在城市园林建设中,对花卉色彩图像进行有效匹配,可以更好的提升城市园林建设水平.对花卉图像的匹配,需要利用未标定的透视图像,从而恢复参数以及标记点的三维空间坐标,完成对花卉色彩图像的匹配.传统方法剔除部分深度误差较大的花卉三维点,计算出整体花卉场景的三维点云,但忽略了恢复相应参数,导致匹配精度偏低.提出基于分水岭的城市园林建设中花卉色彩图像匹配方法.融合于分水岭方法进行花卉场景图像的分割,利用模糊均值方法将相似度较高的花卉形态区域进行合并,得到花卉边缘点匹配约束条件,将花卉轮廓上已匹配的特征角点定义为基准点,并利用未标定的透视图像恢复相机的参数以及标记点的三维空间坐标,以此为依据完成对城市园林建设中花卉场景图像匹配.仿真证明,所提方法花卉色彩图像匹配效果好,为提升城市园林建设水平奠定了结实的基础.%An image matching method of flower color in urban landscape construction is proposed based on watershed.The segmentation for scene image of flour integrated with method of watershed is carried out,and the fuzzy mean method is used to combine flower morphological area with high degree of similarity,and then the constraint condition of peripheral point of flower is obtained.The matched feature angular point on contour of flower is defined as datum point,and the unmarked perspective image is used to recover camera parameter and three-dimensional space coordinates of gauge point.Thus,the image matching is completed.Simulation proves that the method has good matching effect.It lays strong foundation for improving level of urban landscape construction.
展开▼
