一种机器视觉调定玻璃量器液面的方法

A Method of Setting the Liquid Level of Working Glass Container by Machine Vision

针对玻璃量器检定时人工调定凹液面误差大、易疲劳的问题,提出一种基于机器视觉与图像处理技术的玻璃量器液面调定的方法。方法采用了先快速大流量注水、再慢速小流量注水、最后滴注调定的方式。其中,快速大流量注水通过衡量法进行控制,慢速小流量注水和滴注调定则通过机器视觉进行控制。在机器视觉控制注水时,连续采集被检图像并检测图像中调定刻度线下方小容量区域内指定目标特征的灰度增量矩阵。利用液面灰度大的特点,当灰度增量矩阵中最大值大于液面灰度阈值时,与最大值对应的目标特征即可作为液面所在位置的标记,从而为计算机实时识别液面提供方法和决策。在机器视觉控制滴注时,通过不断检测调定刻度线上边缘的灰度,利用液面经过边缘时的灰度变化趋势,将其拐点的确立作为调定结束的条件。该方法在容量瓶校准机器人研究方面得到了实际应用。

Focus on the problems of large error and easy fatigue in manually setting the concave liquid level in the calibration of working glass containers,a method of setting the liquid level of working glass container based on machine vision and image processing technology was proposed.Technical path of water injection was rapid large flow firstly and then slow small flow,finally drip setting.The rapid large flow was controlled by weighing method.Slow small flow and drip setting were controlled by machine vision.In machine vision control of large flow,the image of water injection was taken continuously and the gray increment matrix of the specified target features in the small volume area below the setting line was measured and calculated.Based on lager gray of liquid surface,when the maximum value in the gray increment matrix was greater than a threshold,the feature corresponding to the maximum value could be used as the mark of the position of the liquid surface,so as to provide a method and decision-making for the computer to identify the liquid surface in real time.In machine vision control of drip setting,constantly detected the gray level of the edge of the setting line as the liquid surface passing through the edge.And establishment of the inflection point in the gray level increment was taken as the condition for the end of setting.The method applies in the research on volumetric flask calibrating robot.