基于视觉识别的液压拉伸器四轴对准装置OA

Bolts and nuts are important fasteners for heavy industry,and bolt tensioners are widely used in bolt pretightening due to their uniform force distribution and accurate control over the bolt pretightening force.However,the positioning accuracy required for the tensioner during operation is high,so it is necessary to keep the axis of the tensioner aligned with the bolt.The article designs a tensioner alignment device based on visual recognition.The depth camera image is input into YOLOv5 to obtain sub images,which are then subjected to image enhancement and threshold segmentation based on annealing algorithm.After that,the precise pose and center coordinates of the bolt are obtained by hexagonal fitting based on Hough transform.Finally,the coordinate information of the bolt is transmitted to the actuator to drive the motor for positioning.Experiments verified that the image positioning accuracy and device positioning accuracy can be within 1.5 mm,and the target recognition accuracy can reach over 90%.The designed mechanical structure can complete the alignment work satisfactorily.