A state recognition method for substation protection plate based on HSI color characteristics

Abstract

The process of identifying substation protection plates faces challenges due to the complex lighting factors that are often inadequately considered. Furthermore， connected domain adhesion during image segmentation tends to be overlooked. Consequently， a recognition method for substation protection plate that leverages HIS （hue， saturation， intensity） color characteristics and employs image processing techniques is introduced. Firstly， the images of the captured protection plate are preprocessed to correct the distortions； then， these images are segmented using HIS color characteristics. Secondly， morphological feature analysis is carried out on the connected domains； the boundary information within these domains is used to eliminate the signboards， and the limit corrosion algorithm is employed to separate the adhesive connected domains. Finally， the state of the protection plate is identified through the fusion of two key features： the boundary feature of the simple minimum bounding rectangle and the deflection angle feature of minimum area bounding rectangle. In a comparative test against the traditional RGB （red， green， blue） color segmentation method and a recognition method based on single aspect ratio， the results demonstrate a substantial improvement in the recognition rate. This method exhibits practical value for engineering applications.

Keywords

• protection plate
• hsi color segmentation
• image processing
• bounding rectangle
• state identification