研究目的
To develop and implement an intelligent fault diagnosis system for electrical equipment using IRT images by segmenting the high-temperature zones through an optimized FCM algorithm with MALO for centroid optimization.
研究成果
The proposed FCM-MALO approach effectively segments high-temperature zones in IRT images with high accuracy, outperforming existing methods. Future work should focus on automating segmentation and predicting equipment reliability under full load conditions using hybrid optimization techniques.
研究不足
The study relies on specific IRT images and may not generalize to all electrical equipment types; the optimization and segmentation processes could be computationally intensive and may require further validation in real-time applications.
1：Experimental Design and Method Selection:
The study uses IRT images for fault diagnosis. The methodology involves converting RGB images to grayscale, applying histogram equalization for contrast enhancement, segmenting using FCM with centroid optimization via MALO algorithm, and extracting desired regions using Region Props function in MATLAB.
2：Sample Selection and Data Sources:
Real-time thermal images of electrical equipment (e.g., switches, fuses) are captured using a thermal imager, with over 200 images and 600 areas analyzed.
3：List of Experimental Equipment and Materials:
Thermal imaging camera for image acquisition, MATLAB 2015a software for simulation, computer with 4GB RAM and i5 processor.
4：Experimental Procedures and Operational Workflow:
Steps include image acquisition, grayscale conversion, histogram equalization, FCM segmentation with MALO optimization, and region extraction using Region Props.
5：Data Analysis Methods:
Performance is evaluated using statistical measures (sensitivity, specificity, accuracy, PPV, NPV, FPR, FNR, FDR) derived from confusion matrices, with comparisons to existing methods like RG, ORG, and FCM-ALO.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
