摘要
激光传感器是一种结构简单且技术成熟的用于监控和测量等方面的重要传感摄像仪器，但由于存储数据增长过快，因此文中提出对激光传感器系统进行冗余数据挖掘设计。首先，利用聚类数据挖掘方法对重复冗余数据进行数据划分，针对给定数据对象的数据库，根据给定目的对数据数目进行划分，运用目标函数最小化策略将数据划分为k个簇，依据簇中所包含对象的平均值即为质心，利用对象到质心的距离对数据对象进行簇的划分，反复该过程直至准则函数收敛，完成数据挖掘过程；其次，对聚类后的冗余数据采用边长块分割法对激光传感器系统中的重复冗余数据进行删除操作，依据冗余数据删除原理，针对重复冗余数据删除过程中的影响因素进行量化，同时获知重复冗余特征数据，根据分割粒度的变化实现冗余数据挖掘过程。实验证明，运用基于聚类数据挖掘的重复冗余数据删除方法可有效实现激光传感器冗余数据优化过程。
Loser sensor is an important sensing camera used for monitering and measuring with simple structure and mature technology. Due to the rapid growth of stored data, a redundant data mining method in laser sensor system is designed. First, divide redundant data by olustering mining method. For the database with specified objects, divide it according to specified goals. Divide data into k clusters by minimizing target function strategy. Set average value of the including targets in clusters as centroids, carry out cluster division of data according to the distance between targets and centroids, repeat above step till criterion function converging and complete data mining processing. Meanwhile, delete the clustered redundant data by using length - block segmentation method. Quantify the influence factors during deleting and obtain redundant data features. Realize mining process according to the changes of segmentation granularity. Experimental results show that the redundant data deleting method based on clustering data mining can effectively optimize laser sensor redundant data.
作者
平金珍
王茜
师硕
PING Jinzhen1, WANG Qian1, SHI Shuo2(1. Shijiazhuang Information Engineering Vocational College, Shijiazhuang Hebei 050035, China; 2. College of Computer Science and Software, Hebei University of Technology, Tianjin 300401, Chin)
出处
《激光杂志》
北大核心
2018年第5期138-141,共4页
Laser Journal
基金
河北省科技计划项目(16270343)
