Design And Verification Of High Speed Image Gathering And Processing Platform Based On ZYNQ

The "Industry 4.0"era, whose objective is smart and fully networked manufacturing, is coming. The key of "Industry 4.0"is the advanced factory, which bases on intelligent identification, intelligent detection, intelligent connection, and intelligent control. It is image vision technology containing amount of information that efficiently supports both the intelligent identification and intelligent detection.The traditional image vision technology works as follows. At the frontend, a camera is used to capture pictures, which are subsequently transported to the backend PC through the video-capture card. This approach puts the intelligence at the backend. Referring to the new trend-"put the intelligence into the frontend", this paper introduces a distributed high speed image acquisition and processing platform based on ZYNQ SOPC, which integrates the software programmability of a processor with the hardware programmability of an FPGA. This platform aims at putting processing and analysis of images to the camera end.A prototype of the platform introduced above has been implemented and is described in this paper. On the one hand, the paper introduces the board design of the prototype. With the help of Cadence Oread, the schematic design is done. Moreover, after learning related knowledge of Signal Integrity and Power Integrity, this paper designs the constraints which are used to drive the following layout design. On the other hand, the paper introduces design of the related logic which runs on the Programmable Logic in ZYNQ SOPC and the related software which runs on the Processor Cortex-A9 in ZYNQ SOPC. After studying SerDes and AXI4 protocol, the paper designs a series of IP,High speed image sensor interface IP and High-speed image data path IP, all of which are used to create the logic system based on AXI Interconnect.Finally, the paper illustrates the results obtained through experiment used for prototype verification. The result from fundamental experiment shows that power supply meets the standards, and every component runs well under the proper setting of software. The result from the performance experiment shows that while capturing picture 800 frames per second, the system can still sample the picture of the hands which is turning swiftly and store it without any mistake. With the corresponding algorithm, the rotor speed can be calculated accurately.This platform exhibits more flexibility and higher real-time character. With the corresponding image processing module running on ARM or FPGA, this platform will not only perform image acquisition and processing, but also control related equipment. As a result, it will contribute to the betterment of the modern industry.