.Through the introduction of porosity calculation formula,Through the introduction of porosity calculation formula,the three different types porosity calculation method of goods were explained.The Algorithm Based on the volume of merchandise convex hull of 3D point cloud data proposed by this paper,which can quickly calculate the test and packaging of goods and accurate judgment porosity.By national standards and regulations for comparison,can accurately and quickly determine whether the excessive packaging of goods.Experimental results show that the test results fully meet the quality inspection accuracy and real-time requirements.