关键词: 张量分解, 图像识别, 特征提取, 平行因子分析

Abstract: The correct recognition of flame combustion state is a prerequisite for maintaining stable combustion. The existing methods have low intellectualization level, strong subjectivity and difficulty in the recognition of combustion state. Therefore, in order to improve the accuracy of recognition, a flame image recognition method based on tensor decomposition is proposed in this paper. The distribution characteristics of the flame pixels in the RGB color space model are used to construct a three-dimensional tensor for each channel image, and on this basis, the parallel factor analysis method is used to reduce the dimension and extract the relevant features of the flame image. The analysis results show that compared with the traditional single flame image region recognition and feature fusion method, the flame image tensor is decomposed directly, which not only can find the relationship between the flame image sequence, but also can retain more information to improve the accuracy of image recognition. Compared with the theoretical value, the average error after calculation is within 0.5%, and the accuracy is more than 99.5%. It provides an effective method for continuous flame image recognition.