Development Of A Mosaic Sorting System Based On Machine Vision

Mosaic as a decoration has been widely used in outer wall and floor appearance.Due to gorgeous bright style,it is favored by an increasing number of people in the field of decoration.However,in this stage,mosaic sorting process exits problems which are mainly carried out mainly and low efficiency.Even if a few advanced sorting processes use detection methods of conventional sensor combination to identify sorting target.To a certain extent,it can improve the automation of the sorting system,but can not meet the requirements of mosaic defect detection.With the development of computer graphics technology,machine vision technology that use of cameras to replace the human eye is becoming more and more mature.Because of its visualization,lossless,real-time,versatility and other characteristics,it can compensate for performance of sensor combination detection applied in the mosaic sorting fieldThis article is based on machine vision knowledge to start the research of mosaic target sorting technology.In this process,A set of visual recognition and detection program was proposed,And the research object was identified and detected by detection experiments.1)From the mosaic identification and sorting process,design an system program,build the experimental hardware and control platform,according to the experimental environment,analysis and calculation the parameters of each unit components,select the appropriate model of the product.2)Proposed a set of visual recognition schemes including image gray processing,image filtering,image morphological processing and image segmentation.Based on this visual recognition scheme,the mosaic target contour and surface grayscale histogram feature are extracted,According to the image feature,contour matching and gray value matching are used to identify the mosaic target and detection defect in the image.3)proposed a model based on the movement distance to trigger the camera to take pictures,according to the movement distance of conveyor,camera field size of view and the target size to calculate the camera trigger time,to achieve the target on the mobile belt not missing and as little as possible to repeat acquire image.4)The transformation matrix between the camera pixel coordinate system and the robot coordinate system is solved by the robot internal and external reference calibration technology,and the target real-time tracking and crawling task on the conveyor belt is realized by combining the position data fed back by the conveyor belt encoder.In this paper,machine vision technology is used to identify and sort the mosaic target.After repeated experiments,the mosaic sorting system has high accuracy,real-time strong and high speed,and has good scalability can be applied to other non-mosaic products with obvious characteristics of the identification and sorting tasks.