Detection Algorithm Of Impurity In The Complex Mold Of Pens Production

As the market economy’s rapid development, pens industry has also had rapid developmentand progress in China. In order to improve productivity and increase competitiveness, manydomestic pens enterprises introduce robotic devices applied to the production of pens, whichcomplete the automatic stripping. But in the process of stripping there are a variety ofphenomena that there will be stripping carry-over or some of the products fall down, all of whichresult in the impurity left on the mold. Currently the detection of the impurity is mostly based onartificial detection method, which leads to non-automation in the production line and can notfully play the role of robot. The operator efficiency is low and the labor intensity is stillrelatively large. Therefore, the need to develop detection system of impurity as an alternative tothe artificial detection in pens production is urgent. This system would reduce the labor intensityand production losses, improve the production efficiency and increase the operator efficiency,economic efficiency and enterprises’ competitiveness.The main subject of this work is to study the image processing algorithm of the detectionsystem, analyze the complex features of the impurity and the diversity of the mold, work over theimpurity detection algorithm, as well as to improve and optimize the system’s detectioncapabilities. The research to improve pens production efficiency has important theoretical andpractical significance.Analysis of the actual needs of impurity detection during the pens production and theimplementation of the program, the paper’s main research and results are divided into thefollowing three aspects:(1) Mold image preprocessing. As the equipment operation, people walking and otherenvironmental factors affect the acquired image quality in the pens production site, it needsimage preprocessing. In this paper, the mold image preprocessing depends on the aspects of theselection of the detection zone, the geometric distortion correction of the image, image contrastenhancement, image noise filtering and image sharpening. The process of image noise filtering uses this modified image denoising based on two-dimensional empirical mode decomposition(BEMD).(2) The extraction of the impurity image feature in mold. For the uncertain of the shape andlocation of impurity left on mold, the subject uses the feature extraction method of combiningthe threshold segmentation and edge detection in order to achieve the precise extraction of theimpurity’s characteristics. And the edge detection uses the improved edge detection algorithmbased on the integration of BEMD and Canny operator.(3) The software realization of impurity detection and system in mold. Considering someexternal factors impacting the impurity detection in the production process of pens, this articlepresents a detection method based on the edge features of the image registration in order toreduce the impact of external factors on impurity detection system’s performance. Finally, usingMATLAB programming language archives impurity detection system.The simulation tests verify the impurity detection method proposed in this paper can solvethe phenomenon of noise interference, lighting changes, image dithering and others, all of whichare generated by the light intensity, people’s walking, equipment vibration and etc in the actualproduction process of pens. The method also reduces the external environmental factors’ effecton the performance of impurity detection system and improves the impurity detection system’spracticality and effectiveness.