Vector Mathematical Morphology for Color Image Processing
This paper presents a novel approach to the generalization of the concepts of grayscale morphology to color images. A new vector ordering scheme is proposed based on L*a*b* color space, and color erosion and dilation are defined, and the fundamental color
- PDF / 1,920,863 Bytes
- 7 Pages / 439.37 x 666.142 pts Page_size
- 48 Downloads / 214 Views
Vector Mathematical Morphology for Color Image Processing Bo Tao and Lin Zhang
Abstract This paper presents a novel approach to the generalization of the concepts of grayscale morphology to color images. A new vector ordering scheme is proposed based on L*a*b* color space, and color erosion and dilation are defined, and the fundamental color morphological operations are proposed. The main advantages of the proposed vector ordering are that is compatible to the standard grayscale morphology when it is applied to grayscale images. In addition, it provides improved results in many morphological applications. Experimental results show that the proposed method is useful for color image processing, such as color image filtering. Keywords Vector mathematical morphology ordering Color morphological operators
Color image processing Vector
25.1 Introduction Mathematical morphology is a highly efficient tool in image processing; it has experienced a binary image, gray-scale images and color images of three stages. Binary and gray-scale morphology have been widely used in all areas of image processing, but the research and application of color morphology are not yet ripe [1]. Binary morphology based on set theory above, there are two basic operators: dilation and erosion. The structural elements are defined as a small set according to B. Tao (&) L. Zhang Xi’an Communications Institute, Xi’an 710106, China e-mail: [email protected] L. Zhang e-mail: [email protected]
W. Du (ed.), Informatics and Management Science III, Lecture Notes in Electrical Engineering 206, DOI: 10.1007/978-1-4471-4790-9_25, Ó Springer-Verlag London 2013
187
188
B. Tao and L. Zhang
image the shape and application, and they are used to scan the image. As a result, dilation operation enlarges images while erosion operation contraction images. Dilation and erosion operation maintains the essential characteristics, and removes or suppresses irrelevant content of the images. Based on dilation and erosion operation, we can construct open and close operators. The four basic operators constitute a combination of the entire binary morphological algorithm; it is considered a subset of the infinite. Threshold method or the use of the umbra method can extend binary morphology to grayscale images processing, where the intersection and union operation can be replaced by the maximum and minimum operation in grayscale images processing. This shows that the core idea of theory of mathematical morphology is the ordering relation among the pixels. Because the grayscale images are scalar functions, they are easy to implement, the binary image is a special case of grayscale image so it is easy to promotion from the binary morphological Grayscale morphology [2, 3]. However, color images are vector-valued functions which are not comparable between vectors. Therefore, we can not directly extend gray-scale morphology to color images processing [4]. For the research works that extending grayscale morphology to color images processing, some research results
Data Loading...
