Accurate Tracking Method of Basketball Flight Trajectory Based on YOLO and Multi-level Data Association

There are problems such as large tracking deviation in basketball flight trajectory tracking. To this end, this paper proposes an accurate tracking method of basketball flight trajectory based on YOLO and multi-level data association. Digital imaging equipment is used to collect basketball flight image video sequences, and the frame difference of the basketball flight image video sequence is defined. The spatial invariant feature decomposition method is used in the three-dimensional flight space to obtain the energy function of the binary multi-dimensional image and analyze the inherent mode function of the image. The visual feature extraction method is used to extract the basketball flight trajectory; the Gaussian kernel function is used to set the Gaussian distribution to approximate the two-dimensional convolution operator, and the weight of the product of each neighborhood pixel value and the weight is used as the pixel value of the pixel, and the size of the weight increases as the distance from the center point decreases. The image information is retained. The Gaussian proportion mixture model is used to extract the texture information and noise information existing in the image, and the Rudin–Osher–Fatemi model is used to eliminate the texture information existing in the image to achieve denoising processing of basketball images. The mining algorithm ML_T2 of multi-level association rules is used to mine multi-level data of basketball flight trajectory characteristics. Support and confidence are introduced to describe the importance of mining multi-level data of basketball flight trajectory characteristics. All frequent item sets are found. According to the obtained frequent items Set, generate corresponding strong association rules, set a minimum confidence threshold, and realize multi-level data mining of basketball flight trajectory characteristics. On this basis, build a network structure for basketball flight trajectory tracking and enhance basketball flight trajectory tracking image data. Set the loss function for basketball flight trajectory tracking, and integrate the multi-level data of basketball flight trajectory characteristics to build an accurate basketball flight trajectory tracking model based on YOLO and implement the research. The feasibility of this method was verified through experiments.