The Research And Realization Of The Military Port Objects Classification Platform

As a synthetic subject, remote sensing has gotten rapid development in recent decades. It plays an important role in civil and military applications as it provides a unique perspective from which to observe large regions and acquire abundant useful information in high speed. The development of the remote sensing technology makes us obtain very abundant information of nature, especially with the appearance of high resolution remote sensing image it extends the visual field of the nature. Remote-sensing image processing is dedicated to processing and analyzing digital images acquired by remote sensors. As a primary quantitative means for image analysis, supervised remote-sensing image classification is a crucial issue during the whole procedure of image processing, and has comprehensive applications in many domains. In this paper, the technique of pattern recognition is applied to the recognition field of high resolution remote sensing image of military ports. The algorithm can realize the automatic identification for military port objects, and improve the understanding of remote sensing images.This article mainly completes the following tasks.(1) The high resolution remote sensing image of military ports database is established for the experiment. After thorough studying on the color space of color image and the software of Google Earth, eight US-Japan military ports are selected as the research object. Meanwhile, under the condition of 1 cm scale and 2000-3000 feet high field of view, the military port database is established, in which 150 images are sampled by the software of HyperSnap in the same way.(2) Principal Component Analysis is adopted to target image for feature extraction. In view of great target image information, PCA (principal component analysis, PCA) and 2D PCA (two-dimension principal component analysis,2D PCA) algorithms are used to dimension reduction. Then, it obtains the principal component characteristics which are input parameters for classifier are obtained.(3) K-nearest neighborhood (k-nearest neighborhood, kNN) and support vector machine (support vector machine, SVM) classifier are chosen as classifiers for remote sensing images classification platform of military port objects. Different classifier and different parameters are selected for experiments, and the experimental results are contrasted and analyzed. The experiment results show that SVM classifier has the maximum recognition rate, which is 93.3%, and the recognition rate of kNN classifier is 80.0%. The two classifiers all have good performance.(4) A military port objects classification platform is designed. The SVM model and kNN model are given, and furthermore, the platform interface is simple and reliable.The experimental results provide an effective support for computer auxiliary classification of military port objects.