即時影像物體追蹤之FPGA硬體平台實現

Abstract

在科技日新月異蓬勃發展的現代，智慧型機器人已成為各國研發的重點項目，而在機器人的視覺以及座標定位方面必須使用到影像的擷取及資料的判讀，一般傳統都使用CCD經由usb介面或影像擷取卡先將影像取回PC再做資料分析，而在這方面我們利用低成本的CMOS Sensor及ALTERA10K FPGA搭配Parallel Port 來做影像的擷取分析及傳送，將需要龐大運算量的影像處理先於FPGA分析，再將處理完後較少量的data及影像特徵傳送到Host CPU，降低Host CPU的處理時間。 本論文分為兩部分，第一部份將先介紹整個系統的硬體架構，包括CMOS Sensor周邊電路以及控制介面 之8051單晶片的硬體實現，以及為了實現即時影像擷取所設計之SRAM Control 機制和Parallel Port硬體實現。第二部分則介紹HOST接收端即時影像擷取及Bayer pattern圖素還原法，以及移動物體影像edge detection和Motion detection algorithm硬體實現。

This thesis presents a design that utilizes ALTERA10K FPGA and parallel port to capture images from CMOS sensor. The image process which needs huge computing will be calculated in FPGA. By transmitting only the image characteristics and less data to host CPU, this method can reduce the computing time of host CPU. The thesis is divided into two parts. The first part is the implementation of hardware that include CMOS sensor、microcontroller、SRAM and FPGA . The designs of the SRAM oontroller and hardware implementation of Parallel port are to implement the real-time image capturing. The second part is the hardware implementation of edge detection and Motion detection and the algorithm to recover the original image in the host platform.