Hyperspectral Image Enhancement Using Evolutionary Algorithm

Hyper spectral remote sensors collect image data for a large number of narrow, adjacent spectral bands. Every pixel in hyperspectral image involves a continuous spectrum that is used to classify the objects with great detail and precision. This paper presents hyperspectral image enhancement mechanism using an evolutionary algorithm. 2-D Empirical mode decomposition method is used to divide the hyperspectral image belonging to a specific band into finite number of components called intrinsic mode functions. The last component is called a residue. Each intrinsic mode function is multiplied by a specific weight and summation of these weighted intrinsic mode functions gives the enhanced image. The weight related to each intrinsic mode function is automatically determined using genetic algorithm. The information entropy is used as objective function in the genetic algorithm. This image enhancement increases the classification accuracy of hyperspectral images with an unsupervised classification algorithm.