Gray-level Quantitative Recognition Method Research For Linear Texture

Digital image process is always a major field of computer applications and it's use in production and daily life has promoted much more effect. Besides, the fingerprint image procession, which is based on the technology of gray-level digital image recognition, has become one of the most important binarization methods, and has been of great interests to researchers.The binarization method is considered a key issue in digital image process, which directly affects the analysis of the outcome data. This paper brings forward the general methods of binarization in gray-level image, including global methods and local dynamic methods, with analysis of their advantages and shortcomings. A new binarization method of fingerprint images is then proposed, which is based on the orientation and the dynamic threshold, and it has the excellent capability of noise resistance. This method makes full use of the orientation and the characteristics of grayscale change and obtains the binarization image from the original image by not losing the basic characteristics. Experimental results show better processing qualities.In addition, several elementary means of the gray-level image process are presented in this paper, such as Histogram, GST (Gray Scale Transformation), Threshold transformation, and Equalization. Finally we will introduce the software of "gray-level image process" which is programmed by Visual C++ 6.0.