Infrared Detection Method for Power Equipment Based on Pixel Dimension Reduction

With the constant expansion of the power grid, the requirement for the effectiveness of power equipment state monitoring is increasing. Infrared temperature measurement has been widely used as the state monitoring method, but the detection accuracy is much affected by data noise, easily leading to misjudgments and omissions. To solve the problem, this paper proposed a new infrared detection method based on pixel dimension reduction for power equipment. The pixel data of the infrared temperature measurement image is dimensional reduced to eliminate the noise data, and ensure the consistency of infrared detection data under the same operating state, improving the accuracy of the temperature measurement result recognition. The experimental results show that the pixel dimension reduction algorithm proposed in this paper can improve the consistency of infrared temperature measurement data of power equipment in the same operating state from 73.54% to 88.67%, thereby improving the accuracy of defect recognition from 78.21% to 95.47%, which can provide timely and accurate information on equipment fault types and severity for power equipment operation and maintenance personnel.