With the development of the economy, people's living standards continue to improve, and they have a certain amount of time to improve themselves and cultivate their interests and hobbies. Tennis is a sport and people's attention to it is constantly increasing. In this study, we analyzed tennis match data to answer various questions related to tennis. Firstly, in response to the first question, we constructed a model to capture the game process and evaluate the performance of players. By preprocessing and analyzing the data, including eliminating outliers, filling in missing values, and standardizing the data, we obtained prepared analytical data. Subsequently, using principal component analysis, three representative principal components were identified. In addition, by using the entropy weight method, we calculated the weights, enabling us to derive a momentum formula for visualizing the temporal changes during the competition process. Secondly, regarding the second question, we explored whether there is a correlation between the game results and momentum. By using the momentum model we established, we used Pearson correlation analysis to calculate the correlation coefficient, thereby determining the relationship between momentum and game results.

In order to solve the problem of calculating the efficiency of the heliostat distribution, based on the knowledge of physics and astronomy, this paper establishes a variety of physical parameter calculation models, comprehensively considers the relevant data and constraints of the heliostat field, solves and calculates the parameters of the heliostat field, and calculates the optical efficiency and output thermal power of the heliostat according to the relevant calculation formulas. In order to optimize the design of heliostat field, the optimal results of each parameter of heliostat field are obtained through multiple iterative calculations and comparative analysis by using the simulated annealing algorithm, which realizes the optimization of the heliostat field and ensures the highly efficient performance of the heliostat field.

 Since the 1970s, road disease detection at home and abroad has become a research hotspot. Intelligent detection of road diseases can effectively reduce traffic safety accidents and maintain social safety. Based on R-CNN and YOLO models, our team innovatively proposed a detection method of many algorithms, which further improved the detection efficiency. In terms of hardware, micro high-speed industrial camera, LPMS series and iAHRS-MO attitude sensor are adopted. At present, the team has completed a complete set of detection methods and structures, which can adapt to many future needs and changes.

 In order to realize the purpose of better inspection on the outer surface of high-rise buildings, comprehensively detect the damage of the building surface, so as to find various problems, such as cracks, bulge and falling off, a UAV inspection image acquisition and processing system based on convolutional neural network is designed. The UAV uses open MV to realize the efficient processing of visible light image information, and uses an embedded system for image fusion processing. Using convolutional neural network model and multi-object detection algorithm, extract the feature value of broken position and combine multiple sensors for 3 D position transformation. Data reads and 3 D coordinate map presentation were performed using the app software. For the safety problem of ceramic tiles, a marking device is installed on the UAV. Calculate the threat range and use the application to send location information to maintenance personnel and alert information to residents, and provide real-time location safety monitoring. The experimental results show that the designed uav can realize comprehensive, automatic, intelligent, efficient and preventive inspection of the building health status and problems of buildings through inspection, data collection and processing, intelligent analysis and early warning.