摘 要:數(shù)字圖像邊緣檢測的設(shè)計大多是基于Matlab軟件設(shè)計的,這與工程化還有一定的距離。針對這一不足,給出了數(shù)字圖像邊緣檢測Sobel算子以及程序設(shè)計流程,并在CCS v3.1的軟件仿真環(huán)境下進行仿真實驗。通過3個實例的分析,驗證了基于CCS的數(shù)字圖像邊緣檢測設(shè)計的可行性。結(jié)果表明,使用Sobel算子實現(xiàn)了圖像邊緣檢測,這為圖像處理提供了一種硬件實現(xiàn)的方法。
關(guān)鍵詞:CCS; 邊緣檢測; Sobel算子; 數(shù)字圖像
中圖分類號:TN919.8 文獻標識碼:A
文章編號:1004-373X(2010)14-0056-02
Design of Digital Image Edge Detection Based on CCS
YANG Hong
(School of Electronic Engineering, Xi’an University of Post and Telecommunications, Xi’an 710061, China)
Abstract: The design of edge detection of digital image is mostly based on Matlab software design, so this is still a certain distance from the engineering. In response to this lack, the digital image edge detection algorithm based on Sobel operator and program design process is proposed, and the simulation experiment is done in the CCS v3.1 software simulation environment. The feasibility of this design of digital image edge detection based on CCS by analyzing three application examples is proved. The results show that the image edge detection is achieved by Sobel operator, a hardware implementation approach is provided for the image processing.
Keywords: CCS; edge detection; Sobel operator; digital image
基金項目:國家自然科學(xué)基金資助項目(40974078);陜西省自然科學(xué)基金資助項目(2007D15)
0 引 言
邊緣是圖像最基本的特征,圖像的輪廓、細節(jié)基本都存在于圖像的邊緣部分,它是圖像的最基本特征,在圖像分析中借助它能大大地減少所要處理的信息,又保留了圖像中物體的形狀信息[1-3]。因此,邊界檢測在圖像處理、模式識別和機器視覺等領(lǐng)域中有很重要的作用,它是底層視覺處理中最重要的環(huán)節(jié)之一。邊緣能勾畫出目標物體,使觀察者一目了然,邊緣蘊含豐富的內(nèi)在信息(方向、階躍性質(zhì)和形狀等)。從本質(zhì)上來說,圖像邊緣是圖像局部特征不連續(xù)性(灰度突變、顏色突變和紋理結(jié)構(gòu)突變等)的反映,它標志著一個區(qū)域的終結(jié)和另一個區(qū)域的開始。
兩個具有不同灰度值的相鄰區(qū)域之間總存在邊緣,邊緣是灰度值不連續(xù)的結(jié)果,這種不連續(xù)性通常可以利用求導(dǎo)數(shù)的方法方便地檢測到,一般常用一階導(dǎo)數(shù)和二階導(dǎo)數(shù)來檢測邊緣。邊緣檢測的基本思想是首先利用邊緣增強算子,突出圖像中的局部邊緣,然后定義像素的“邊緣強度”,通過設(shè)置門限的方法提取邊緣點集[4-5]。……