低分辨率激光光谱遥感图像弱小目标识别方法设计

Design of weak target recognition method in low resolution laser spectral remote sensing image

为了突破图像分辨率对弱小目标识别工作的限制,优化设计低分辨率激光光谱遥感图像弱小目标识别方法。根据激光光谱吸收原理,生成遥感图像,通过大气校正、平滑去噪等步骤,完成初始图像的预处理。通过图像增强,提升遥感图像的分辨率,分割激光光谱遥感图像区域,通过特征提取与匹配得出弱小目标的识别结果。实验结果表明,优化设计方法的平均目标数量正确识别率为94.4%和目标尺寸识别误差的平均值为0.12 m^(2)。

In order to break through the limitation of image resolution on dim and small target recognition,the method of dim and small target recognition in low resolution laser spectral remote sensing image is optimized.Accord-ing to the principle of laser spectral absorption,the remote sensing image is generated,and the initial image is prepro-cessed through atmospheric correction,smoothing and denoising.Through image enhancement,the resolution of re-mote sensing image is improved,the laser spectral remote sensing image area is segmented,and the recognition result of small and weak targets is obtained through feature extraction and matching.The experimental results show that the average correct recognition rate of object quantity is 94.4%and the average error of object size recognition is 0.12 m^(2).