基于Otsu图像分割的颗粒与气泡碰撞检测方法

    A collision detection method of particle and bubble based on Otsu image segmentation

    • 摘要: 为了准确地检测出浮选过程中颗粒与气泡是否发生碰撞,提出了一种基于形态学图像处理和最大类间方差法(Otsu)的颗粒与气泡碰撞检测方法。首先利用图像的预处理操作优化图像,通过最大类间方差法分割出目标区域与背景区域,然后结合最小二乘法将颗粒与气泡拟合成圆,方便求取坐标信息及半径,最后将所有的颗粒位置提取到一幅图像中,实现了对颗粒位置的跟踪,避免了连续帧图像跟踪的复杂程度。结果表明,该算法能够快速地检测出颗粒与气泡是否发生碰撞,有效地节省了人工试验的工作时间。

       

      Abstract: In order to detect whether the collision occurs between particle and bubble in flotation accurately, a collision detection method of particle and bubble, which based on morphological image processing and Otsu method, is presented. First of all, It utilizes image pre-processing to optimize images, the target area and the background area are segmented by Otsu method. Then, particles and bubbles are fitted to a circle combined with the least-square method, it’s convenient to obtain the coordinate information and radius. Finally, the position of all particles is extracted to an image, which achieves the particle tracking and avoids the complexity of the continuous frame image tracking. The results indicate that this algorithm can detect the collision between particle and bubble quickly, and the working hours of manual test are saved effectively.

       

    /

    返回文章
    返回