The Study Of High Image Processing On Vision Sensor

With the rapid development of modern automation manufacture, it is necessary to develop new inspection ways to ensure the quality of the products. Many traditional measuring methods are not suitable for modern industrial product any more. However, visual inspection based on computer vision meet the demands. Visual inspection equipments can offer non-contact, high-precision, automatic, and 100-percent-on-line measurements, so now it has been used in monitoring and measuring in industrial field. The speed of inspection network can be increased by simplifying its structure and improving the speed of image processing. In this paper, the idea of smart vision sensor is presented, which is based on the COMS camera chips and FPGA technology. It aims to improve the inspecting velocity of the vision inspection system and make it to fulfill 100% real time on-line inspection.This paper focuses on the three aspects below and achieves some results.The study on the hardware architecture design of the Smart vision sensorBy analyzing the advantages and disadvantages of traditional vision inspection system, the system based on CMOS camera and FPGA is brought forward. And the schematic design has been finished.The study on image collection using CMOS cameraThe system uses CMOS camera as its eyes. This paper discusses how to use the camera and gives a VHDL model to initialize the camera..The study on low-level image processing in FPGAOn the basis of analyzing the main ways of image processing and the resource of FPGA, some real-time low-level hardware image processing algorithm such as Median Filter in VHDL is present. It also discussed the usage of the CPU which is embedded in FPGA.