Real time runway detection in satellite images using multi-channel PCNN

This paper proposes a pulse coupled neural network with multi-channel (MPCNN) linking and feeding fields for multispectral image processing. Different from the conventional PCNN, pulse based RBF units are introduced into the model neurons of PCNN to determine the fast links among neurons with respect to their spectral feature vectors and spatial proximity. This MPCNN can be implemented in parallel on a FPGA chip to perform real-time image segmentation and edge detection. Based on the output of the neural circuits, modified Hough Transform and landmark feature extraction algorithm is designed to perform airport runway detection in satellite images. Experimental results show that the proposed parallel MPCNN circuits drastically improve the processing speed over the popular seeded region growing (SRG) algorithm for segmentation of RGB satellite images. Furthermore the detection accuracy of the proposed scheme remains competitive.