Hardware Design Of Embedded Machine Vision System Based On ARM And DSP

To overcome issues such as machine vision systems based on PC or PLC are bulky, and their applications are limited. This paper tries to establish a micro-embedded machine vision systems hardware platform to expand the application of machine vision. System uses the dual-core ARM+DSP architecture, the main processor ARM9 S3C2410 is the control core of the system, it completes image acquisition, storage, transmission, human-computer interaction, interrupt management, alarm, multi-task and so on.Co-processor DSP TMS320CDM642 is the core of data computing and processing, it receives video image information from ARM to complete a variety of image processing algorithms, and translate characteristics of information to the ARM processor, as judgments of decision-making of control, to realize tasks such as stepper motor control, camera position and orientation adjustment, visual tracking.Paper's study focuses on ARM and DSP-based dual-core machine vision hardware platform program and completed the software and hardware design, embedded-Linux operating system's cutting and transplantation, file system configuration, driver develop and debugging, the video image's real-time collection and display and so on, built a highly integrated, real time and multi-task embedded machine vision development platform. this paper Made a deply study on data exchange method between ARM and DSP on embedded machine vision systems and built a highly efficient and stable hardware platform for the application of machine vision systems and algorithms study.It made a useful attempt on building method of machine vision system.