激光雷达距离像噪声抑制及方法评价  

Noise suppression and the method evaluation for ladar range images

作　　者：于加其[1] 杨树兴[1] 朱伯立 

机构地区：[1]北京理工大学宇航学院飞行器动力学与控制教育部重点实验室,北京100081 [2]北京特种机电技术研究所,北京100012

出　　处：《光电子．激光》2015年第6期1215-1221,共7页Journal of Optoelectronics·Laser

摘　　要：针对激光成像雷达距离像距离反常噪声的噪声抑制问题,提出了一种改进环圈滤波(IDF)算法并进行了评价。首先,根据激光雷达距离像噪声的构成,建立了一般化的距离像噪声仿真模型,模型中包含了地面、目标和细小结构等成分,可模拟距离反常、失落信息和内部噪声等噪声干扰;其次,提出了一种IDF噪声抑制算法的一般形式,利用仿真模型对中值滤波(MF)、IDF等噪声抑制算法进行了分析,对计算结果从整体、非异常值和细小结构等多个方面,评价了算法的噪声抑制能力和保护目标细节信息能力;最后,采用实测距离像对算法进行了验证。研究结果表明,利用建立的距离像噪声仿真模型,可以有效地评价不同噪声抑制算法的能力;同时,利用IDF可以根据目标特性选取校正系数,在满足算法对保护目标细节信息能力要求的前提下,提高噪声抑制能力。An improved donut filter （IDF） algorithm is proposed and evaluated to restrain the outliers in the ladar range images. First, a generalized noise simulation model of ladar range image is proposed based on the analysis of the composition of the noises in ladar range image, which takes the ground, objects, and detailed parts into account, and can simulate such noise disturbance as outliers, dropouts and internal noise as well. Secondly,an improved donut filter （IDF） algorithm to suppress the noise is proposed. Different noise suppression algorithms, such as median filter and improved donut filter, are analyzed by means of the simulation model and the results are also discussed from the points of the entirety, no outliers and detailed parts. Their capabilities in suppressing noise and protecting the detailed information are evaluated as well. Finally, the improved donut filter algorithm is verified by analyzing the measured data. The results show that the capabilities of different noise suppression algorithms can be evaluated effectively based on the proposed simulation model and by means of mathematical analysis. Meanwhile, the correction coefficient of the improved donut filter can be obtained based on the target features, which can improve the capability of suppressing noise and protect the detailed information.

关 键 词：图像处理 激光雷达 改进环圈滤波(IDF) 质量评价 距离像噪声模型 

分 类 号：TN958.98[电子电信—信号与信息处理]