作者: Yangxing Liu , Satoshi Goto
DOI: 10.1016/J.COMPELECENG.2014.04.019
关键词: Pixel 、 Mathematics 、 Time complexity 、 Artificial intelligence 、 Computer vision 、 Color image 、 Image gradient 、 Edge detection 、 Boundary (topology) 、 Enhanced Data Rates for GSM Evolution 、 Object detection
摘要: A fully automatic approach of color image edge detection is proposed.The algorithm integrated with a novel circle efficiently.Three parameters are estimated attractively low time complexity and memory requirement.Objects incomplete or concentric boundary in real images without any prior knowledge can be located. Object critical step many recognition systems. In this paper, we discuss the problem circular shape object still images. An isotropic detector merged spatial information region based analysis employed to extract obtain accurate gradient pixels, which assures high accuracy subsequent detection. Then three efficiently only one 2-dimensional accumulator array 1-dimensional array, greatly reduces storage requirements our approach. Experimental results show that method robust locating objects complete knowledge.