Affiliations: [a] College of Computer, Pingdingshan University, Pingdingshan, China
| [b] Department of Economics, University of Southern California, Los Angeles, CA, USA
Correspondence:
[*]
Corresponding author. Gaiyun Zhou, College of Computer, Pingdingshan University, Pingdingshan, 467000, China. E-mail: [email protected].
Abstract: Currently, the method was not applicable to the requirement of feature extraction of different types of grayscale images, resulting in the feature extraction results with low accuracy, long time consumption, low clarity and poor flexibility. In this article, a method of extracting feature of gray image based on fuzzy clustering algorithm was proposed. The grayscale, the median filtering, the edge detection and mathematical morphology processing were carried out for the color image of CCD camera collected by acquisition card. Then, sample feature object of target object gray level image and object of target feature were obtained. The similarity between sample feature object of target object gray level image and object of target feature was obtained through calculation. Moreover, the feature conforming to the set threshold was selected. Meanwhile, the grayscale image feature extraction results with different requirements were obtained through adjusting gray level image matrix and similarity parameters. From comparison and analysis of experimental result, we can see that the correctness, effectiveness and flexibility of proposed method are proved for different types of gray level image feature extraction. The extraction result has high definition and short running time.
