无人船运行轨迹图像细化分割方法
|
摘要
为了提高对无人船运行轨迹的规划和识别能力,需要对运行轨迹图像进行细化分割,提出一种基于块匹配和边缘轮廓特征分割的无人船运行轨迹图像的细化分割方法。对采集的无人船运行轨迹灰度图像进行模板分块处理,对分块图像进行二值化分离,采用小波分解方法进行图像的边缘轮廓特征分解和尺度融合,结合统计特征重构方法进行运行轨迹的自适应信息增强,以增强后的轨迹图像为模板进行轮廓套索,实现轨迹图像的边缘轮廓特征提取,在分块图像中以边缘轮廓特征为基准线,实现无人船运行轨迹图像的细化分割。仿真结果表明,该方法进行无人船运行轨迹图像分割的特征分辨力较好,分割精度较高,挺高了图像的输出辨识能力。
In order to improve the planning and the recognition ability of the unmanned ship trajectory, need to track image thinning and segmentation, this paper proposed a segmentation method for refinement of unmanned ship trajectory image segmentation and contour block matching based on feature. The unmanned ship trajectory gray image processing template block, sub block image banalization, edge contour feature using wavelet image decomposition method and decomposition scale adaptive information fusion, combining the statistical characteristics of the trajectory reconstruction method in trajectory enhancement, enhanced contour like lasso as template, extract the edge contour feature trajectory images, in image blocks with edge features as a baseline, to achieve refinement unmanned ship trajectory image segmentation. The simulation results show that the unmanned ship trajectory image segmentation using the method the feature resolution is better, the segmentation precision is high, and the output identification ability of the image is higher.
In order to improve the planning and the recognition ability of the unmanned ship trajectory, need to track image thinning and segmentation, this paper proposed a segmentation method for refinement of unmanned ship trajectory image segmentation and contour block matching based on feature. The unmanned ship trajectory gray image processing template block, sub block image banalization, edge contour feature using wavelet image decomposition method and decomposition scale adaptive information fusion, combining the statistical characteristics of the trajectory reconstruction method in trajectory enhancement, enhanced contour like lasso as template, extract the edge contour feature trajectory images, in image blocks with edge features as a baseline, to achieve refinement unmanned ship trajectory image segmentation. The simulation results show that the unmanned ship trajectory image segmentation using the method the feature resolution is better, the segmentation precision is high, and the output identification ability of the image is higher.
引文
[1]伊力哈木·亚尔买买提.基于改进的自适应分水岭图像分割方法研究[J].计算机仿真,2013,30(2):373-377.
[2]李惠光,孙思佳.融合边缘与区域信息的水平集分割算法[J].计算机工程与应用,2013,10(1):1210-1216.
[3]郑伟,张晶,李凯玄,等.基于相位一致性改进的活动轮廓分割模型[J].激光技术,2016,40(2):296-302.
[4]卞乐,霍冠英,李庆武.基于Curvelet变换和多目标粒子群的混合熵MRI图像多阈值分割[J].计算机应用,2016,36(11):3188-3195.
[2]李惠光,孙思佳.融合边缘与区域信息的水平集分割算法[J].计算机工程与应用,2013,10(1):1210-1216.
[3]郑伟,张晶,李凯玄,等.基于相位一致性改进的活动轮廓分割模型[J].激光技术,2016,40(2):296-302.
[4]卞乐,霍冠英,李庆武.基于Curvelet变换和多目标粒子群的混合熵MRI图像多阈值分割[J].计算机应用,2016,36(11):3188-3195.