A Groove Parameter Processing Method Based on Hierarchical Clustering Analysis

In order to solve the problem of inaccurate calculation of groove parameters caused by cutting defects of blade, offset of cellulose paper position, uneven filling of tow and other factors in the measurement of grooved filter rod, a processing method of groove parameters based on hierarchical clustering analysis is proposed. In this method, the spatial distance similarity between the detected data points is used to cluster the detected groove parameter data layer by layer, then one or several data points with the greatest difference are filtered out. After this method was used the false detection rate is reduced from 4.64% to 0.7%.