The Application Of Image Analysis In Industrial Production

Empty bottle inspector is mainly used to detect the finish, sidewall and base of empty bottle whether it is in compliance with the requirement for quality of beer production. It is important and widely used in beer filling line, it leds to the development and innovation of domestic empty bottle inspector. This paper designs the detection system based on machine vision, and analyzes in detail the composition of the structure and working process. Then it discusses the key technologies of image acquisition and gives the detection algorithm about the finish, base and sidewall of empty bottle. Especially for overdue bottle detection, it gives two kinds of detection methods. In order to improve the detection rate, it uses the method based on OpenCV and completes the interface of system and the realization of function module with the software development platform. In this paper, the main research work is as follows:First on the basis of relevant research results at home and abroad, combined with the actual demand of empty bottle detector, it gives the overall design scheme of the system and introduces the various components of hardware platform and software platform in detail.Then based on the analysis of the characteristics of images, this paper designs the lighting scheme and appropriate preprocessing algorithm. In order to reduce the processing time, it deals with the place where only reflect the characteristics of the defect. Due to the position of bottle in the image will change of various reasons. It studies the positioning methods respectively.Through the analysis of defect on the finish, it can be divided into three categories, include fracture, damage and abnormity. And for these defects, the paper gives the methods respectively and then comprehensive discriminant finally. For fracture detection it uses the method of radial integral projection. For damage detection, it uses the method of double circular. For abnormity, it identifies through the change of the distance between outer contour of white ring area and center.According to the characteristics of the sidewall of bottle, it proposes to eliminate expired bottle first. So we can save time and reduce security risks. First it proposes a method which determines whether it is expired bottle by the wear ring on the sidewall of bottle. After analyzing the advantages and disadvantages of this method, it gives a recognition method by year of character on the bottle. Then it uses the method of multiple partitions testing for the small defects on the sidewall of the bottle. For each small area, it uses practical detection algorithm based on edge and connected domain. This approach used on small area is also applicable to the small defects on the base of bottle.Finally this paper studies the human machine interface of the system, it is developed by the Visual Studio integrated development environment, embedded image recognition algorithm. It consists of four modules: image acquisition, online image detection, offline image detection, database management. The system realizes the human-computer interaction, and verifies the practicability of the algorithm. In addition it can satisfy the requirement of speed and accuracy.