Font Size: a A A

Research And Implement Of Imaging Development Platform Based On Xilinx Zynq-7000 Chip

Posted on:2019-01-03Degree:MasterType:Thesis
Country:ChinaCandidate:H J LiuFull Text:PDF
GTID:2428330572961056Subject:Optical Engineering
Abstract/Summary:PDF Full Text Request
The image data acquired from image sensors account for the large proportion of the data needed by the research of artificial intelligence,machine learning,machine vision and so on.However,some traditional solutions which are applied to embedded image system have failed to meet the growing demand.For example,some common microprocessors used as master chip do not have enough ability for signal processing.They have a bottleneck of development of multiprocessor,which is difficult to deal with massive data.The solution of multichip and multiprocessor is high-power with high cost and it takes large space,so it does not conform to the demand of microminiaturization of embedded system.Moreover,the integrated circuit design has limited ability of updating and difference,which means less possibility on rapid extension of functionality and interfaces.Therefore,a brand new embedded image processing platform can provide a development opportunity for developers.It can meet the demand for functionality and reduce the development cycle,which is really meaningful for engineering.The dissertation discusses the schematic diagram and builds the system of image acquisition,processing,transmission and display,based on the Xilinx Zynq-7000-Soc produced by Xilinx Company.In terms of image acquisition,we realize that the serial image data with LVDS level is able to be transferred into parallel data based on the configurable logic of Xilinx Zynq-7000-Soc.After processed by the module of image processing pipeline,the data can be transmitted to PC by the interface of USB,or to PS of Xilinx Zynq-7000-Soc by the interface of AXI.The function of image acquisition can finally be used by the application level on the operation system through V4L2 of Linux system which is transplanted onto the ARM of Xilinx Zynq-7000-Soc.In terms of image processing,the image processing pipeline module,including dead pixel removal,denoising,demosaicing,automatic white balance,automatic exposition,color correction,color space transform,image enhancement,gamma correction and so on,which are realized on PC traditionally,would be working by being transferred to the programmable logic on Xilinx Zynq-7000.In this way,the algorithm hardware can be accelerated and the burden of PC can be eased,which means the processing speed will be greatly increased.In terms of image transmission,with the help of the CYUSB3014 transmission chip of Cypress Company and the success of the transmission interface based on the programmable logic of Xilinx Zynq-7000-Soc,the dissertation makes it happen that the image data can be transferred to PC by the interface of USB3.0 and USB2.0.After the image acquisition,PC can display the image and control the system through the software writing on MFC.In terms of image display,the image data which is obtained from USB interface can be displayed on PC.Also,the image can be displayed on HDMI screen.By transplanting the Linux operation system onto the ARM of Xilinx Zynq-7000-Soc and using the DRM framework of Linux to realize multi-plane superposition on hardware,we can design man-machine interface on HDMI screen by Qt and control the system by mouse.A variety of display methods meet the requirements of the real applications.The platform system discussed in the dissertation can meet the needs of high-performance,low cost,low power,high flexibility and so on.Compared to the traditional image system,the platform system not only can retain the original basic functions,but also can meet the growing new demand with some new features,such as hardware acceleration,extension of interface and flexibility.In a word,the dissertation accomplishes the whole system with a satisfying experiment result.Combining the features of Xilinx Zynq-7000-Soc,the platform system can solve the current system dilemma and meet the related demand,which is of great value of engineering reference and application.
Keywords/Search Tags:Image acquisition, Image processing pipeline, Image display, Mutli-interface, Hardware acceleration
PDF Full Text Request
Related items