摘要
传统激光雷达图像特征分割方法,存在仿射变换方向的随机性,无法有效抵御斑噪声的干扰,得到的图像特征分割结果质量较差。因此,提出基于分形维数的激光雷达图像特征分割方法,通过基于实际差分盒维数的计算方法,获取激光雷达图像的分形维数。采用基于分形维数的激光雷达图像分割方法,通过提取的激光雷达图像分形维数,融合K均值聚类和形态学处理方法,完成激光雷达图像特征的准确分割。实验结果说明,该种方法对激光雷达图像特征进行分割时,具有较高的分割效率和识别率。
Traditional segmentation method, laser radar image features are affine transformation direction of randomness, unable to effectively resist spot noise interference, image characteristics that the poor quality of segmentation result. Therefore, based on laser radar image characteristics of fractal dimension method, based on the actual difference of box dimension difference of box dimension calculation method, obtain the laser radar image fractal dimension. Based on the fractal dimension of the laser radar image segmentation method, through the extraction of laser radar image fractal dimension, fusion, k-means clustering and morphological processing method to complete the laser radar image characteristics of accurate segmentation. Experimental results indicate that the method of laser radar image charac-teristics of dividing, has the high efficiency of segmentation and recognition rate.
出处
《激光杂志》
北大核心
2016年第11期154-157,共4页
Laser Journal
基金
教育局资助项目(CXY201405)
关键词
激光雷达
图像特征
分割
laser radar
image features
segmentation
