| With the development of machine vision technology, the real-time online and high accuracy inspection based on vision has gradually become the key technology in the process of industrial production. Since the machine vision inspection has such characteristics like high degree of automation, excellent performance in real time, high accuracy in inspection and non-contact, it has become a reliable new technology to replace the defect detection system based on human vision. The paper takes the defect detection based on machine vision as a research and application background, the design and implement of a defect detection system defects the process of powder coating of energy-saving lamps filament in the industrial production.In this paper, we choose the TMS320DM6467 with ARM+DSP architecture as core processor in embedded platform, and use the Monta Vista Linux operating system as system software platform. By using the Davinci software framework’s DVSDK kit and the algorithm program of EMCV transplant, we accomplish the transplant of embedded platform.The main work and achievements of this paper are as follows:1) By introducing the DM6467 and using DVSDK kit, we design the defect detection system based on embedded platform.2) We use the OpenCV to compile the preparation of the algorithm, and then complete embedded algorithm transplant by modifying the source file of EMCV.3) Through analyzing the industrial production environment and designing an optical back-reflexion device which combining with the camera’s API library, we complete a high-quality image acquisition in programming.4) By modifying DVSDK kit’s source code and combining with the EMCV modified algorithm library, we implement the defect detection based on embedded platforms.Finally, through image acquisition and processing of samples in the industrial field, we verify the feasibility, high efficiency and real-time characteristic of the project. |
PDF Full Text Request