We propose a deep learning (DL) image denoising method to solve the problem of poor image quality and serious noise interference in the infrared thermal imaging. We develop a DL model to process the image, fit the noise residual of the image and remove the noise from the original image to get a clear image. The model adopts the convolutional neural network (CNN) architecture and uses the second-order attention mechanism and non-local modules at the regional level to improve the extraction of image features and fit the noise residual. We implement experiments on four different datasets to analyze the performance of the algorithm in different noise environments. The experimental results show that the proposed method can effectively remove the noise in the infrared image, performs well in different types of noise, and retain a lot of image details.

我们提出了一种深度学习（DL）图像去噪方法，以解决红外热成像中图像质量差和噪声干扰严重的问题。我们开发了一个DL模型来处理图像，拟合图像的噪声残差并从原始图像中去除噪声以获得清晰的图像。该模型采用卷积神经网络（CNN）架构，并在区域级别使用二阶注意机制和非局部模块来改善图像特征的提取并拟合噪声残差。我们在四个不同的数据集上进行实验，以分析算法在不同噪声环境下的性能。实验结果表明，该方法能够有效去除红外图像中的噪声，在不同类型的噪声中表现良好，并保留了大量图像细节。