Centroiding and Connected Component Labeling for Radar Images Using Image Processing Algorithms

Abstract

The main objective of this paper is to identify centroids for the given image using connected component labeling technique. Imaging radars are used to acquire the images of the required target and it produces 2D images. The imaging radar used here is wall penetrating radar which explores the targets behind the wall. The images obtained from this imaging radar are subjected to some processes like thresholding and filtering in order to improve the image characteristics. The region of interest is then defined using a technique called connected component labeling where the image is scanned and the pixels grouped into components based on pixel connectivity. All the pixels in the connected component of an image exhibit same pixel intensity values which are connected together. When all the groups in an image have been identified then each pixel will be labeled with a gray level or color labeling based on the component it was specified. Finally, these individually identified objects are located with centroid, these centroids show the presence of the targets. The main goal is to remove noise/clutters in an image in an efficient manner and to locate individual object in an image by a single point called centroid.