A Novel Hybrid Edge Detection Technique: Abc-Fa (original) (raw)
International Journal for Research in Applied Science & Engineering Technology, 2021
Image edge detection mainly involves in identifying points in an image at which the image brightness changes sharply or more formally has discontinuities. The points at which image brightness changes sharply are typically organized into a set of curved line segments termed edges. Edges give important structural information about the images which would play a vital role in various fields of image processing and computer vision. In order to optimize the process of image edge detection and produce better edges for images, we have implemented edge detection using two swarm intelligence techniques. In this chapter we will study in detail how ant colony optimisation and particle swarm optimization techniques were utilised to identify edges of an image. I. INTRODUCTION The concept of edge detection mainly involves in identifying local variation in brightness levels in images. Image edge detection is mainly employed for image processing and computer vision. Computer vision helps in extracting useful information from the detected objects such as images, videos. In last two decades numerous techniques and variations have been developed. Despite them being carried out, main skeleton of these techniques remains similar. With varying application domains, most of these techniques have been modified to fit into various applications. The changes undergo mostly in perspective of encoding scheme, parameter tuning and search strategy. The edge detection task is to find the boundaries of image regions based on properties such as intensity and texture [3]. It is a critical low-level procedure of image processing because edges carry a lot of information. The edge detection process generally includes five steps: 1) Filtering: Filtering out noise from the image and improving the performance of the edge detector. 2) Enhancement: Emphasising pixels which have important change in local intensity. 3) Detection: Identifying the edges and thresholding. 4) Link: Linking the broken edges (such as hysteresis thresholding techniques). 5) Localisation: Locating the edge accurately and estimating the edge orientation (edge and orientation map). In the modern age computer vision concepts are useful for detecting checks through images and in automated driving cars. Edge detection helps in identifying the important structural object of the image, which thereby is used for identifying important information from image such as number, size, and relative location of objects in an image. If we compare black and white grayscale images the variation in colour for our eyes is very sudden but at pixel level, it is not very rapid. The edge detected is a constant change of pixel values when we travel from one bright level to another bright level in the space of pixels of an image. Usually, the technique for identifying variations in the pixel values to identify edge variation is calculated for each pixel value in image space which is not an optimistic solution for identifying an image. So, to optimize the solution for this scenario in this paper we will further discuss about implementing swarm intelligence for detecting edges. Swarm intelligence mainly involves the study of collective and individual behaviour of biological systems which help in improving efficiency for solving complex problems such as identifying shortest path for food source or identifying their enemies in nature. Even though all the insects are unsophisticated and unorganised they make wonders by adapting to their problems in environment and solving their daily life scenarios.