Image processing is the process of analyzing and manipulating the digitized image to visualize the hidden objects and improve its quality. Occlusion is an obstacle or insignificant object that disturbs the matching or recognizing process. A stereovision system captures the images from parallel cameras as inputs. However, occlusion removal is a challenging issue in many computer vision algorithms and the image is not viewed clearly. Therefore, the image quality is still to be improved by removing occlusion in an image. In order to improve the image quality, Prewitt Texture Synthesis based Image Occlusion Removal (PTS-IOR) mechanism is introduced. PTS-IOR mechanism is used to remove large objects from image and replace visually reasonable backgrounds with minimal occlusion detection time. Initially, the input image is pre-processed using Wiener filtering technique to remove the noise artifacts in images. After that, an occluded object in an image is detected using best-first search algorithm. Then Prewitt edge detector is used to find edge of occluded objects in an image. Finally, region-filling algorithm is used for texture synthesis process to fill the edge detected occluded regions for improving the quality of images. A texture is synthesized through each pixel for misplaced region in texture synthesis. Experimental results show that the proposed PTS-IOR mechanism significantly improves the image quality in terms of Peak Signal to Noise Ratio (PSNR), true positive rate and computational complexity with respect to image size and number of images.
Volume 11 | 07-Special Issue
Pages: 1052-1061