基于机器视觉的文件扫描机器人缺失数据填补

Missing Data Filling of Document Scanning Robot Based on Machine Vision

由于扫描环境噪声过高,导致机器人扫描文件图像内部分信息无法识别读取,为此,提出一种基于机器视觉的文件扫描机器人缺失数据填补方法。采集机器人的历史扫描缺失数据,划分不完备数据集及容差属性数据集,利用对数非自然函数识别数据集内的缺失数据,自适应函数值越大的数据,估计缺失的概率越高。采用机器视觉技术结合扫描机器人激光映射特点,根据扫描点的时间序列关系,提取扫描区域内的中心值。根据中心值判定邻近范围内的数据是否存在噪声影响,采用收敛模型实现填补。仿真实验证明,所提方法填补效果极佳、缺失数据识别率较高。

Because the scanning environment noise is too high,some information in the file image scanned by the robot can not be recognized and read.Therefore,a missing data filling method of file scanning robot based on machine vision is proposed.It collects the historical scanning missing data of the robot,divides the incomplete data set and the tolerance attribute data set,and uses the logarithmic unnatural function to identify the missing data in the data set.The larger the value of the adaptive function is,the higher the probability of missing is estimated.Using machine vision technology and laser mapping characteristics of scanning robot,the center value in the scanning area is extracted according to the time series relationship of scanning points.According to the central value,it is determined whether there is noise influence on the data in the adjacent range,and the convergence model is used to fill.Simulation results show that the proposed method has excellent filling effect and high recognition rate of missing data.