论文部分内容阅读
通过对激光雷达成像的光学图像分析,实现地面建筑物的识别和分类,面临抗噪性能差,特征可变性强等问题。提出一种基于自适应局部噪声滤波和仿射不变矩特征的建筑物激光雷达光学成像分类识别算法,首先构建激光雷达光学成像的采集模型,采用横向扫描、块扫描方法进行图像的特征点扫描,根据自适应局部噪声滤波算法进行光学图像的降噪提纯处理,提取滤波输出图像的仿射不变矩特征,采用BP神经网络分类器进行特征分类,实现建筑物激光雷达光学成像的分类识别。仿真结果表明,采用该方法进行光学图像的降噪滤波以及特征提取的准确性较好,输出图像的峰值信噪比较高,对建筑物分类的精度较高,计算开销小于传统方法,误分率等各项指标性能优越。
Through the optical image analysis of laser radar imaging, the identification and classification of ground buildings are realized, and the problems of poor anti-noise performance and strong feature variability are faced. A classification and recognition algorithm of building lidar based on adaptive local noise filtering and affine moment invariant features is proposed. First, an acquisition model of lidar optical imaging is constructed, and the feature points of the image are scanned by horizontal scanning and block scanning methods The adaptive local noise filtering algorithm is used to denoise and degrade the optical image. The affine moment invariant features of the filtered output image are extracted. The BP neural network classifier is used to classify the features, so as to realize the classification and recognition of the building Lidar optical imaging. The simulation results show that the proposed method has better accuracy of noise reduction filtering and feature extraction, the peak signal-to-noise ratio of output image is higher, the accuracy of classification of buildings is higher, the computational cost is less than the traditional method, Rate and other indicators of superior performance.