The Infrared Image Enhancement Processing Algorithm

During World War II, infrared imaging was applied in the military fields for the first time. The subsequent period of time, the technology is mainly used in military fields. But in recent years, with the continuous development of infrared detectors, infrared imaging technology has not only used in military field, but more and more widely used in civilian areas, such as, remote sensing, medical, video surveillance, transportation as well as industrial inspection and other civilian areas.For the infrared imaging system is very vulnerable to interference by the environment and the own deficiencies of system, so the infrared image has low contrast, low signal to noise ratio and high background characteristics, and it brings people the inconvenience to identification and detection on the image. Therefore, in order to get good quality infrared images, to find good ways to improve the infrared image are the hope for people. Generally, the research is doing from two aspects, one is improving the system hardware; other is enhancing the image to the infrared image. Improvements in the hardware are complexity and cost is large. However, improvements in the software are very flexible. So, we usually take the image enhancement processing method.From the digital image processing knowledge, the enhanced process can generally be divided into two categories, one is the frequency domain processing and the other is spatial processing. Although frequency domain method was good at processing in frequency, the computational amount was large. Spatial domain method has a small computational and it possess a good adaptability and real-time quality, but the control effect is not good, and it can not reflect from the perspective of the various parts. For example, histogram equalization was a typical method. In view of the faults and merits of traditional infrared image enhancement algorithm, a new method integrating of time-frequency characteristics is presented. A part of frequency factors are applied to original histogram to form the weighted histogram, and then balancing the weighted histogram with a platform threshold,finally, using another frequency factors to adjust each pixel of image that balanced, then the image was enhanced. Experimental results show that, the infrared image has been enhanced well, the contrast of image has been enhanced, dynamic range of the gray level has been optimized. So this new method has a good effect and with large value for practicality.