Platen State Recognition Method Based on S-component Clustering Segmentation
Download as PDF
DOI: 10.23977/EECTM2020.042
Author(s)
Quan Lu, Chengcheng Pan and Likun Hu
Corresponding Author
Quan Lu
ABSTRACT
The inspection of the state of the platen in substation relay rooms is a tedious task of power operation and maintenance. Machine vision can effectively increase inspection efficiency, but its recognition accuracy requires improvement. Considering this problem, a platen state recognition method based on S-component clustering segmentation is proposed in the present study. Perspective transformation is first conducted to rectify the original platen image, and the image is then converted to HSV space. Next, the effective platen area is segmented accurately through S-component clustering, and morphological filtering and feature analysis are then conducted. Finally, the state information of the platen is judged by analysing the feature of the minimum circumscribed rectangle of the platen. Experiments were carried out on platen images of different cabinets, and the results demonstrate that the proposed method can quickly and accurately recognize the platen state.
KEYWORDS
Platen state recognition, rectify image, HSV color space, cluster segmentation, feature analysis