Cap Defect Detection System Of Assembly Line Based On TMS320DM6467

Defects based on machine vision inspection system, with its high degree of automation, non-contact, high real-time, the advantages of high accuracy, is an alternative to the artificial eye to the product defect detection of new technology. The subject of machine vision-based inspection system and its application to the background, proposed caps for high-speed beverage production line defect detection system solution.Topics selected TI’s dual-core multimedia processor core processors as TMS320DM6467, CCS compiler algorithm simulation platform as, Monta vista Linux operating system as the system software platform, using Davinci software framework and multi-threading technology, mainly to complete a common defect detection system structures, and beverage bottles defect detection algorithm design, and beverage bottle defect detection system implementation and verification.The main subject of work and innovation are as follows:1) The subject of defect detection for common scenarios to build a common hardware platform and software architecture, and design of specific cap defect detection algorithm, and finally loaded into the bottle defect detection algorithm testing system, applied to high-speed assembly line. This development model to maximize therate of increase code reuse, reduce duplication of development.2) The subject of multi-threaded parallel processing technology, software development, the detection system ofimage acquisition, image processing, defect detection and test results back, and several units can execute concurrently, to improve system efficiency, reducing hardware requirements.3) The subject line for high speed image acquisition to the jitter characteristics of large, defect detection before thepre-designed and targeted key regional location, the results showed good defect detection system can identify thefast movement of the cap.4) Use of the hardware platform based on the subject, made a ping-pong double buffer, embedded C code andassembly line optimization, three algorithms optimization method. Implement an algorithm that is optimized to reduce the time required from437.14Mcycles to9.72Mcycles, a45-fold performance improvement.