Research And Development On Panel Furniture Location Detection System

For a long time,in panel furniture production of many processes,the furniture of the chop on material handling such as physical processes are done by manual,not only low efficiency increase labor costs,and due to the large furniture plate,easy to cause the workers were injured.Companies want to automate these processes,but they are difficult to implement.In order to realize automatic cutting and unloading,the edge profile of the plate and its center of gravity must first be obtained,which is to obtain the precise three-dimensional coordinates of the board furniture on the conveyor belt.However,it is difficult to realize the automatic acquisition of coordinates accurately,and the difficulty lies in that the three-dimensional positioning of the board furniture is large,the illumination is uneven,the image noise is complex and the accuracy is higher.In view of the above difficulties,this paper discusses and studies the method of obtaining the 3d coordinate of board furniture based on image processing and Kinect2.0 sensor in industrial production line.The main work of this paper is as follows:Firstly,the background,research status and significance of the research are introduced.The structure function and imaging principle of the Kinect2.0 sensor and the advantages of the sensor are introduced in detail.This paper discusses the requirements of the software development platform of this system,and obtains the color image and depth image by developing the Kinect for Windows SDK.Secondly,the image preprocessing algorithm of depth image and color image was detailed designed.By comparing the practical effects of the average filtering,median filtering,the bilateral filter filtering algorithm,eventually choose bilateral filtering algorithm to process the depth image.This method in removing noise at the same time,to preserve the image edge information.The image segmentation method of color image is compared with Canny,Sobel and Laplace.This paper segment the color image based on depth data.By using Camshift method we calculates the center point of the plate.Combined with the depth data obtained,the 3d coordinates of the plate are obtained.Based on the analysis of the principle of small hole imaging,the method of coordinatetransformation in camera imaging is studied.The calibration method of Kinect2.0 sensor color camera and depth camera is introduced,at the same time this paper analyzes the causes of the aberration and introduces the solution of the distortion.Respectively calculate the internal matrix and the outside matrixthe of color cameras and depth cameras outside,and puts forward a kind of by a rotation matrix and a translation vector on the depth of color image and image registration method.Finally,the experimental platform was built,and the three-dimensional coordinates of the plates were obtained through experiments.Compared with the actual measured plate coordinate data.Analyze the cause of error and obtain the depth data error distribution.Adjust the error according to the actual production situation of the factory.The final image field size is 1500 mm × 2000 mm,with the installation height of the Kinect2.0 sensor at about 1200 mm,the precision of the 3d coordinate of the plate is 5mm to meet the industrial demand.