基于形态学和光流方程的光斑中心计算方法

针对图像序列中微小光斑中心位置的计算,本文提出了一种结合灰度形态学原理和光流估计方程的中心计算方法。在对图像序列逐一进行灰度膨胀后,根据光流方程对连续三帧图像进行位置转换并叠加,完成图像中微小目标的能量增强,从而使传统的重心算法在处理这一类图像时能够达到亚像素精度。经过误差分析和实验证明,这一方法有效的解决了微小目标的中心位置计算问题。

Spot centre arithmetic based on morphology and optical flow equation

To calculate the centre of small light points in images, an effective centre arithmetic which combines morphology with optical flow estimation is proposed. After dilating image sequences, three consecutive frames of the image are operated by using optical flow equation, then they are accumulated to enhance the small object energy in the images. This arithmetic makes the classical centroid method reach subpixel precision when it processes this kind of images. By error analysis and experiments, it can solve the problem of small target centre determination.