为了提高无人机侦察识别能力, 提出基于小波变换方法的无人机载光电与SAR的图像融合技术。经时间配准算法生成图像配准源, 采用SIFT算法提取图像特征点, BBF算法计算生成匹配点集, 依据匹配点集计算图像间透视变换模型完成图像配准, 利用小波变换算法实现配准图像融合。经实验验证以及利用Matlab分析图像灰度直方图和计算信息量, 结果表明: 融合图像保留了光电图像95.7%的细节(熵), 相比于光电图像平均梯度提高了1.52倍, 增强了光电图像目标区对比度, 降低了随机性噪点; 融合图像相比于SAR图像信息量提高了1.44倍。

In order to improve reconnaissance capability of UAVs, an image fusion technology of unmanned aerial vehicle(UAV) borne opto-electronic and SAR based on wavelet transform is proposed. Firstly image registration source is generated by time registration algorithm, and image feature points are extracted by SIFT algorithm. Then BBF algorithm is used to calculate matching point set, and image registration is done according to perspective of matching point set. Finally registration image fusion is achieved by wavelet transform algorithm. Results show that fusion image preserves 95.7% of details (entropy) of opto-electronic image, which is 1.52 times higher than average gradient of opto-electronic image, and opto-electronic image target area contrast is enhanced and random noise is reduced. Compared to SAR image, fusion image information is increased by 1.44 times.