Raw infrared image enhancement via an inverted framework based on infrared basic prior

Existing raw infrared image enhancement methods can effectively compress image and improve contrast. However, there still exist limitations. First, enhanced images are often over-exposed. Second, a high contrast but low noise enhanced image is difficult to be obtained due to the fact that the noise level increases with contrast. Third, the targets are not sufficiently salient in enhanced image. In this paper, we design an inverted enhancement framework to address the three limitations simultaneously. Specifically, we analyse the widely recognized features of raw infrared image and call them infrared image basic prior. That is, infrared detectors are often used to detect targets under special conditions and our attention mechanism is to focus on high radiation objects, but there are few targets in the scene. Then we modify the traditional image enhancement framework into an inverted framework based on infrared image basic prior and design an inverted nonlinear gray mapping curve, which avoids over-exposure and noise over-amplification. Furthermore, result is further improved by using layer decomposition model and gamma correction. Enhanced result yields the targets of our main interest. Finally, the extended applications are performed and show ability to stimulate the effectiveness of algorithms of related fields. Experiments show that our approach yields better results than state-of-the-art methods. A video of results is provided at https://github.com/wangyuro/Datashare1.