Research And Implementation Of Ship Target Detection Algorithm In GF-3 SAR Imaging

The detection of ship target in synthetic aperture radar image is one of the needs of radar image interpretation application.High score three satellite is a phased array satellite which is independently developed and successfully launched in China.It has the features of multimode and multi-function,wide imaging amplitude,long running life and high image quality.The microwave imaging system can realize all weather and all day monitoring and remote sensing of the earth and ocean.Therefore,it is of great theoretical significance and practical value to study the fast detection algorithm of ship targets based on the radar image data provided by the high score three satellite.This paper mainly studies the fast detection algorithm of ship target and develops a software system for ship target detection used in high score three satellite image.The main work and research results of this paper are as follows:In order to solve the problem of excessive false alarm in the complex background of land and sea boundary,the second chapter studies the sea land segmentation algorithm based on the maximum inter class variance algorithm.This paper mainly introduces the image preprocessing method,realizing the maximum inter class variance algorithm,the morphological operation and the coastline acquisition process.At last,the experiment is carried out with the high score three satellite image.The experimental results show that the algorithm can accurately obtain the external contour curve at the boundary between sea and land,so that the land area can be eliminated effectively.The third chapter is the first level coarse detection algorithm in the whole ship detection process.The traditional detection algorithm is the process of detection by sliding window.This process is very time-consuming.The paper provides a frequency domain algorithm to implement the whole detection process,which can provide the time of order of order.First,the principle of the traditional constant false alarm rate detection algorithm model and the fast CFAR detection algorithm model is introduced,and the comparison and analysis of the core process calculation in the traditional algorithm and the fast detection algorithm is given.In the actual high score three satellite image scene,a fast detection method based on Rayleigh distribution is proposed in the case of the sea clutter background distribution model obeys Rayleigh distribution.Considering the Gauss distribution of sea clutter background distribution model,a fast ship detection method based on Gauss distribution is proposed in this paper.Experimental results show that the proposed fast coarse detection algorithm can greatly reduce detection time when the detection effect is the same as traditional methods.The fourth chapter is about the second level precise detection algorithm in the whole ship detection process.The CFAR detection algorithm based on the background sea clutter obeys the different distribution is based on the amplitude difference,and in the process of detection,there will be a lot of difficulties in the false alarm target.The algorithm based on the support vector machine classifier is proposed to implement the hierarchical histogram of the direction of the ship target.This paper mainly introduces the working principle of feature and the whole process of feature acquisition.Due to the overlap phenomenon in the process of feature acquisition,interpolation is introduced to correct feature vectors.Secondly,the process of the precision detection algorithm is introduced,including the introduction of the working principle of the algorithm,the design of the SVM classifier and the cross validation training method to verify the effect of the classifier.Finally,the algorithm is verified by using five meters resolution and one meters resolution image.The experimental results show that by classifying the candidate targets based on the fast detection algorithm based on Gauss distribution to achieve multilevel detection of ship targets,the performance of the whole detection system can be greatly improved.Fourth,the core work of the fifth chapter is to use the fast detection algorithm based on Gauss distribution and the classifier based ship target precision detection algorithm as the core algorithm of ship detection.A set of interactive software for processing high score three satellite image is developed.The software can accurately detect ship targets in high score three image and extract ship position and related parameters according to users' needs.The experimental results show that the detection algorithm module in the software can complete the basic detection requirements,and also provides a convenient interface for the user interface of ship detection.