Color/Near Infrared Guided Thermal Image Super-resolution

Infrared images have many outstanding advantages that visible-light images and near infrared images don't have,such as strong anti-jamming capability,the ability of working at night and in extreme weather conditions,and temperature sensitity.Because of these remarkable characteristics,it is widely used in various fields of military and civilian,military weapons,satellite,medical,security,fire and navigation and other areas.However,infrared images also have their own limitations,such as low resolution,low contrast and details of fuzzy and other shortcomings.The main reason for these defects is that the infrared camera and the visible light camera have much difference in their principle of imaging.The infrared camera is dependent on the thermal radiation to capture the object to the image.High-resolution infrared images can show more image information,so they have great practical value.But the current situation limits its further development.Due to the limitations of industrial technology now,it is difficult to upgrade the resolution of infrared cameras through the hardware like the visible light camera.Therefore,it is an inevitable choice to realize the low-resolution infrared image super-resolution reconstruction with the software algorithm.In the experiment,an innovative method is proposed to support the low-resolution far-infrared images by high-resolution visible or near-infrared images to obtain infrared images with high resolution and good water quality.The idea is based on a strong correlation between visible or near-infrared and far-infrared images.A visible light camera or near-infrared camera a greater advantage than thermal image in price and resolution,so it is practical to use the former to assist the latter for super-resolution.The key idea of the algorithm is to divide the information of the visible light image into two parts.One part of the information is related to the infrared image and the other part is not related to the infrared image.We use the related information to assist in the super-resolution reconstruction of the infrared image.But at the same time,we need to avoid the irrelevant data bringing the error impact to the process of super-resolution reconstruction.In this paper,the whole system is implemented,including image acquisition,camera calibration,image registration and the most critical step—super-resolution processing.