Aiming at the difficulty of garbage cleaning in small or medium-sized waters, a garbage detection algorithm for surface cleaning robot based on machine vision is proposed by analyzing the characteristics of water surface images. First, the mean-shift algorithm is used to smooth ripples as much as possible while maintaining the edge of the target; Then the color difference gray model and the improved OTSU threshold method are used to segment the water surface target, which can effectively suppress the problems of uneven illumination and reflection on the water surface; Next, the morphological processing of close and open operation is performed, the small holes are filled, and the isolated noise points are eliminated; Finally, the center-of-mass coordinates, the minimum external moment of each target object is calculated, the targets are marked to provide the basis for further actions of the robot. The experimental results show that this method can overcome the complex environmental impact of water and extract the floating garbage on the surface quickly and accurately.%针对中小型水域漂浮垃圾清理困难问题,通过分析水面图像的特点,提出基于视觉的水面机器人垃圾检测方法.使用均值漂移(Meanshift)滤波平滑算法,在保持目标边缘的基础上尽可能平滑波纹的干扰;使用基于色差灰度模型与改进的最大类间方差(OTSU)法分割出水面目标,能有效地抑制水面光照不均和水面倒影的问题;采用先闭运算后开运算的形态学处理,填充小孔、消除孤立噪声点;最后计算各目标对象的质心坐标、最小外界矩,对目标进行标记,为机器人的进一步动作提供依据.实验结果表明该方法能够较好地克服水面复杂环境影响,快速、准确地提取出水面的漂浮垃圾.
展开▼
