Tennis automatic ball-picking robot based on image object detection and positioning technology

Abstract With the quick advancement of technology, tennis ball-picking robots have been maturely applied. However, currently, automatic tennis ball-picking robots often have low accuracy in positioning. To improve the accuracy of tennis automatic robot positioning, a weighted factor and virtual reference label improved indoor positioning algorithm is proposed, combined with radio frequency identification (RFID) technology. This method is applied to the automatic tennis ball-picking robot and constructs a tennis automatic ball-picking robot model based on improved indoor positioning algorithm and RFID technology. Comparing the effectiveness of the proposed improved algorithm, it was found that the precision and average accuracy of the algorithm were 0.983 and 94.6%, respectively, which were better than the comparison algorithms. In addition, an effectiveness comparison analysis was conducted on the tennis automatic robot model based on this algorithm, and it was found that the positioning effectiveness of the model was significantly better than the comparison model. Moreover, a positioning comparison experiment was conducted on the model, and the results showed that the positioning accuracy of the model was better than that of the comparison model. The above outcomes illustrate that the improved algorithm proposed in the study and the tennis automatic ball-picking robot model based on this algorithm have good practical value, which is conducive to improving the accuracy of the robot’s image target positioning.