Multisensor Data Fusion for Life Detection in Cluttered Environments

A multisensor data fusion algorithm is proposed in this study for life detection in cluttered environments. The method consists of four steps: data acquisition, data preprocessing, data fusion, and result sort. First, four kinds of data (RGB image, RGB video, infrared image, and radar wave) are acquired from three different sensors (RGB camera, infrared camera, and life detection radar). Then, four specified algorithms are used on each data for data preprocessing. Third, an ordered weighted average (OWA)-based data fusion is used to fuse the four single results. The result is a sort of four possible results ( $\textit {Alive}$ , $\textit {Dead}$ , $\textit {Animal}$ , and $\textit {None}$ ). A series of cluttered environments are designed to verify the effectiveness of the proposed method. The results show that the data fusion method can effectively improve the success rate of life detection in cluttered environments. Specifically, the average accuracies of 97.8%, 96.7%, and 89.7% under different distances, light intensities, and shelter, respectively, are obtained.