【摘要】 在远距离激光测距领域，由于测量目标的反射面特性不同，造成相同距离下墙、线等典型目标的测量结果存在明显差异。如果能够自动区分墙线目标，对不同目标采用不同的数据处理方法，就可以提高远距离目标测距系统的精度和普适性。提出了一种基于机器学习的激光测距墙线目标自动分类方法，采用自适应k-Means聚类算法，根据优化条件，从原始测距数据中筛选出相似性最高的一组有用数据并剔除其余异常数据，根据回波信号特征从统计角度构建特征值空间，然后将XGBoost和逻辑回归算法进行结合实现特征数据的墙线分类模型训练。实测结果表明，本文提出的自动分类方法准确率超过98%,测距精度也明显提高。更多还原

【Abstract】 In the field of long-distance laser ranging, due to the different reflective surfaces of the measurement targets, there are obvious differences in the measurement results of typical targets such as walls and lines at the same distance. Distinguishing wall and line targets automatically, and using different data processing methods for large and small targets respectively can improve the accuracy and precision of laser ranging. An objective classification method for wall and line target based on machine learning algorithm for laser ranging is proposed. The adaptive k-Means clustering algorithm is used to filter out the useful data from the original ranging data, and the feature space are established according to comprehensive characteristics of the echo laser signal. The XGBoost+LR based classification algorithm is combined to realize the training of ranging feature data and finish automatic classification. The actual measurement results verify that the accuracy the proposed method of automatic classification of wall and line objects based on machine learning is more than 98%,and the ranging accuracy is also significantly improved.更多还原