| Photoelectric imaging technologies have been developing substantially overtime,among which is the thermo-infrared image processing. With rapid improvement of thistechnology, quality evaluation of infrared images are becoming more and more focusedin research for its convenience in monitoring obtaining and processing infrared images,and it means remarkably to functional qualitative evaluation of the infrared imagingsystems.This study, mainly focusing on the quality of infrared images and evaluation ofimaging, first explores detailedly the features of infrared images and main components ofnoise. In bringing into focus two factors, namely Nonuniformity and Bad Pixel that cansubstancially influence the quality of infrared images, the author explains what are theyand why they occur, and resorts to bad pixel detection of infrared images, using a meanvalue of a number of thresholds that appear in multiple images in a time sequence. Andfinally based on previous studies, this paper elaborates on a modular design of simulationsystem for quality evaluation of infrared images, including process design and design ofeach of its modules on their functional and interface levels.A group of experiments are conducted using univariant factor analysis. Byevaluating three groups of images, i.e. the original image, image that is only affected byone single noise and image by noise complex, this study finally indicates a degree ofinfluence on imaging by each kind of noise. And a mean value of thresholds obtained inmultiple images rather than one single image shows better decision-making guidancewhen it comes to detecting bad pixels. |
PDF Full Text Request