凹点搜寻下重叠区域图像自动分离方法仿真
|
摘要
重叠区域图像分离是图像处理技术中的一个重要环节,在医疗诊断、工业加工和军事防御等众多领域占据重要地位。针对当前方法在对重叠区域图像进行分离时,过于依赖极限腐蚀运算,一旦极限腐蚀运算没有较好的将图像重叠区域腐蚀开,无法得到准确的距离标记,将会导致重叠区域图像无法有效分离。提出一种基于凹点搜寻的重叠区域图像自动分离方法,在低尺度空间上,对重叠区域图像轮廓曲率进行计算;利用高斯函数的宽度参量控制重叠区域轮廓的平滑程度;计算获得重叠区域每个边缘像素的曲率后,去除曲率大于0的凸点,选取曲率极小值的点作为候选凹点,同时去除虚假凹点,完成重叠区域图像的凹点检测;根据重叠区域图像凹点个数与目标对象个数之间的对应关系确定重叠类型,给出了重叠区域图像的不同凹点匹配规则;通过连接的匹配凹点,实现重叠区域图像的自动分离。仿真测试结果证明,所提方法能够实现图像不同重叠类型的有效分离,分离对象具有较高的信噪比,结构相似性和品质,且没有发生较大形变。
This paper proposes an automatic image separation method in overlapping area based on concave spot search. In low-scale space,the contour curvature of image in overlapping area is calculated. The width parameter of Gaussian function is used to control the smoothness of contour in overlapping area. After calculating each curvature of edge pixel in overlapping area,we remove the salient point whose curvature is greater than 0,and select the point with the minimum curvature as the candidate concave point. At the same time,the false concave point is removed to complete the detection of false concave point in the overlapping area. According to the corresponding relation between number of concave points of image in overlapping area and the number of the target objects,the overlapping type is determined and the matching rule of different concave points of image in overlapping area is given. Thus,the automatic separation of overlapping images is realized through connected matching concave points. Simulation results prove that the proposed method can effectively separate different overlapping types in images. The separated objects have high signal-to-noise ratio,structural similarity and quality.
This paper proposes an automatic image separation method in overlapping area based on concave spot search. In low-scale space,the contour curvature of image in overlapping area is calculated. The width parameter of Gaussian function is used to control the smoothness of contour in overlapping area. After calculating each curvature of edge pixel in overlapping area,we remove the salient point whose curvature is greater than 0,and select the point with the minimum curvature as the candidate concave point. At the same time,the false concave point is removed to complete the detection of false concave point in the overlapping area. According to the corresponding relation between number of concave points of image in overlapping area and the number of the target objects,the overlapping type is determined and the matching rule of different concave points of image in overlapping area is given. Thus,the automatic separation of overlapping images is realized through connected matching concave points. Simulation results prove that the proposed method can effectively separate different overlapping types in images. The separated objects have high signal-to-noise ratio,structural similarity and quality.
引文
[1]桑园,王莉丽.公路路面摩擦力属性图像纹理特征的识别方法研究.[J].科技通报,2016,7(32):188-192.
[2]秦长泽,刘琼.基于改进分水岭算法的球团矿图像分割[J].武汉科技大学学报,2015,38(2):134-137.
[3]关涛,等.基于稀疏轮廓点模型的彩色重叠细胞图像分割[J].计算机研究与发展,2015,52(7):1682-1691.
[4]彭晏飞,周娟,訾玲玲.于卫星地域遥感图像分割方法研究仿真[J].计算机仿真,2017,34(1):332-335.
[5]王丹丹,等.融合K-means与Ncut算法的无遮挡双重叠苹果目标分割与重建[J].农业工程学报,2015,31(10):000227-234.
[6]邱瑞,祝日星,许宏科.基于改进分水岭算法的图像分割算法[J].吉林大学学报:理学版,2017,55(3):629-634.
[7]周静波,任永峰,严云洋.基于视觉显著性的非监督图像分割[J].计算机科学,2015,42(8):52-55.
[8]徐越,等.基于Snake模型与角点检测的双果重叠苹果目标分割方法[J].农业工程学报,2015,31(1):196-203.
[9]苗中华,等.自然环境下重叠果实图像识别算法与试验[J].农业机械学报,2016,47(6):21-26.
[10]赵福娟,李玲,辛化梅.凝胶图像重叠蛋白质点分割算法研究[J].计算机工程与应用,2017,53(5):212-215.
[2]秦长泽,刘琼.基于改进分水岭算法的球团矿图像分割[J].武汉科技大学学报,2015,38(2):134-137.
[3]关涛,等.基于稀疏轮廓点模型的彩色重叠细胞图像分割[J].计算机研究与发展,2015,52(7):1682-1691.
[4]彭晏飞,周娟,訾玲玲.于卫星地域遥感图像分割方法研究仿真[J].计算机仿真,2017,34(1):332-335.
[5]王丹丹,等.融合K-means与Ncut算法的无遮挡双重叠苹果目标分割与重建[J].农业工程学报,2015,31(10):000227-234.
[6]邱瑞,祝日星,许宏科.基于改进分水岭算法的图像分割算法[J].吉林大学学报:理学版,2017,55(3):629-634.
[7]周静波,任永峰,严云洋.基于视觉显著性的非监督图像分割[J].计算机科学,2015,42(8):52-55.
[8]徐越,等.基于Snake模型与角点检测的双果重叠苹果目标分割方法[J].农业工程学报,2015,31(1):196-203.
[9]苗中华,等.自然环境下重叠果实图像识别算法与试验[J].农业机械学报,2016,47(6):21-26.
[10]赵福娟,李玲,辛化梅.凝胶图像重叠蛋白质点分割算法研究[J].计算机工程与应用,2017,53(5):212-215.