The Research of Recognition on Oceanic Internal Waves Based on Gray Gradient Co-Occurrence Matrix and BP Neural Network

The gray gradient co-occurrence matrix and BP neural network were presented about the issue of the recognition of oceanic internal waves in a MODIS remote sensing image. First, the gray gradient co-occurrence matrix was used to extract the texture features of internal waves. Second, the appropriate eigenvalues extracted were selected as components of the input vector of the BP neural network.. Third, the hidden layer and output layer of the BP neural network were obtained through experiments. At last, the peaks of internal waves were recognized. The recognition rate of internal waves in those subgraphs with internal waves was approximately 90.6% and the one in those subgraphs without internal waves was approximately 83%. The result indicated that the gray gradient co-occurrence matrix and BP neural network adopted to recognize oceanic internal waves in a MODIS remote sensing image were feasible.