西南师范大学学报（自然科学版）
西南師範大學學報（自然科學版）
서남사범대학학보（자연과학판）
JOURNAL OF SOUTHWEST CHINA NORMAL UNIVERSITY
2013年
12期
95-100
,共6页
图像分割%边缘轮廓识别%梯度算子%模糊形态学%分水岭算法%IFT分割算法
圖像分割%邊緣輪廓識彆%梯度算子%模糊形態學%分水嶺算法%IFT分割算法
도상분할%변연륜곽식별%제도산자%모호형태학%분수령산법%IFT분할산법
image segmentation%edge contour recognition%gradient operator%fuzzy morphology%water-shed algorithm%IFT segmentation algorithm
图像边缘轮廓提取在国计民生、军事等领域有广泛应用，为了克服噪声等影响导致的图像过度分割现象，探讨了一种基于模糊形态学的图像分割改进算法。该方法借助模糊形态学的开、闭运算，首先对原始图像进行平滑处理，然后基于形态学梯度算子进行梯度计算，最后基于改进分水岭算法与 IFT 分割算法的融合，对梯度图像进行分割得到期望的边缘轮廓图像。仿真以图像识别为例，实验显示基于改进的图像分割算法可较好地消除过分割现象，正确识别图像，实现目标与背景的分离。研究结果表明，提出的图像分割算法是合理可行的。
圖像邊緣輪廓提取在國計民生、軍事等領域有廣汎應用，為瞭剋服譟聲等影響導緻的圖像過度分割現象，探討瞭一種基于模糊形態學的圖像分割改進算法。該方法藉助模糊形態學的開、閉運算，首先對原始圖像進行平滑處理，然後基于形態學梯度算子進行梯度計算，最後基于改進分水嶺算法與 IFT 分割算法的融閤，對梯度圖像進行分割得到期望的邊緣輪廓圖像。倣真以圖像識彆為例，實驗顯示基于改進的圖像分割算法可較好地消除過分割現象，正確識彆圖像，實現目標與揹景的分離。研究結果錶明，提齣的圖像分割算法是閤理可行的。
도상변연륜곽제취재국계민생、군사등영역유엄범응용，위료극복조성등영향도치적도상과도분할현상，탐토료일충기우모호형태학적도상분할개진산법。해방법차조모호형태학적개、폐운산，수선대원시도상진행평활처리，연후기우형태학제도산자진행제도계산，최후기우개진분수령산법여 IFT 분할산법적융합，대제도도상진행분할득도기망적변연륜곽도상。방진이도상식별위례，실험현시기우개진적도상분할산법가교호지소제과분할현상，정학식별도상，실현목표여배경적분리。연구결과표명，제출적도상분할산법시합리가행적。
The extraction of image edge contour is widely applied in the fields such as national economy and the people's livelihood,military affairs,etc.In order to overcome the influence of image over segmentation phenomenon resulted by noise and so on,an improved algorithm for image segmentation based on fuzzy morphology has been explored in this paper.By means of open and closed operation of fuzzy morphology, the algorithm first made smooth processing for the original image,and then it took the gradient calculation based on morphological gradient operator,and based on the fusion of improved watershed algorithm and IFT segmentation algorithm,and finally,it made the segmentation for gradient image to obtain the expec-ted edge contour image.The simulation with image recognition as an example has demonstrated that it could effectively eliminate the over segmentation phenomenon by improved image segmentation based algo-rithm,distinguish the image correctly,and realize the separation of target and background.The research result shows that the proposed algorithm of image segmentation improvement is reasonable and feasible.
