基于椭圆重叠区域的蚕茧图像融合方法

针对现有线性权值分配方法对蚕茧图像进行融合时会在高纬度区域出现蚕茧边缘分界线的问题，提出一种基于椭圆重叠区域的蚕茧图像融合方法。首先，利用最小二乘拟合算法得到待拼接图像中蚕茧表面的等效椭圆，输出其在该图像坐标下的位置方程；再根据图像配准过程中得到的横、纵轴位移距离建立蚕茧图像融合的数学模型，获取等效椭圆重叠区域的点集；最后通过椭圆重叠区域的最大宽建立改进的三角函数权重算法，对椭圆重叠区域进行融合处理。结果表明：该方法的融合效果比渐入渐出算法和三角函数权重算法更优，能够有效消除蚕茧边缘分界线，得到具有良好视觉效果且包含更多信息量的蚕茧融合图像。

Cocoon image fusion method based on ellipse overlapping area

Aiming at the appearance of boundary lines of the cocoon in the high latitude area when using the existing linear weight assignment method to fuse cocoon images, a cocoon image fusion method based on elliptical overlapping area was proposed. The equivalent ellipse of the cocoon surface in the image to be spliced was obtained by using the least squares fitting algorithm, and its position equation in the image coordinates was output. A mathematical model of cocoon image fusion was then established according to the displacement along the horizontal and vertical axes obtained in the image registration process, and the point set of the overlapping area of the equivalent ellipse was obtained. An improved trigonometric function weight algorithm was established through the maximum width of the ellipse overlapping area to fuse the ellipse overlapping area. The experimental results show that the fusion effect of this method is better than the fading in and out algorithms and trigonometric function weight algorithms, which can effectively eliminate the boundary line of the cocoon, and obtain a cocoon fusion image with good visual effect and more information.