以FPGA實現圖像邊緣運算偵測障礙物

本論文使用友晶科技DE2-115系統開發板，外接相機及多點觸控螢幕，將相機即時道路影像，透過手滑觸控螢幕，引導前進方向與判斷方向中之障礙物，以全硬體IC電路架構設計，加速影像處理以及影像辨識之速度。相機即時擷取鏡頭影像，並將資料傳送至DE2-115開發板，開發板將影像處理成解析度為640x480圖像，存入SDRAM記憶體。本論文將管理有限容量的開發板記憶體SDRAM與SRAM，進行影像資料灰階化與Sobel邊緣偵測，並即時將處理完之圖像資料顯示於觸控螢幕上，完成與使用者之間的互動，進行圖像邊緣之偵測、方向判定及行進方向中障礙物之判斷。圖像處理辨識出障礙物後，將障礙物的位置標示出來，處理完成後的圖像資料儲存至SRAM記憶體之中，並將其顯示觸控螢幕上。使用者同時也能以手指於多點觸控板上滑出行進方向之軌跡，多點觸控螢幕將手指觸控的位置座標傳至DE2-115開發板，手指觸控點之座標資料會經由手指軌跡電路模組的處理，讓使用者滑出軌跡可以即時顯示至螢幕上；最後以手指軌跡為引導，模擬行車的路徑，判斷引導指示的路徑是否會與障礙物發生碰撞。本論文提出實現的方式，將可運用在駕駛輔助系統上，透過螢幕指引出車輛行進之路線，可運用在速度不快的代步電動車，以及治療手眼不協調的病患。

關鍵字

圖像邊緣偵測； Sobel邊緣偵測； FPGA

並列摘要

In our thesis, we adopted the DE2-115 system development board to build the hardware system of the image processing. We will get the image from the camera, and process the one to display in the multi-touch LCD (MTL) in real time. We also extract the sliding directions on the MTL. In the directions, there is the judgment of the obstacle on the image. In the hardware development and design, we speeded up the image processing and recognition. In our thesis, we will manage the memory capacity of the board on the system. The picture extracted from the camera is transferred into the resolution of 640x480 and stored into the SDRAM. We design the hardware circuit on the FPGA to speed up the image grayscale processing and Sobel edge detection. We use the countable capacity of the memory to store the processing image and control the memory to display on the touch screen. The system will process the camera picture and show the obstacle on the MTL. The user can slide out the directions by fingers on the touch screen, the coordinates of the finger positions will be transmitted into the development board. The system with the developed circuit module of the finger track can trace the coordinates of the fingers. The finger track will become as a reference of the directional path. The system will judge the obstacles on the path. In our implementation, we can build the assistant driver system. We can determine the driving direction on the screen. It can be applied on the low-speed electric vehicles, as well as the treatment of hand-eye uncoordinated patients.

並列關鍵字

FPGA ； Edge Detect ； Sobel Edge Detect

參考文獻

[3] 黃莫凱, “全方位智慧型車輛--前車追撞及車道偏離警示系統,” 碩士論文,國立台灣大學, 2006.

[5] 張盛唐, “以電腦視覺為基礎的速限與偏移車道偵測,” 碩士論文,中原大學, 2017.

[7] 華春和, “影像邊緣偵測之參數化FPGA架構設計,” 碩士論文,國立台灣師範大學, 2003.

[12] I.Yasrix, N.H.Hamid, and V.V.Yap., “Performance Analysis of FPGA Based Sobel Edge Detection Operator,” IEEE International Conference on Electronic Design, 2008.

[14] G. Anusha, T.J. Prasad, and D.S. Narayana ,“Implementation of SOBEL Edge Detection on FPGA,” International Journal of Computer Trends and Technology, vol. 3, no. 3, 2012.

國際替代計量

以FPGA實現圖像邊緣運算偵測障礙物

全文下載

主題瀏覽