Pre-processing study of multi-sensor data

For multi-sensor data with multi-source heterogeneity, data pre-processing is required for data fusion. This paper proposes a data pre-processing method, firstly, the image data is binarized using the designed image binarization algorithm and flame area extraction algorithm, and the flame area information in the fire image is extracted into one-dimensional data, then the five one-dimensional data of ambient temperature, temperature-sensitive cable, CO2, CO and flame area are noise reduced using the designed wavelet noise reduction algorithm, and finally the final pre-processed data is calculated by the selected normalization formula. The results of the study show that the method can improve the effectiveness of multi-sensor data fusion.