红外与可见光图像融合的空间碎片识别方法

针对暗弱空间环境中空间碎片的识别问题,提出了一种光照不均匀环境中的空间碎片识别方法。不同于现有识别方案,该方法从光照不均匀导致空间碎片图像源细节丢失造成识别性能下降的角度出发,首先将空间碎片的红外和可见光图像进行深度融合,并建立空间碎片融合图像数据库,然后基于训练样本采用深度学习技术训练得到空间碎片识别模型。算法分析表明,该图像融合方案具有高度的细节保留能力,识别模型具有在暗弱环境中高精度目标识别能力。最后进行了仿真实验,实验结果表明,该识别方案在姿态变化、图像源亮度变化等干扰条件下都具有较好的鲁棒性。

Visible and Infrared Image Fusion for Space Debris Recognition

On account of space debris recognition in a dim space environment, a space debris recognition method in nonuniform illumination environment is proposed. Different from the existing recognition scheme, the infrared and visible images of the space debris are fused firstly, from the perspective of the loss of details of the space debris image source caused by the nonuniform illumination. Secondly, the fused space debris image database is built, and the deep learning technique is used to obtain the space debris recognition model based on the training samples. The analysis result shows that the superior ability in detail preservation, and the recognition model has high-precision object recognition performance in dim environment. Finally, the experimental results show that the proposed method has good robustness under the interference conditions of attitude change and brightness variation of image source