在线产品智能搬运机械手系统视觉识别技术

针对采用机械手从传送带上抓取位置不固定的产品并将产品按照固定位姿放置而需要复杂的视觉图像处理且成本较高的问题,提出了一种基于辅助定位机构的视觉识别技术。该技术首先以苹果手机为具体研究对象,在非极端光照条件下采集了不同光照下11组44张手机图像,并进行了预处理；然后根据各图像直方图的灰度分布特征,通过研究各图像中灰度值出现次数最多的灰度值和所有像素的灰度值均值的变化情况,确定了分段线性阈值分割准则；最后,采用面积特征找到合适的目标,再采用目标的位置特征对图像进行位姿识别,并给出了识别算法。以普通视频聊天摄像头为实验设备进行了算法适应性验证,实验结果表明,在非极端光照且光源环境基本稳定的情况下,该技术能够稳定可靠地实现手机位姿识别。该识别技术相对于多自由度机械手+摄像机的识别技术在成本上至少降低50%以上,可广泛应用于类手机形状产品的生产线上。

Visual Recognition Technology of Manipulator System for Intelligent Grasping Online Product

As grasping online products with unfixed location from the conveyor belt and placing them to the fixed position with the decided posture by the intelligent Manipulator usually need complex visual image processing and high cost of the intelligent Manipulator system, a visual recognition technology based on auxiliary positioning device was proposed. Firstly, with the Apple Mobile Phone as the specific research object, under non-extreme lighting conditions, 44 mobile phone images of 11 groups under different illumination conditions were collected and were pre-processed. Secondly, according to the grayscale distribution characteristics of histogram of each image, though research on the changing situation of the grayscale value with the maximum number of gray values and the mean of the gray value of all pixels for each image, the piecewise linear threshold segmentation criterion was induced. Finally, the right targets were found by applying area characteristic, and the position feature of the target is used to identify the image, and then the characteristic recognition algorithm were induced. With ordinary video camera as the experimental equipment, an algorithm adaptability experiment was implemented. The experimental results show that the visual approach can identify mobile posture stably and reliably under the non-extreme lighting condition and the stable light source environment. The recognition technology is more than 50% lower in cost than the multi-degree-of-freedom manipulator + camera recognition method, which can be widely used in the production line of the shape products of mobile phones.