Color Image Segmentation Optimization: Threshold Edge Detection with Harmonic and Wiener Filter Enhancements

Abstract

Digital photos can be segmented to find objects, borders, and other relevant information. Segmentation can be done in a variety of ways, including Watershed Segmentation, Region-Based Segmentation, Edge-Based Segmentation, Threshold-Based Segmentation, and Cluster-Based Segmentation. These methods produce a segmented image, which is a compilation of every pixel in the image. Pixels represent the color, texture, and other elements of an image. An image is divided into separate objects or areas. Color Image Segmentation uses threshold-based edge detection, and it’s improved through harmonic and Wiener filters to reduce noise. This method simplifies and changes the representation of an image into something (line, curve drawing that highlights the intensity change, which is more important and easier to explore. In this concept, convert a color image into a gray-scale image and apply different filters (Robert, Prewitt, Sobel, log, and Canny) with edge detection techniques. This applied edge detection technique is also useful with a filter that gives acute and exact results. This research computes the Threshold 0.67 values for color images to check for better performance. For this purpose, use MATLAB. Our technique will help to improve edge detection by the combination of other types of filters, namely Hormonic and Weiner, to eliminate the noise from the image.