Reseach On Key Technology Of Optoelectronic Detection Sysytem For The Bottom Of Bullet Surface

The production of bullet has become automated with the rapid development of automated production equipment. But the surface of bullet is detected manually in China, such as by hand or visual detection, it is laborious and time-consuming, and the stability of the result is poor. Automated inspection system of bullet surface has great prospects and significance for military enterprises and public security departments. Therefore, we need to develop real-time and photoelectric detection system for the bottom of bullet surface as soon as possible.We researched on some key technical related to the system.1) Structural design of the systemThe framework and working process of bullet surface detection system has been built, structure of the photoelectric detection system for the bottom of bullet surface has also been analyzed and designed, some relative devices are selected and some modules are designed in this dissertation.2) Image preprocessingThe type of noise presented in the image of bullet bottom is analyzed, various algorithms for image filtering are compared and finally the code of de-noising algorithm is developed for the image of bullet bottom. According to the characteristics of low contrast for the image of the bullet bottom, methods for image enhancement is researched. Algorithm of image segmenting is researched, a local dynamic threshold method is improved for the image of bullet bottom, and threshold for the segmenting method is automatically selected.3) Defect identification and classificationThe method to detect circle of bullet bottom is researched in the size detection term, the Sobel edge detection method and the improved method of Hough transform is used, radius and center of the circle is measured, and the size defect is identified. In the stage of defect identification and classification, method to extract characteristics of defects and the character recognition method are studied, and the defect classification criteria is developed based on defect feature of the bullet bottom.4) The effect of the system is testThe effect of the whole system is experimented, a large number of bullets are detected by the system, false detection rate of the experiment is less than 3percent, miss rate is zero, the detection time is less than 0.7 seconds per bullet, customer requirements are meet, and the detection speed can match the production rate.Through the study of these key technologies in the dissertation, a series of real-time detection scheme for complex image is proposed, these methods have reusability and extensibility, have great value and instructive for surface inspection of industrial components, such as detection system for glass, bearing surface defect detection and other surface defects detecting system can draw or reuse some related programs in the system.