Abstract
Robot control based on visual information perception is a hot topic in the industrial robot domain and makes robots capable of doing more things in a complex environment. However, complex visual background in an industrial environment brings great difficulties in recognizing the target image, especially when a target is small or far from the sensor. Therefore, target recognition is the first problem that should be addressed in a visual servo system. This paper considers common complex constraints in industrial environments and proposes a You Only Look Once Version 2 Region of Interest (YOLO-v2-ROI) neural network image processing algorithm based on machine learning. The proposed algorithm combines the advantages of YOLO (You Only Look Once) rapid detection with effective identification of ROI (Region of Interest) pooling structure, which can quickly locate and identify different objects in different fields of view. This method can also lead the robot vision system to recognize and classify a target object automatically, improve robot vision system efficiency, avoid blind movement, and reduce the calculation load. The proposed algorithm is verified by experiments. The experimental result shows that the learning algorithm constructed in this paper has real-time image-detection speed and demonstrates strong adaptability and recognition ability when processing images with complex backgrounds, such as different backgrounds, lighting, or perspectives. In addition, this algorithm can also effectively identify and locate visual targets, which improves the environmental adaptability of a visual servo system
Highlights
Robot visual servo is a type of technology that uses visual sensors to obtain environmental information and give the corresponding movement command to the robot controller, so it can be considered as an imitation of human eyes and arms [1]
There are many types of images in the field of vision of visual systems in the industrial environment, especially for a mobile robot arm installed on an automated guided vehicle (AGV) [22]; namely, color, shape, size, and texture of a target object are similar to a certain extent, so how to reasonably classify targets is a prerequisite for realizing the function of robot intelligent perception in a visual servo system
In order to test the effectiveness of the proposed algorithm, an industrial robot visual servo platform was built for verification in the laboratory
Summary
Robot visual servo is a type of technology that uses visual sensors to obtain environmental information and give the corresponding movement command to the robot controller, so it can be considered as an imitation of human eyes and arms [1]. The Overfeat algorithm proposed by Sermanet et al [8] uses convolutional networks to extract image features, performs target detection on each sliding window, and completes the precise positioning task of the target. Based on the requirements for the real-time processing of environmental information by environmental perception systems, this paper designs a joint architecture for detection and segmentation; i.e., target detection and semantic segmentation share the same feature extraction network through joint training while reducing the inference time, effectively improving the performance of subtasks.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
