Infrared image true colorization method based on local contrast learning

The invention belongs to the technical field of image processing and deep learning, and particularly relates to an infrared image true colorization method based on local contrast learning. Comprising the following steps: obtaining a colored true color image, constructing local contrast learning loss to ensure that local content of a coloring result is consistent with that of an original infrared image, training an infrared coloring network, and obtaining a corresponding true color image. According to the method, under the condition of non-paired data, after the infrared image is mapped to the true color domain, the problems of content change, detail missing and the like exist, and it is guaranteed that the coloring result is true and natural through the thought of attention mechanism-based multi-receptive-field feature fusion and local contrast learning. 本发明属于图像处理技术与深度学习技术领域，具体涉及一种基于局部对比学习的红外图像真彩化方法，包括以下步骤：得到着色后的真彩图像、构建局部对比学习损失以保证着色结果局部内容与原始红外图像一致、训练红外着色网络以及得到相对应的真彩化图像。本发明在非配对数据的情况下，红外图像映射到真彩域后会存在内容改变、细节缺失等问题