针对单源舰船检测识别率偏低的问题,基于DS证据理论,提出了一种通过权值更新,改进DS证据的基本概率赋值的方法,并结合冲突解决方案,对SAR和可见光舰船检测得出的特征数据进行特征级融合检测。实验结果表明,该算法大大提高了舰船检测的识别率,降低了误判率。
As for the low accuracy of single source ship detection,based on Dempster-Shafer theory,this paper introduces an improved basic probability assign method which updated by weights.Using the solutions of conflicting evidences,the data from the ship detection of SAR and optical imagery are fused on feature level with the improved Dempster-Shafer theory,and the experiment results show that this method can improve the accuracy of ship detection greatly.