Cis Acquisition System Based On Gigabit Ethernet Design

As an emerging industry of photovoltaic technology, the machine vision has been widely used in microelectronics, automotive, medical, packaging, printing, electronics, scientific research, military and many other industries. The image acquisition is the basis of machine vision, and the high-quality image quality is the basis for subsequent analysis and processing. The most important thing to the industrial applications is upload the target object image as soon as possible.This dissertation introduce a system of image acquisition, which is based on a contact image sensor and gigabit ethernet. The system use three contact image sensors to capture the target object image, and use a gigabit ethernet to transfers the data to host computer. Acquisition module and transmission module are separate logically, so it is east to system upgrades in the future.The system is divided into two parts, one is hardware design, and another is embedded Linux transplantation.Hardware design consists of three modules. The first module is image acquisition module, which is based on contact image sensor, AD9822 and Xilinx FPGA. The way of using three contact image sensors parallel can increase the width of the scanning greatly. AD9822 is used to translate the analog voltage signal to digital signal. Xilinx FPGA control the data flow precise as the control central, and the image data is cached in the on-chip memory of FPGA. The second module is the shared memory between FPGA and ARM. The on-chip memory of FPGA is mounted on storage controller of ARM as a Bank, and the image data is transfered from the on-chip memory of FPGA to the external memory of ARM using the DMA. The third module is network transmission, which is include the ARM S3C2440A as a microprocessor, AX88180 as a MAC, and 88E1111 as a PHY chip.Embedded Linux transplantation consists of two modules, the first is the Linux kernel transplantation, and the second is the development of network driver.