Random deployment of wireless sensor network (WSN) leads to holes and coverage problem. Thus, the detection and healing of coverage holes have become major issues and obstacles for achieving the satisfactory coverage of the WSNs. For this purpose, comparison is made among tree-based coverage hole detection, Delaunay triangulation and Voronoi algorithm in terms of their performances. It helps to detect the coverage holes in WSNs. The area that is checked for holes may be more than necessary. A method is proposed that will reduce the number of sensors, so that only the desired area is checked for holes. The area that is checked for holes may be more than necessary. A method is proposed that will reduce the number of sensors, so that only the desired area is checked for holes. The proposal is to do this using image recognition technique. The image will be scanned and using the pixel occurrence of black and white colour in the image, and we should be able to identify the desired area and deploy the required number of additional sensors accordingly. We propose to compare the performance of existing tree-based algorithm and triangulation algorithms in this domain, with particle swarm optimization (PSO). © 2021, Springer Nature Singapore Pte Ltd.