Research on Fire Detection Theory and Method Based on Millimeter Wave Radiometer

Based on the existing 3mm/8mm radiometer system, this paper conducts an in-depth study of the mechanism of target detection and recognition, various detection methods and methods for extracting target image features at the ignition point of the fire site. The method of converting the optical image of the scene into millimeter wave radiation image is studied, the optical image is segmented based on color clustering, different segmentation areas are specified with different features, and then the conversion relationship between optical image pixels and millimeter radiation image pixels is obtained according to the imaging mechanism of millimeter radiometer and the imaging mechanism of optical image, on this basis, the calculation and simulation of the ignition point detection method of typical forest fire scenes is realized, and the simulated millimeter wave radiation map is obtained. Finally, the edge feature extraction based on differential operator is carried out on the millimeter wave radiation image, and the clear edge of the fire point is obtained. Fusing the simulated millimeter-wave radiation image obtained from visible light with the millimeter-wave radiation image enables the aggregation of target features and their display in a single image. This can significantly improve the target recognition rate and contribute to making scientifically sound and correct decisions.