Study On Infrared Image Capturing, Enhancement And Segmentation Processing

In modern infrared technology application, civil infrared inspection system and military infrared detection system have been widely used today, but for the poor quality of infrared instrument and working environment's influence, the result of infrared image is not as good as expected. So it's necessary to process the captured infrared image in order to make the images more suitable for human visual system, and the segmented infrared image to identify the faint objectives.In this paper, active infrared image capturing system has been designed and built, it mainly makes up of infrared light, low-luminance CCD camera, infrared filter, video capture integration block and computer. Performances of three kinds of infrared image capturing system are evaluated, and for the captured images, depicting and judgment of their quality are shown.In the procedure of capturing infrared images, under the influence of infrared detecting instrument and outside lightness, the unwanted luminance drop and halo will appear. So in this paper, through analyzing and comparing the classic algorithms, the Retinex theory is introduced to process infrared signal. The Retinex theory and the wavelet are combined to enhance infrared images; the new algorithm can not only denoise the image and modify luminance of the whole infrared image, but also can improve the contrast and impression of image.A target extraction method based on pre-processing of infrared image is presented to detect certain target; it is implemented through the artificial selection of sub-block which includes the infrared target. Then deciding the threshold based on modified Ostu principle to complete the object's coarse detection, the accurate target location is obtained by edge detection in this object region, morphological algorithms are applied to avoid small shortcomings of segmented infrared objectives finally.