基于最优信噪比的饮水机动态水位识别方法

利用机器视觉技术识别通用饮水机水位能为家政服务机器人赋予更广阔的应用空间。通过分析饮水机水桶的颜色特征，利用彩色图像的单一通道实现水桶和背景的分割，利用水桶被动震动后液体水和固体容器达到稳定状态的时间长短差异实现水位线信号和噪声的分离。然后，利用提出的基于最优信噪比的灰度投影法实现水位线的定位，达到水桶内水量探测的目的。实验结果表明，该方法是有效的，并且其识别率较高。

Dynamic Water Level Identification for Dispenser Based on Optimal SNR

The domestic service robots can be given a broader application space by applying the machine vision technol-ogy to identify the water level of dispenser. Through the analysis of the color characteristics for bucket dispenser, the method is proposed by using single channel of color image, the separation is performed between bucket and back-ground. When the dispenser gets vibration,the length of time for the liquid and container reaching back to steady state is different,by using the time length difference,the separation of the water level signal and noise can be accomplished. Then, the position of water level is achieved and the goal of water level detection can be reached by using our pro-posed optimal SNR based grey projection method. Experimental results show the proposed method is effective and has high identification rate.