The Algorithm Of Bottle Defect Detection Based On Machine Vision And System Development

It is inevitably that beverage bottle will be dimensional deviation, surface scratches, cracks or other defects in the manufacture. With the continuous improvement of the level of automation, the traditional manual testing methods have been unable to meet the large-scale production. The lag of the detection technology will lead to lower production efficiency. Machine vision has the advantage of efficient, non-contact, which has been widely quoted in various fields of modern industry. In recent years, manual inspection will replaced by machine vision detection technology. It will greatly improve the detection process efficiency and quality and making the entire production efficiency improvement. Therefore, combination of practical application, the study of this paper is the defect detection system of the bottle defects based on machine vision. In order to defect bottle detection achieve of effective and rapid detection.In this paper, first introduces the principle of machine vision, its state of development at home and abroad. Then introduce the algorithms of image processing in the machine vision system. Combined bottle characteristics of defects with the election algorithm in image processing, this paper developed a bottle defect detection algorithm, and then in-depth study of the algorithm. Completing the soft and hardware to establish a bottle defect detection system based on machine vision.This study mainly include:election algorithm for solving the optimization problem of the convergence checking on the various types of objective function to verify the election algorithm has excellent convergence. In the design of software algorithms, according to the bottle defect characteristics, design the key technologies of the detection algorithm. Image segmentation is a very important step in image processing. In this paper, the election algorithm is applied to the Otsu threshold for fast and efficient image segmentation. Combined with the advantages of the election algorithm, this proposed automatic extraction of the ROI of an image. This method can quickly and automatically extracting image ROI. In the part of defect analysis, the method of linear scanning used to scan the image defects. The defects detected by the statistical analysis. Choose the right camera, light source, lens and other hardware to build an experimental platform; and use VS2005as development tools. Preparation of efficient and simple system software which based on OpenCV and defect detection algorithm combined with the design of this article. And this gives a flow chart of the software and detailed steps.By the experimental results and analysis of authentication, the defect detection system of bottle based on machine vision has a good robustness and can be provide technical guidance for the subsequent detection of factory automation.