Improve the new edge detection algorithm with increasing the fullcolor image contrast in CIE Lu'v' color space

Since the edge detection of images is one of the most important phases of preprocessing of computer vision, the success of this phase can have a great impact on the next phases, data mining and its desired features. However there has not been a general procedure to be able to process an image in all color spaces, grayscale or black/white spaces and in the meantime, to be able to extract all the details and information of the edges of the image. Moreover, the appropriate procedure for the operation is always selected by considering the usage and type of the image. One of the procedures that can be helpful in improving the edge detection of the image is to enhance the contrast of the image, however, this can cause the elimination of some real edges and also creation of some extra edges in the image. To the best of our knowledge, most of the procedures used for improving the image contrast have not had enough ability to protect the edges of images. In this paper, the features of enhancing image contrast are used to improve the edge detection and by comparing our results with the available literature, not only their weaknesses are eliminated, but also by offering a more efficient algorithm the edges of the image are protected while the contrast of the image (saturation of the pixels) has enhanced. To achieve this goal we make recommendations for changes in optimum filtering and search for efficient pixels which is caused to achieve the ideal results.

Keywords:

Edge detecting , Saturating , CIE Lu'v' color space , White point