A comprehensive and precise system and method for AI-powered physical training based on intelligent perception.

Abstract

This invention discloses an AI-powered intelligent perception system and method for comprehensive and precise physical training, relating to the field of physical training technology. The key technical solution includes the following steps: acquiring the test area for which physical training performance needs to be tested; statistically analyzing the influencing factors on the physical training performance of each object in the test area to obtain related influence characteristic factors; processing and analyzing the characteristic data of each feature factor to obtain a comprehensive influence correlation relationship; comparing and judging the comprehensive influence correlation relationship with the actual correlation influence relationship between the physical training performance of each object in the test area, and then outputting detection information; selecting key movement segments and core training periods from each object in the test area based on the detection information. The effect is that it can accurately evaluate physical training performance from multiple dimensions, providing a comprehensive and reliable basis for training effect evaluation.

Description

Technical Field

[0001] This invention relates to the field of physical training technology, and more specifically, to an AI-powered intelligent perception system and method for comprehensive and precise physical training. Background Technology

[0002] With increasing emphasis on health and physical fitness, physical training has been widely applied in various fields, such as sports competitions, military training, and fitness shaping. However, traditional physical training methods have limitations in evaluating and optimizing training effects. Past assessments of physical training performance often focused on only a few intuitive indicators. For example, in sports training, they might only focus on basic data such as an athlete's running speed or the weight used in strength training. This singular assessment method cannot comprehensively reflect the trainee's true physical fitness level or the combined influence of various factors during training. Summary of the Invention

[0003] To address the shortcomings of existing technologies, the present invention aims to provide an AI-powered intelligent perception-based system and method for comprehensive and precise physical training.

[0004] To achieve the above objectives, the present invention provides the following technical solution:

[0005] A comprehensive and precise method for physical fitness training based on AI intelligent perception, comprising the following steps:

[0006] Step S1: Obtain the test area where physical fitness training performance needs to be tested, and statistically analyze the influencing factors of physical fitness training performance testing for each object in the test area to obtain the associated influencing characteristic factors.

[0007] Step S2: Process and analyze the feature data of each feature factor in the correlation influence feature factor to obtain the comprehensive influence correlation relationship 1. Compare and judge the comprehensive influence correlation relationship 1 with the actual correlation influence relationship 1 between the physical training performance of each object in the test area and output the detection information.

[0008] Step S3: Select key movement segments and core training periods from each object in the test area based on the detection information, detect the physical training data of key movement segments and core training periods respectively, and perform comprehensive statistics to obtain preprocessed physical fitness performance data. Statistically analyze the correlation between the performance data in the preprocessed physical fitness performance data to obtain the actual correlation and influence relationship 2.

[0009] Step S4: Evaluate the comprehensive influence correlation 2 to which the preprocessed physical fitness performance data belongs. After comparing the similarity values ​​between the comprehensive influence correlation 2 and the actual influence correlation 2, divide the test area into local region 1 and local region 2. Based on the preprocessed physical fitness performance data, perform sequential physical fitness training performance analysis on local region 1 and local region 2 respectively, and output the test results.

[0010] Preferably, step S1 specifically includes the following steps:

[0011] Obtain the test area where physical training performance needs to be tested, and detect the physical training performance data of each object in the test area to obtain the physical training performance test value.

[0012] The correlation between the various performance data in the physical training performance test values ​​is evaluated to obtain the actual correlation influence relationship.

[0013] The associated influencing feature factors are obtained by statistically analyzing the speech features, action posture features, training time series features, and environmental spatial features of each object in the test area.

[0014] Preferably, step S1 specifically includes the following steps:

[0015] Based on the speech features, action and posture features, training time series features, and environmental spatial features in the correlation influence feature factors, the correlation degree relationship between the physical training results of each object in the test area is extracted to obtain the speech influence correlation relationship, posture influence correlation relationship, time series influence correlation relationship and spatial influence correlation relationship;

[0016] A comprehensive correlation assessment is conducted on the correlations of speech, posture, temporal, and spatial effects to obtain a comprehensive correlation relationship 1. The similarity value between the comprehensive correlation relationship 1 and the actual correlation relationship 1 between the physical training performance of each object in the test area is then calculated.

[0017] If the similarity value between the overall influence correlation one and the actual influence correlation one between the physical training performance of each object in the test area is less than the preset similarity judgment threshold, then the detection information will be output.

[0018] Preferably, based on the speech features, posture features, training time series features, and environmental spatial features in the correlation influence feature factors, the correlation degree between the physical fitness training scores of each object in the test area is extracted to obtain the speech influence correlation, posture influence correlation, time series influence correlation, and spatial influence correlation. Specifically, this includes the following steps:

[0019] Based on the speech features in the correlation influence feature factors, the degree of correlation between the various performance data in the physical training performance test values ​​under the influence of speech features is evaluated to obtain the speech influence correlation relationship;

[0020] Based on the action posture features in the correlation influence feature factors, the correlation between the various performance data in the physical training performance test values ​​under the influence of action posture features is evaluated to obtain the posture influence correlation relationship.

[0021] Based on the training time series characteristics in the correlation influence feature factors, the correlation between the various performance data in the physical training performance test values ​​under the influence of the training time series characteristics is evaluated to obtain the time series influence correlation relationship;

[0022] Based on the environmental spatial characteristics in the correlation influence characteristic factors, the spatial influence correlation is obtained by evaluating the degree of correlation between the various performance data in the physical training performance test values ​​under the influence of environmental spatial characteristics.

[0023] Preferably, if the similarity value between the overall influence relationship 1 and the actual influence relationship 1 between the physical training performance of each object in the test area is less than a preset similarity threshold, then the detection information is output, specifically including the following steps:

[0024] An assessment of the comprehensive correlations of speech, posture, temporal, and spatial influences yields Comprehensive Influence Correlation 1.

[0025] The comparison result 1 is obtained by comparing the comprehensive impact relationship 1 and the actual impact relationship 1.

[0026] A preset similarity threshold is set. If the similarity value in the comparison result is greater than or equal to the similarity threshold, the test results are output after the physical training performance of each object in the test area is finely processed according to the physical training performance test value and the actual correlation influence.

[0027] If the similarity value in the comparison result is less than the similarity judgment threshold, then the physical training performance data of each object in the test area will be detected and the detection information will be output.

[0028] Preferably, step S3 specifically includes the following steps:

[0029] Based on the detection information, key movement segments and core training periods are selected from each subject in the test area; wherein, the key movement segments refer to the combination segments of movements in the physical training process that have a key impact on performance, and the core training periods refer to the time periods in the entire training cycle that have an impact on physical fitness improvement and performance.

[0030] Physical training data one is obtained by detecting physical training data of the key movement segments, and physical training data two is obtained by detecting physical training data of the core training period.

[0031] The preprocessed physical fitness performance data is obtained by performing comprehensive statistical calculations on the physical fitness training data one and physical fitness training data two.

[0032] The correlation between the preprocessed physical fitness data is evaluated to obtain the second actual correlation influence relationship.

[0033] Preferably, the second comprehensive impact correlation of the pre-processed physical fitness performance data is assessed, specifically as follows:

[0034] The correlation between preprocessed physical fitness performance data and the physical fitness training performance of each subject in the test area under the influence of speech features, movement posture features, training time sequence features, and environmental spatial features is obtained as a comprehensive influence correlation relationship 2.

[0035] Preferably, after comparing and judging the similarity values ​​between the comprehensive influence relationship 2 and the actual influence relationship 2, the test area is divided into local region 1 and local region 2, specifically including the following steps:

[0036] The comparison result 2 is obtained by comparing the comprehensive impact relationship 2 and the actual impact relationship 2;

[0037] If the similarity value in the comparison result 2 is greater than or equal to the similarity judgment threshold, then the test area is divided into local region 1 and local region 2 according to the actual association and influence relationship 2. The arrangement position of the actual association and influence relationship of each object in local region 1 is before the arrangement position of the actual association and influence relationship of each object in local region 2.

[0038] Preferably, the test results are output after analyzing the physical training performance of local region one and local region two respectively based on the preprocessed physical fitness performance data. Specifically, this includes the following steps:

[0039] After processing the physical training results of the local area 1 based on the preprocessed physical fitness data, the processing result 1 is output.

[0040] Based on the preprocessed physical fitness performance data and the preset interval processing time of processing result one, the physical fitness training performance of the local area two after the interval processing time is processed and then the processing result two is output.

[0041] The test result is formed by combining the first processing result and the second processing result.

[0042] A comprehensive and precise physical training system based on AI intelligent perception includes:

[0043] The statistics module retrieves the test area where physical training performance needs to be tested, and statistically analyzes the influencing factors of physical training performance testing for each object in the test area to obtain the associated influencing characteristic factors.

[0044] The processing and analysis module processes and analyzes the feature data of each feature factor in the correlation influence to obtain the comprehensive influence correlation relationship 1. It then compares and judges the comprehensive influence correlation relationship 1 with the actual correlation influence relationship 1 between the physical training performance of each object in the test area and outputs the detection information.

[0045] Select the statistics module, select key movement segments and core training periods from each object in the test area based on the detection information, detect the physical training data of key movement segments and core training periods respectively, and perform comprehensive statistics to obtain preprocessed physical fitness performance data. Statistically analyze the correlation between the performance data in the preprocessed physical fitness performance data to obtain the actual correlation and influence relationship.

[0046] The output module evaluates the comprehensive influence correlation 2 to which the preprocessed physical fitness performance data belongs. After comparing the similarity values ​​between the comprehensive influence correlation 2 and the actual influence correlation 2, it divides the test area into local region 1 and local region 2. Based on the preprocessed physical fitness performance data, it performs sequential analysis and processing of physical fitness training performance in local region 1 and local region 2, and then outputs the test results.

[0047] Compared with the prior art, the present invention has the following beneficial effects:

[0048] This technical solution comprehensively analyzes the influencing factors of physical training performance tests for various subjects in the test area, covering voice features, movement and posture features, training time series features, environmental and spatial features, etc., forming correlated influencing feature factors. Based on this, it deeply analyzes the correlation between the feature data, enabling accurate evaluation of physical training performance from multiple dimensions, making the evaluation results closer to the actual training level, and providing a comprehensive and reliable basis for evaluating training effectiveness.

[0049] This approach selects key movement segments and core training periods from each subject within the testing area. Key movement segments are combinations of movements in the physical training process that significantly impact performance, while core training periods are timeframes within the entire training cycle that have a substantial impact on fitness improvement and performance. For subjects in different local areas, pre-processed fitness performance data is used to analyze and process their sequential fitness training results. This method enables the development of training plans tailored to individual differences, providing personalized fitness training guidance, improving training efficiency and effectiveness, and helping trainees more effectively improve their fitness and performance.

[0050] By comparing the similarity values ​​of the comprehensive influence correlation and the actual influence correlation, the test area is divided into Local Region 1 and Local Region 2. Based on this division, training resource allocation can be optimized in a targeted manner according to the characteristics of objects in different regions. This on-demand resource allocation method avoids unreasonable allocation and waste of resources, improves the efficiency of training resource utilization, and allows limited training resources to play a greater role. Attached Figure Description

[0051] Figure 1 This is a flowchart illustrating an AI-powered intelligent perception method for comprehensive and precise physical training proposed in this invention.

[0052] Figure 2 This invention presents a schematic diagram of a module for an AI-powered intelligent perception-based, multi-dimensional, and precise physical training system. Detailed Implementation

[0053] Reference Figures 1 to 2 As shown.

[0054] Example 1 further illustrates the AI-powered intelligent perception-based all-dimensional precision physical training system and method proposed in this invention.

[0055] A comprehensive and precise method for physical fitness training based on AI intelligent perception, comprising the following steps:

[0056] Step S1: Obtain the test area where physical fitness training performance needs to be tested, and statistically analyze the influencing factors of physical fitness training performance testing for each object in the test area to obtain the associated influencing characteristic factors.

[0057] Step S2: Process and analyze the feature data of each feature factor in the correlation influence feature factor to obtain the comprehensive influence correlation relationship 1. Compare and judge the comprehensive influence correlation relationship 1 with the actual correlation influence relationship 1 between the physical training performance of each object in the test area and output the detection information.

[0058] Step S3: Select key movement segments and core training periods from each object in the test area based on the detection information, detect the physical training data of key movement segments and core training periods respectively, and perform comprehensive statistics to obtain preprocessed physical fitness performance data. Statistically analyze the correlation between the performance data in the preprocessed physical fitness performance data to obtain the actual correlation and influence relationship 2.

[0059] Step S4: Evaluate the comprehensive influence correlation 2 to which the preprocessed physical fitness performance data belongs. After comparing the similarity values ​​between the comprehensive influence correlation 2 and the actual influence correlation 2, divide the test area into local region 1 and local region 2. Based on the preprocessed physical fitness performance data, perform sequential physical fitness training performance analysis on local region 1 and local region 2 respectively, and output the test results.

[0060] This application first identifies the testing area where physical training performance needs to be tested, and then uses tools such as sensors, video surveillance, and data recording devices to detect the physical training performance data of each object within that area and obtain the test values. For example, in a gym's physical training scenario, the data recording functions built into treadmills and strength training equipment, as well as installed cameras to capture movements, are used to obtain performance data such as the trainee's running speed, strength training weight, and number of repetitions.

[0061] By analyzing the speech features of each object (such as breathing rhythm and shouting commands during training), movement and posture features (position, angle, and movement trajectory of various body parts during training), training time series features (training intensity and rest intervals at different time periods), and environmental spatial features (temperature, humidity, and size of the training venue), the associated influencing feature factors can be obtained.

[0062] Based on the different features in the associated influence factors, the aforementioned speech, posture, temporal, and spatial influence relationships are comprehensively evaluated, and a multi-factor fusion analysis algorithm is used to obtain the first comprehensive influence relationship. Then, the first comprehensive influence relationship is compared with the actual influence relationship obtained by directly analyzing physical training performance test values, and the similarity value between the two is calculated. If the similarity value is less than a preset similarity threshold, it indicates that the comprehensive influence relationship obtained based on feature factor analysis differs significantly from the actual situation. In this case, detection information is output, suggesting that further analysis and adjustment of the training situation are needed.

[0063] Based on the detection information, key movement segments and core training periods are selected from each subject in the testing area. For example, in gymnastics training, by analyzing a large amount of training data and competition results of elite gymnasts, key movement segments such as flips and jumps that have a significant impact on performance are identified as key movement segments.

[0064] Using high-precision sensors and data acquisition equipment, physical training data 1 is obtained by detecting key movement segments, such as force output and joint angle changes in key movement segments; physical training data 2 is obtained by detecting physical training data during core training periods, such as heart rate and energy consumption during those periods.

[0065] The preprocessed physical fitness performance data is obtained by combining and statistically analyzing the first and second sets of physical fitness training data. Then, the correlation between the various performance data in the preprocessed physical fitness performance data is evaluated to obtain the second set of actual correlation and influence relationships.

[0066] The second comprehensive influence correlation of the preprocessed physical fitness performance data is to evaluate the correlation between the physical fitness training performances of each object in the test area under the influence of the preprocessed physical fitness performance data on the speech features, movement posture features, training time sequence features and environmental spatial features of each object.

[0067] Based on the pre-processed physical fitness performance data, the physical fitness training performance of local area one was processed to obtain processed result one. Then, based on the pre-processed physical fitness performance data and processed result one, a pre-set interval processing period was established. Taking into account factors such as training fatigue recovery and knowledge and skill consolidation, the interval length was determined, and the physical fitness training performance of local area two after the interval processing period was processed to obtain processed result two. Finally, processed result one and processed result two were combined to form the test result, providing a scientific basis for evaluating the effectiveness of physical fitness training and adjusting training plans.

[0068] Step S1 specifically includes the following steps:

[0069] Obtain the test area where physical training performance needs to be tested, and detect the physical training performance data of each object in the test area to obtain the physical training performance test value.

[0070] The correlation between the various performance data in the physical training test results was evaluated to obtain the actual correlation influence relationship 1;

[0071] The associated influencing feature factors are obtained by statistically analyzing the speech features, action posture features, training time series features, and environmental spatial features of each object in the test area.

[0072] By utilizing various sensors and data acquisition devices, the testing area for physical training performance evaluation is determined. For example, in a gym setting, using the built-in sensors of equipment such as treadmills and dumbbells, as well as additional motion tracking sensors, the physical training performance data of each individual in the testing area is detected, obtaining test values ​​such as running speed, strength training weight, and repetitions.

[0073] The correlation between the various data points in the acquired physical fitness training performance test values ​​is evaluated. For example, analyzing the correlation between the number of repetitions at different weights and the final strength improvement in a set of strength training exercises yields the first actual correlation influence relationship, thus providing a preliminary understanding of the interaction relationships among the various data points within the physical fitness training performance.

[0074] The speech characteristics of the target are statistically analyzed using speech recognition technology, such as the rhythm and commands shouted during training; action posture characteristics are obtained using computer vision and posture recognition, such as the angles and trajectories of different body parts during training movements; training time series characteristics are determined by time recording systems and data analysis, including the duration and intensity changes of different training stages; environmental spatial characteristics are recorded using environmental monitoring equipment, such as the temperature, humidity, and size of the training venue, and these characteristics are combined to obtain the associated influencing factor.

[0075] Step S1 specifically includes the following steps:

[0076] Based on the speech features, action and posture features, training time series features, and environmental spatial features in the correlation influence feature factors, the correlation degree relationship between the physical training results of each object in the test area is extracted to obtain the speech influence correlation relationship, posture influence correlation relationship, time series influence correlation relationship and spatial influence correlation relationship;

[0077] A comprehensive correlation assessment is conducted on the correlations of speech, posture, temporal, and spatial effects to obtain a comprehensive correlation relationship 1. The similarity value between the comprehensive correlation relationship 1 and the actual correlation relationship 1 between the physical training performance of each object in the test area is then calculated.

[0078] If the similarity value between the overall influence correlation one and the actual influence correlation one between the physical training performance of each object in the test area is less than the preset similarity judgment threshold, then the detection information will be output.

[0079] Speech recognition technology is used to analyze speech features among the influencing factors. For example, during physical training, speech information such as the trainee's shouts and breathing rhythm is identified. Speech signal processing is used to evaluate the degree of correlation between this speech information and the various performance data in the physical training results under the influence of speech features, thus obtaining the speech influence correlation.

[0080] Computer vision and posture recognition are used to process movement posture features. Posture information such as angles and trajectories of various body parts during training movements is captured. The correlation between posture information and various performance data in physical training results is analyzed under the influence of movement posture features, thus obtaining the correlation between posture and performance.

[0081] By combining training time-series characteristics, information such as training intensity and rest intervals at different time periods during training is recorded. The correlation between these time-related information and various performance data in physical fitness test values ​​is analyzed under the influence of training time-series characteristics, thus obtaining the correlation of time-series influence. For example, the impact of arranging high-intensity interval training followed by low-intensity endurance training on the final physical fitness performance can be explored.

[0082] By establishing a correlation model between environmental spatial characteristics and physical training performance, the environmental spatial characteristics are analyzed. Information such as temperature, humidity, and size of the training venue is obtained using environmental monitoring equipment. The spatial influence correlation is then assessed by evaluating the degree of correlation between these environmental factors and various performance data in physical training under the influence of these environmental spatial characteristics. For example, the impact of high-temperature environments on physical training endurance performance is analyzed.

[0083] The comprehensive correlation evaluation of the above-obtained speech influence correlation, posture influence correlation, temporal influence correlation and spatial influence correlation yields the comprehensive influence correlation one, which fully reflects the comprehensive influence of multi-dimensional features on physical training performance.

[0084] Meanwhile, the actual correlation between the physical training scores of each object in the test area has already been obtained. Similarity (such as cosine similarity) is used to calculate the similarity value between the comprehensive correlation and the actual correlation.

[0085] A preset similarity threshold is set, and the calculated similarity value is compared with this threshold. If the similarity value is less than the preset threshold, it indicates that the overall influence relationship obtained based on the analysis of various features differs significantly from the actual situation. In this case, detection information is output, suggesting that the training process needs further investigation and adjustment.

[0086] Based on the correlation factors of speech features, action and posture features, training time series features, and environmental spatial features, the correlation between the physical training scores of each object in the test area is extracted to obtain the correlation relationships of speech influence, posture influence, time series influence, and spatial influence. The specific steps include:

[0087] Based on the speech features in the correlation influence feature factors, the degree of correlation between the various performance data in the physical training performance test values ​​under the influence of speech features is evaluated to obtain the speech influence correlation relationship;

[0088] Based on the action posture features in the correlation influence feature factors, the correlation between the various performance data in the physical training performance test values ​​under the influence of action posture features is evaluated to obtain the posture influence correlation relationship.

[0089] Based on the training time series characteristics in the correlation influence feature factors, the correlation between the various performance data in the physical training performance test values ​​under the influence of the training time series characteristics is evaluated to obtain the time series influence correlation relationship;

[0090] Based on the environmental spatial characteristics in the correlation influence characteristic factors, the spatial influence correlation is obtained by evaluating the degree of correlation between the various performance data in the physical training performance test values ​​under the influence of environmental spatial characteristics.

[0091] Speech recognition is used to collect speech information during training, such as athletes' shouts and breathing sounds. This speech information is then converted into quantifiable data, such as feature values ​​like volume, frequency, and rhythm. Next, considering the influence of these speech features, the correlation between these feature values ​​and physical training performance metrics (such as running speed and strength training weight) is determined, thus establishing the relationship between speech influence and performance. For example, the correlation between the frequency of athletes shouting slogans during high-intensity training and the improvement in endurance performance can be analyzed.

[0092] Computer vision and posture recognition are used to capture athletes' movements and postures in real time during training. Postural data, such as the position, angle, and trajectory of various body parts, are extracted. Under the influence of these postural features, the correlation between this data and physical training performance is evaluated to determine the relationship between posture and performance. For example, in basketball shooting training, the correlation between posture data such as arm extension angle and body tilt during shooting and shooting accuracy is determined.

[0093] Training time-series data, including time points, changes in training intensity, and rest intervals, are recorded at different stages of the training process using time-tracking devices. Considering the influence of training time-series characteristics, the degree of correlation between these time-related data and physical fitness test values ​​is determined, thus obtaining the temporal influence correlation. For example, the correlation between speed distribution and endurance performance at different time periods in long-distance running training can be analyzed.

[0094] Environmental spatial data of the training venue, such as temperature, humidity, air pressure, venue area, and venue layout, are acquired through environmental monitoring equipment. Under conditions influenced by these environmental spatial characteristics, the correlation between this data and physical training performance is assessed to determine the spatial impact correlation. For example, this study investigates the correlation between temperature changes and a decline in athletes' endurance performance during outdoor physical training in high-temperature environments.

[0095] If the similarity value between the overall influence correlation one and the actual influence correlation one between the physical training performance of each object in the test area is less than the preset similarity judgment threshold, the detection information is output, specifically including the following steps:

[0096] An assessment of the comprehensive correlations of speech, posture, temporal, and spatial influences yields Comprehensive Influence Correlation 1.

[0097] The comparison result 1 is obtained by comparing the comprehensive impact relationship 1 and the actual impact relationship 1.

[0098] A preset similarity threshold is set. If the similarity value in the comparison result is greater than or equal to the similarity threshold, the test results are output after the physical training performance of each object in the test area is finely processed according to the physical training performance test value and the actual correlation influence.

[0099] If the similarity value in the comparison result is less than the similarity judgment threshold, then the physical training performance data of each object in the test area will be detected and the detection information will be output.

[0100] This application comprehensively processes the correlations between speech, posture, temporal, and spatial influences. By integrating the impacts of different dimensional features on physical training performance, a comprehensive influence correlation (Relationship 1) is obtained, fully reflecting the influence pattern of multiple dimensional features on physical training performance.

[0101] Similarity metrics (such as cosine similarity and Euclidean distance) are used to compare the overall influence relationship 1 and the actual influence relationship 1. The actual influence relationship 1 is the correlation between performance data obtained by directly analyzing physical training performance test values. The comparison result 1 is obtained after comparison, which quantifies the degree of similarity between the two.

[0102] A preset similarity threshold is set. If the similarity value in the comparison result is greater than or equal to the similarity threshold, it indicates that the comprehensive influence correlation constructed based on multi-factor analysis is largely consistent with the actual situation. At this point, based on the physical training performance test values ​​and the actual correlation influence, the physical training performance of each object in the test area is refined, such as by standardizing the performance and correcting errors. The test results are then output, providing a reliable basis for evaluating training effectiveness and developing subsequent training plans.

[0103] If the similarity value in the first comparison result is less than the similarity judgment threshold, it indicates that there is a significant deviation between the overall influence correlation and the actual situation. In this case, to find the cause of the deviation, the physical training performance data of each object in the test area is re-tested, the performance data is collected and analyzed again, and the test information is output to indicate that there may be a problem with the current training model of the trainee or coach.

[0104] Step S3 specifically includes the following steps:

[0105] Based on the test information, key movement segments and core training periods were selected from each subject in the test area. Among them, key movement segments refer to the combination of movements in the physical training process that have a key impact on performance, and core training periods refer to the time periods in the entire training cycle that have an impact on physical fitness improvement and performance.

[0106] Physical training data 1 was obtained by detecting physical training data of key movement segments, and physical training data 2 was obtained by detecting physical training data of core training periods.

[0107] The preprocessed physical fitness performance data is obtained by performing comprehensive statistical calculations on physical fitness training data 1 and physical fitness training data 2.

[0108] The correlation between the various performance data in the preprocessed physical fitness data was evaluated to obtain the second actual correlation influence relationship.

[0109] By comparing the performance changes of trainees under different movement segments, key movement combinations that significantly impact performance are identified, known as critical movement segments. For example, in gymnastics training, analyzing extensive training data and competition results identifies flips and somersaults as critical movement segments. On the other hand, studying the trainee's physical fitness changes and performance improvement trends throughout the entire training cycle identifies the periods with the greatest impact on fitness improvement and performance, known as core training periods. For instance, in long-distance running training, the intensive training period in the month leading up to the competition is identified as the core training period.

[0110] For specific key movement segments, real-time monitoring of physical training data such as strength, speed, and angle during the movement yields physical training data one. For example, in the squat key movement segment, data such as knee joint angle changes and leg force output are monitored. For core training periods, wearable devices (such as smart bracelets and sports watches) and training site monitoring equipment are used to collect physical training data such as the trainee's heart rate, energy consumption, and training intensity during this period, yielding physical training data two. For example, recording heart rate changes and energy consumption at different training stages during the core training period.

[0111] The physical training data (data 1 and data 2) were statistically analyzed to integrate the data from key movement segments and core training periods, resulting in preprocessed physical fitness performance data. The correlation between the various performance data within the preprocessed data was evaluated to uncover the intrinsic connections between performance data from key movement segments and core training periods, thus obtaining the second set of actual correlation and influence relationships.

[0112] The second comprehensive impact correlation of the pre-processed physical fitness performance data was assessed, specifically as follows:

[0113] The correlation between preprocessed physical fitness performance data and the physical fitness training performance of each subject in the test area under the influence of speech features, movement posture features, training time sequence features, and environmental spatial features is obtained as a comprehensive influence correlation relationship 2.

[0114] Using speech recognition equipment, motion capture systems, time recording tools (such as training log recording software), and environmental monitoring instruments (such as temperature and humidity sensors and GPS positioning devices), data are collected on the following aspects of each object in the test area: speech characteristics (such as shouts and commands during training), motion and posture characteristics (position, angle, and trajectory of different body parts during training), training time series characteristics (duration, intensity changes, and rest intervals of different training stages), and environmental spatial characteristics (temperature, humidity, size, and geographical location of the training site).

[0115] The collected multi-dimensional feature data is integrated with preprocessed physical fitness performance data to obtain the correlation between various physical fitness training performance data. For example, the correlation between the trainee's movement standardization (movement posture features) and shouting rhythm (vocal features) during a specific training period (training time series features) under different ambient temperatures (environmental spatial features) and the improvement of physical fitness training performance is explored, thus obtaining the second comprehensive influence correlation, which fully reveals the comprehensive mechanism of the multi-dimensional features on physical fitness training performance.

[0116] After comparing and judging the similarity values ​​between the comprehensive impact correlation 2 and the actual impact correlation 2, the test area is divided into local region 1 and local region 2, which specifically includes the following steps:

[0117] The comparison result 2 is obtained by comparing the comprehensive impact relationship 2 and the actual impact relationship 2;

[0118] If the similarity value in the comparison result 2 is greater than or equal to the similarity judgment threshold, then the test area is divided into local region 1 and local region 2 according to the actual association and influence relationship 2. The arrangement position of the actual association and influence relationship of each object in local region 1 is before the arrangement position of the actual association and influence relationship of each object in local region 2.

[0119] Similarity metrics (such as cosine similarity and Euclidean distance) are used to compare the second comprehensive influence correlation and the second actual influence correlation. The second comprehensive influence correlation is the correlation of physical training performance obtained after considering multi-dimensional features, while the second actual influence correlation is obtained by evaluating the correlation between various performance data in the preprocessed physical training performance data. By comparing the two, the degree of similarity between them is quantified, resulting in the second comparison result.

[0120] A preset similarity threshold is established, and the similarity value in the second comparison result is compared with this threshold. If the similarity value is greater than or equal to the similarity threshold, it indicates that the comprehensive influence relationship obtained based on multi-dimensional features is largely consistent with the actual situation. At this point, based on the actual influence relationship, the test area is divided into Local Region 1 and Local Region 2 according to the training performance. Furthermore, objects with better actual influence relationships are assigned to Local Region 1, and their positions are placed before the actual influence relationships of objects in Local Region 2, so as to facilitate subsequent targeted training analysis and resource allocation based on the characteristics of different regions.

[0121] Based on the preprocessed physical fitness performance data, the physical fitness training performance of local region one and local region two are analyzed and processed sequentially before the test results are output. The specific steps include:

[0122] Based on the preprocessed physical fitness data, the physical fitness training results of local area 1 are processed and the processed result 1 is output.

[0123] Based on the preprocessed physical fitness data and the preset interval processing time of processing result one, the physical fitness training results of local area two after the interval processing time are processed and then the processing result two is output.

[0124] The test result is formed by combining the results of processing result one and processing result two.

[0125] The goal is to uncover training patterns behind preprocessed physical fitness data, such as identifying key factors contributing to rapid performance improvement and identifying weaknesses. The output results then provide data support for adjusting and optimizing subsequent training plans for specific regions or individuals.

[0126] Based on the preprocessed physical fitness data and the processing result of local region one, and combined with the principles of sports training, considering factors such as the training fatigue recovery cycle and the time for knowledge and skill consolidation, a reasonable interval processing period is preset. After the interval processing period, the physical fitness training data of the subjects in local region two are processed, focusing on the impact of the interval period on training performance and the differences between their performance and that of the subjects in local region one. The processed result two is then output.

[0127] The results from processing results one and two are integrated to form a complete test result. During the integration process, the training performance characteristics, strengths, and weaknesses of the two local areas are compared, as well as the differences in effectiveness under different processing methods. The final test results comprehensively reflect the physical training status of the subjects within the tested area, providing a strong basis for trainers or coaches to develop scientific, reasonable, and personalized training plans.

[0128] Example 2: An AI-powered intelligent perception-based comprehensive and precise physical training system, comprising:

[0129] The statistics module retrieves the test area where physical training performance needs to be tested, and statistically analyzes the influencing factors of physical training performance testing for each object in the test area to obtain the associated influencing characteristic factors.

[0130] The processing and analysis module processes and analyzes the feature data of each feature factor in the correlation influence to obtain the comprehensive influence correlation relationship 1. It then compares and judges the comprehensive influence correlation relationship 1 with the actual correlation influence relationship 1 between the physical training performance of each object in the test area and outputs the detection information.

[0131] Select the statistics module, select key movement segments and core training periods from each object in the test area based on the detection information, detect the physical training data of key movement segments and core training periods respectively, and perform comprehensive statistics to obtain preprocessed physical fitness performance data. Statistically analyze the correlation between the performance data in the preprocessed physical fitness performance data to obtain the actual correlation and influence relationship.

[0132] The output module evaluates the comprehensive influence correlation 2 to which the preprocessed physical fitness performance data belongs. After comparing the similarity values ​​between the comprehensive influence correlation 2 and the actual influence correlation 2, it divides the test area into local region 1 and local region 2. Based on the preprocessed physical fitness performance data, it performs sequential analysis and processing of physical fitness training performance in local region 1 and local region 2, and then outputs the test results.

[0133] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0134] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0135] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An AI intelligent perception physical fitness training full-dimensional precision method, characterized in that, The method comprises the following steps: Step S1, obtaining a to-be-tested area requiring physical training performance testing, and counting physical training performance testing influence factors of each object in the to-be-tested area to obtain associated influence characteristic factors; Step S2, processing and analyzing each characteristic data in the associated influence characteristic factors to obtain a comprehensive influence association relationship, and comparing and judging between the comprehensive influence association relationship and an actual associated influence relationship one between the physical training performances of each object in the to-be-tested area to output detection information; Step S3, selecting key action segments and core training periods from each object in the to-be-tested area according to the detection information, respectively detecting physical training data of the key action segments and the core training periods and comprehensively counting to obtain preprocessed physical performance data, and counting a correlation degree relationship between each performance data in the preprocessed physical performance data to obtain an actual associated influence relationship two; Step S4, evaluating a comprehensive influence association relationship two to which the preprocessed physical performance data belongs, comparing and judging a similarity value between the comprehensive influence association relationship two and the actual associated influence relationship two, and dividing the to-be-tested area into a local area one and a local area two, and outputting a test result after respectively performing physical training performance analysis and processing on the local area one and the local area two according to the preprocessed physical performance data; Step S1 specifically comprises the following steps: Obtaining a to-be-tested area requiring physical training performance testing, and detecting physical training performance data of each object in the to-be-tested area to obtain physical training performance detection values; Evaluating a correlation degree relationship between each performance data in the physical training performance detection values to obtain an actual associated influence relationship one; Counting voice characteristics, action posture characteristics, training time sequence characteristics, and environmental space characteristics of each object in the to-be-tested area to obtain associated influence characteristic factors; Step S2 specifically comprises the following steps: Respectively extracting a correlation degree relationship between physical training performances of each object in the to-be-tested area according to voice characteristics, action posture characteristics, training time sequence characteristics, and environmental space characteristics in the associated influence characteristic factors to obtain a voice influence association relationship, a posture influence association relationship, a time sequence influence association relationship, and a space influence association relationship; Comprehensively evaluating the voice influence association relationship, the posture influence association relationship, the time sequence influence association relationship, and the space influence association relationship to obtain a comprehensive influence association relationship one, and calculating a similarity value between the comprehensive influence association relationship one and the actual associated influence relationship one between the physical training performances of each object in the to-be-tested area; If the similarity value between the comprehensive influence association relationship one and the actual associated influence relationship one between the physical training performances of each object in the to-be-tested area is less than a preset similarity threshold, detection information is output.

2. The AI intelligent perception full-dimensional precision training method according to claim 1, characterized in that, Respectively extracting a correlation degree relationship between physical training performances of each object in the to-be-tested area according to voice characteristics, action posture characteristics, training time sequence characteristics, and environmental space characteristics in the associated influence characteristic factors to obtain a voice influence association relationship, a posture influence association relationship, a time sequence influence association relationship, and a space influence association relationship, specifically comprising the following steps: According to the voice feature in the correlation influence characteristic factor, the correlation degree relationship between each performance data in the physical training performance detection value under the influence of the voice feature is evaluated to obtain a voice influence correlation relationship; According to the action posture feature in the correlation influence characteristic factor, the correlation degree relationship between each performance data in the physical training performance detection value under the influence of the action posture feature is evaluated to obtain a posture influence correlation relationship; According to the training time sequence feature in the correlation influence characteristic factor, the correlation degree relationship between each performance data in the physical training performance detection value under the influence of the training time sequence feature is evaluated to obtain a time sequence influence correlation relationship; According to the environmental space feature in the correlation influence characteristic factor, the correlation degree relationship between each performance data in the physical training performance detection value under the influence of the environmental space feature is evaluated to obtain a space influence correlation relationship.

3. The AI intelligent perception full-dimensional precision training method according to claim 2, characterized in that, If the similarity value between the comprehensive influence correlation relationship one and the actual correlation influence relationship one between each object physical training performance in the test area is less than the preset similarity threshold, the detection information is output, which specifically includes the following steps: The voice influence correlation relationship, the posture influence correlation relationship, the time sequence influence correlation relationship, and the space influence correlation relationship are evaluated to obtain a comprehensive influence correlation relationship one; The comprehensive influence correlation relationship one and the actual correlation influence relationship one are compared to obtain a comparison result one; The preset similarity threshold is set, and if the similarity value in the comparison result one is greater than or equal to the similarity threshold, the test result is output after the physical training performance of each object in the test area is processed according to the physical training performance detection value and the actual correlation influence relationship one; If the similarity value in the comparison result one is less than the similarity threshold, the physical training performance data of each object in the test area is detected, and the detection information is output.

4. The AI intelligent perception full-dimensional precision training method according to claim 1, wherein, Step S3 specifically includes the following steps: According to the detection information, a key action segment and a core training period are selected from each object in the test area; wherein the key action segment refers to an action combination segment that is critical to the performance in the physical training action process, and the core training period refers to a time period that affects the physical improvement and performance in the entire training cycle; The key action segment is detected to obtain physical training data one, and the core training period is detected to obtain physical training data two; The physical training data one and the physical training data two are comprehensively statistically operated to obtain preprocessed physical performance data; The correlation degree relationship between each performance data in the preprocessed physical performance data is evaluated to obtain an actual correlation influence relationship two.

5. The AI intelligent perception full-dimensional precision training method according to claim 4, characterized in that, The comprehensive influence correlation relationship two to which the preprocessed physical performance data belongs is evaluated, which specifically includes: The correlation relationship between each physical training performance under the influence of the voice feature, the action posture feature, the training time sequence feature, and the environmental space feature of each object in the test area to which the preprocessed physical performance data belongs is obtained to obtain the comprehensive influence correlation relationship two.

6. The AI intelligent perception full-dimensional precision training method according to claim 5, wherein, The test area is divided into the local area one and the local area two after comparing and judging the similarity degree value between the comprehensive influence correlation two and the actual influence correlation two, and specifically includes the following steps: The comprehensive influence correlation two is compared with the actual influence correlation two to obtain a comparison result two; If the similarity degree value in the comparison result two is greater than or equal to the similarity threshold value, the test area is divided into the local area one and the local area two according to the actual influence correlation two, and the arrangement positions of the actual influence correlations of the objects in the local area one are all located before the arrangement positions of the actual influence correlations of the objects in the local area two.

7. The AI intelligent perception full-dimensional precision training method according to claim 6, characterized in that, The test result is output after the local area one and the local area two are respectively processed for the physical training performance according to the preprocessed physical performance data, and specifically includes the following steps: The local area one is processed for the physical training performance according to the preprocessed physical performance data, and a processing result one is output; The local area two is processed for the physical training performance after an interval processing period according to the preprocessed physical performance data and the processing result one, and a processing result two is output; The processing result one and the processing result two are combined to form the test result.

8. An AI intelligent perception physical training full-dimensional precision system applied to the AI intelligent perception physical training full-dimensional precision method of any one of claims 1-7. It includes: An acquisition statistical module acquires a test area for physical training performance test, and statistically obtains correlation influence characteristic factors of physical training performance test influence factors of objects in the test area; A processing analysis module processes and analyzes each characteristic data in the correlation influence characteristic factors to obtain a comprehensive influence correlation one, compares and judges the comprehensive influence correlation one and an actual influence correlation one between the physical training performance of the objects in the test area, and outputs detection information; A selection statistical module selects key action segments and core training periods from the objects in the test area according to the detection information, respectively detects physical training data of the key action segments and the core training periods, and statistically obtains preprocessed physical performance data. The correlation degree relationship between each performance data in the preprocessed physical performance data is statistically obtained to obtain an actual influence correlation two; An output module evaluates a comprehensive influence correlation two to which the preprocessed physical performance data belongs, compares and judges the similarity degree value between the comprehensive influence correlation two and the actual influence correlation two, divides the test area into the local area one and the local area two, and outputs a test result after the local area one and the local area two are respectively processed for the physical training performance according to the preprocessed physical performance data.

Images