Multi-Sensor Fusion Based on Local Activity Measure

In this paper, a novel yet simple, multi-sensor fusion technique realizes in spatial domain is proposed. The core idea is to obtain the local activity features assuming that all the input images are random fields. This local activity feature matrix, which essentially incorporates the local features of the image, is then used to construct the fused image. In the penultimate step of the proposed technique, the homogeneity of the final fused image is verified by the consistency verification process. The performance of the proposed technique is validated subjectively and objectively by extensive experiments on different multi-sensor images. Furthermore, the comparative analysis with state-of-the-art methods confirmed the escalating improvement of the proposed technique.