基于图像匹配的多船舶图像拼接方法

Multi-ship image mosaic method based on image matching

传统多船舶图像拼接方法在使用实践过程中,存在无法对多艘高度相似的船舶进行特征区分,导致在拼接过程中拼接位置出现错误的问题。提出基于图像匹配的多船舶图像拼接方法,首先通过引入M-KAZE算法,对高相似度的船舶进行细节的非线性尺度特征识别计算,使每张船舶图像的特征更为清晰;接着,引入稀疏特征像素点拼接算法对多张船舶图像的稀疏特征进行关联计算,从而完成多艘船舶图像的准确拼接;最后,通过仿真实验证明提出方法的可行性。

In the process of using traditional Multi-ship image mosaic methods,it is impossible to distinguish the features of several ships with similar height,which results in the wrong mosaic position in the process of mosaic.A Multi-ship image mosaic method based on image matching is proposed.Firstly,the M-KAZE algorithm is introduced to recognize and calculate the non-linear scale features of ships with high similarity,which makes the features of each ship image clearer.Then,the sparse feature points mosaic algorithm is introduced to correlate the sparse features of several ship images,so as to achieve the accuracy of Multi-ship images.Finally,the feasibility of the proposed method is proved by simulation experiments.