Adaptive fusion algorithm of heterogeneous sensor networks under different illumination conditions

• We present an adaptive image fusion method used in heterogeneous sensor networks.
• A fuzzy system is applied to classify the imaging illumination conditions.
• An adaptive fusion scheme based on the classification results is proposed.
• The proposed algorithm has better applicability to different illumination conditions.

This paper presents an adaptive fusion method used in heterogeneous sensor networks, which consist of infrared sensors and visible sensors. A fuzzy system is applied to classify the imaging illumination conditions. We select three image quality evaluation parameters as the inputs of the fuzzy system, and the system’s output will determine the fusion weights of the two kinds of images. It means that the classification results are used to guide the fusion process. An adaptive fusion scheme according to blended basis functions is used to fuse the images. Our adaptive fusion rule is compared with the weighted average fusion scheme in the experiments under both high illumination conditions as well as low illumination conditions. The simulation results on the Matlab platform verify that our adaptive fusion algorithm has better applicability to different illumination environments when compared with the fixed weights fusion algorithm.