AN IMPROVED BACKPROPAGATION NEURAL-NETWORK FOR DETECTION OF ROAD-LIKEFEATURES IN SATELLITE IMAGERY

This paper presents an application of backpropagation neural network f or the detection of linear structures in remote-sensing images. The pu rpose of the approach is two-fold. First, to exploit the advantages of a neural network classifier over the tranditional ones. Second, to av oid the strategic phases of enhancement and thresholding. Once the net work is learnt, the classification scheme is real-time. Two critical i ssues in the present approach an the selection of the network architec ture and the rate of convergence of learning. Solutions to these two p roblems are proposed. Experimental results on IRS and SPOT images are presented. Satisfactory classification results have been obtained usin g the network.