应用短程线主动轮廓线的图像多目标面积同时测量

Simultaneous measurement of area for multiple image objects based on geodesic active contour

提出了一种新颖的图像多目标面积同时测量的方法,用于精确测量多个不同形状的图像目标面积。该方法应用了短程线主动轮廓线模型,分2个步骤进行。首先利用水平集函数φ的迭代使主动轮廓线由初始位置向各个目标的轮廓边缘收敛。其次,对于收敛后的主动轮廓线,分别计算出各目标边界的亚像素面积和图像目标的内部像素个数,从而同时求出各个图像目标的面积。实验结果表明,该方法的测量重复性误差<±0.5%;和传统的面积测量方法相比,具有测量效率高(同时测量多个目标面积)和测量精度高的优点。

A novel scheme to the area measurement of multiple image objects is proposed for simultaneous and exactly measurement of multiple objects＇ areas with different shapes. By using a model of geodesic active contour, this scheme consists of two steps. Firstly, the active contour converges from an initial position to the object＇s contour edge via the iteration of level set functional φ. Secondly, for the converged active contour, a sub-pixel area and the number of internal pixels of image objects are computed to obtain every object＇s area simultaneously. The experimental results indicate that measurement repetition error of proposed scheme can be decreased to ±0.5 %. Compared to conventional area measurement schemes, this scheme has highly efficient and highly accurate.