Calibration system, method, and computer program product for adjusting the seat and handlebar range of riding training device

Abstract

A calibration system, method, and computer program product for the adjustable range of a riding training device's seat and handlebars. The system uses a photographic device to capture riding images of a rider on a personal bicycle and analyzes the rider's joint positions, limb lengths, and time series of hip, knee, and ankle joint angles using human posture recognition technology. The system pre-stores at least one file of adjustable range data for the riding training device, including calibration object lengths and adjustment parameters for the seat and handlebars. By comparing the geometric relationship between the personal bicycle and the training device, this invention can perform riding posture mapping and simulate joint movement states under different device settings to find the device settings that best approximate the personal bicycle posture, and output optimized seat and handlebar adjustment suggestions. Therefore, this invention can improve the accuracy and efficiency of training device settings, and improve the rider's performance and comfort on the training device.

Claims

1. A calibration system for the adjustable range of a riding training device's seat and handlebars, which calculates the vehicle geometry and rider posture based on image data to generate corresponding device adjustment parameters. The calibration system comprises: a photographic device for recording riding images of a rider riding a bicycle, wherein the lens of the photographic device is aimed at a predetermined reference area of ​​the bicycle; an information processing device electrically connected to the photographic device, and the information processing device having a matching processing module, the matching processing module comprising: a data storage area for pre-storing at least one adjustable range data file of the riding training device, the adjustable range data file including at least: the length data of the calibration object of the riding training device, seat adjustment parameters, and handlebar adjustment parameters; a positioning unit for setting the alignment calibration of the lens of the photographic device and the bicycle; a recording unit for receiving and storing the riding images captured by the photographic device; and an image recognition unit for recognizing, processing, and analyzing the riding images to generate image feature data. A mapping unit is used to perform mapping processing based on the adjustable range data file and the image feature data pre-stored in the data storage area to generate a corresponding mapping result; and an output unit is used to output or provide corresponding information based on the mapping result, wherein the information includes at least one form of device adjustment parameters or their corresponding relationship; wherein the information processing device is further provided with a sampling processing module for establishing the adjustable range data file, the sampling processing module including: an alignment unit for setting the alignment of the lens of the camera device with a predetermined reference area of ​​the riding training equipment, so that the camera device records a riding training equipment image and transmits it to the information processing device for display; A marking unit is used to set marking data on the displayed image of the riding training equipment. The marking data includes: a calibration object in a predetermined reference area of ​​the riding training equipment, the actual length of the calibration object, and the positions of the saddle and handlebars. The image pixels are converted into actual length units by using the length of the calibration object. A parameter setting unit is used to set the adjustable range of the saddle and handlebar positions on the riding training equipment image or by numerical input. The setting content includes: saddle longitudinal adjustment data, saddle lateral adjustment data, handlebar longitudinal adjustment data, and handlebar lateral adjustment data. The setting content and the actual length units are used to create an adjustable range data file of the riding training equipment.

2. The correction system for the adjustable range of the seat and handlebars of the riding training equipment as described in claim 1, wherein, The mapping unit uses the length data of the calibrated object as a proportional reference to establish a first geometric model of the riding training equipment. The image feature data includes at least the geometric parameters of the bicycle body, and a second geometric model is established based on this. The first geometric model and the second geometric model are then compared to calculate the geometric differences between the models and establish an optimal correspondence. After calculating the geometric differences between the models, the parameter values ​​of the first geometric model are corrected according to the optimal correspondence to form an adjusted geometric model corresponding to the second geometric model, which is the mapping result.

3. The correction system for the adjustable range of the seat and handlebars of the riding training equipment as described in claim 1, wherein, The image feature data includes at least human posture parameters and motion feature vectors transformed from these human posture parameters through signal processing and dimensionality reduction. The motion feature vectors contain the main motion features used to characterize riding movements. The human posture parameters include the rider's joint positions and the time series of hip, knee, and ankle joint angles. The mapping unit simulates the rider's joint movements under different parameter combinations based on the adjustable range data file, generates simulated feature vectors using a time series learning model, compares them with the motion feature vectors, and selects the closest parameter combination as the mapping result under set feature selection and boundary conditions.

4. The correction system for the adjustable range of the seat and handlebars of the riding training equipment as described in claim 3, wherein the image recognition unit employs a deep learning-based human pose estimation algorithm to identify at least one set of joint positions in the rider image, and the algorithm is selected from deep learning-based human pose estimation technology.

5. The correction system for the adjustable range of the riding training equipment seat and handlebars as described in claim 4, wherein the deep learning human pose estimation technology includes: OpenPose, MediaPipe Pose, AlphaPose, or HRNet.

6. The correction system for adjustable range of seat and handlebars of the riding training equipment as described in claim 3, wherein the image recognition unit includes a deep learning architecture comprising a convolutional neural network (CNN), a high-resolution network (HRNet), or an equivalent variation thereof, to improve the accuracy of hip, knee, and ankle joint detection of the rider.

7. The correction system for the adjustable range of the seat and handlebars of the riding training equipment as described in claim 3, wherein, The image recognition unit includes a signal smoothing module, which includes a Kalman filter, spline interpolation, a low-pass filter, or equivalent variations thereof, to remove sensing noise during the image recognition process and smooth the time-series data.

8. The correction system for the adjustable range of the seat and handlebars of the riding training equipment as described in claim 3, wherein, The image recognition unit further calculates the time series of the rider's limb length, hip joint angle, knee joint angle and ankle joint angle, and through extreme value analysis, finds the time points when the maximum and minimum joint angles occur, so as to establish complete angle time series feature data for subsequent processing or conversion into motion feature vectors.

9. The correction system for the adjustable range of the seat and handlebars of the riding training equipment as described in claim 8, wherein, The image recognition unit further includes a principal component analysis (PCA) module, which is used to reduce the dimensionality of the high-dimensional time series data of hip, knee and ankle angles and calculate the correlation of changes among the three to generate motion feature vectors, which serve as the basis for comparison by the mapping unit.

10. The correction system for the adjustable range of the seat and handlebars of the riding training equipment as described in claim 3, wherein, The recording unit stores motion images of at least one complete pedaling cycle of the rider, and the image recognition unit uses a Kalman filter to smooth the data in order to reduce noise and ensure the stability of the joint angle time series, and uses interpolation to align the data to further improve the comparison accuracy.

11. A correction system for the adjustable range of the seat and handlebars of a riding training device as described in claims 1, 2, or 3, wherein, The predetermined reference area is the bottom bracket position of the vehicle body, and the calibrated object is the crank.

12. The correction system for the adjustable range of the seat and handlebars of the riding training equipment as described in claim 1, wherein, The output unit outputs suggested values ​​for seat adjustment parameters and handlebar adjustment parameters.

13. The correction system for the adjustable range of the seat and handlebars of the riding training equipment as described in claim 1, wherein, The saddle adjustment parameters and handlebar adjustment parameters are stored in a data structure, which includes a matrix structure. The saddle adjustment parameter matrix is ​​formed based on the longitudinal and lateral adjustment data of the saddle, and the handlebar adjustment parameter matrix is ​​formed based on the longitudinal and lateral adjustment data of the handlebar. The index position of each data structure corresponds to the settable physical position of the saddle or handlebar on the riding training equipment.

14. A method for calibrating the adjustable range of a riding training device seat and handlebars, executed by a calibration system for the adjustable range of a riding training device seat and handlebars, the calibration method calculating the vehicle geometry and rider posture based on image data to generate corresponding device adjustment parameters, the calibration method comprising at least the following processing steps: The calibration system pre-stores at least one adjustable range data file of the riding training device, the adjustable range data file comprising at least: the length data of the calibration object of the riding training device, the seat adjustment parameters, and the handlebar adjustment parameters; The calibration system uses a camera device to record a riding image of a rider riding a bicycle, wherein the lens of the camera device is aimed at a predetermined reference area of ​​the bicycle; The calibration system sets the alignment calibration between the lens of the camera device and the bicycle to receive and store the riding image captured by the camera device, and performs identification processing and analysis on the riding image to generate image feature data; The correction system performs mapping processing based on a pre-stored adjustable range data file and the image feature data to generate a corresponding mapping result; and the correction system outputs or provides corresponding information based on the mapping result, wherein the information includes at least one form of device adjustment parameters or their corresponding relationships; wherein... The steps for creating the adjustable range data file include the following processing: The calibration system aligns the lens of the camera device with a predetermined reference area of ​​the riding training equipment, causing the camera device to record an image of the riding training equipment; The calibration system displays the riding training equipment image and sets marking data on the riding training equipment image, the marking data including: a calibration object of the predetermined reference area of ​​the riding training equipment, the actual length of the calibration object, and the positions of the saddle and handlebars. By using the length of the calibration object, the calibration system converts image pixels into actual length units; and The calibration system sets the adjustable range of the saddle and handlebar positions on the riding training equipment image or by numerical input, the setting including: saddle longitudinal adjustment data, saddle lateral adjustment data, handlebar longitudinal adjustment data, and handlebar lateral adjustment data. The calibration system then creates an adjustable range data file for the riding training equipment based on the setting content and the actual length units.

15. The method for calibrating the adjustable range of the seat and handlebars of the riding training equipment as described in claim 14, wherein, The calibration system performs mapping processing by using the length data of the calibrated object as a proportional reference to establish a first geometric model of the riding training equipment. The image feature data includes at least the geometric parameters of the bicycle body, and a second geometric model is established based on this. The first geometric model and the second geometric model are then compared to calculate the geometric differences between the models and establish an optimal correspondence. After the calibration system calculates the geometric differences between the models, it corrects the parameter values ​​of the first geometric model according to the optimal correspondence to form an adjusted geometric model corresponding to the second geometric model, which is the mapping result.

16. The method for calibrating the adjustable range of the seat and handlebars of the riding training equipment as described in claim 14, wherein, The image feature data includes at least human posture parameters and motion feature vectors transformed from these human posture parameters through signal processing and dimensionality reduction. The motion feature vectors contain the main motion features used to characterize riding movements. The human posture parameters include the rider's joint positions and the time series of hip, knee, and ankle joint angles. The correction system performs mapping processing by simulating rider joint movements under different parameter combinations based on the adjustable range data file, and generates simulated feature vectors using a time series learning model. The correction system compares these motion feature vectors with the data and selects the closest parameter combination as the mapping result under set feature selection and boundary conditions.

17. A method for calibrating the adjustable range of the seat and handlebars of a riding training device as described in claims 14, 15, or 16, wherein, The saddle adjustment parameters and handlebar adjustment parameters are stored in a data structure, which includes a matrix structure. The saddle adjustment parameter matrix is ​​formed based on the longitudinal and lateral adjustment data of the saddle, and the handlebar adjustment parameter matrix is ​​formed based on the longitudinal and lateral adjustment data of the handlebar. The index position of each data structure corresponds to the settable physical position of the saddle or handlebar on the riding training equipment.

18. A method for calibrating the adjustable range of the seat and handlebars of a riding training device as described in claims 14, 15, or 16, wherein, The predetermined reference area is the bottom bracket position of the vehicle body, and the calibrated object is the crank.

19. A computer-readable storage medium storing program instructions that, when executed by a processor, cause the processor to perform the processing steps described in the calibration method for the adjustable range of the seat and handlebars of the riding training equipment as described in claims 14, 15, or 16.

20. The computer-readable storage medium as described in claim 19, wherein, The saddle adjustment parameters and handlebar adjustment parameters are stored in a data structure, which includes a matrix structure. The saddle adjustment parameter matrix is ​​formed based on the longitudinal and lateral adjustment data of the saddle, and the handlebar adjustment parameter matrix is ​​formed based on the longitudinal and lateral adjustment data of the handlebar. The index position of each data structure corresponds to the settable physical position of the saddle or handlebar on the riding training equipment.

21. The computer-readable storage medium as described in claim 19, wherein, The predetermined reference area is the bottom bracket position of the vehicle body, and the calibrated object is the crank.

22. A computer program product comprising program instructions that, when loaded and executed by a computer, cause the computer to perform the processing steps described in the calibration method for the adjustable range of the seat and handlebars of the riding training equipment as described in claims 14, 15, or 16.

23. The computer program product as described in claim 22, wherein, The saddle adjustment parameters and handlebar adjustment parameters are stored in a data structure, which includes a matrix structure. The saddle adjustment parameter matrix is ​​formed based on the longitudinal and lateral adjustment data of the saddle, and the handlebar adjustment parameter matrix is ​​formed based on the longitudinal and lateral adjustment data of the handlebar. The index position of each data structure corresponds to the settable physical position of the saddle or handlebar on the riding training equipment.

24. The computer program product as described in claim 22, wherein, The predetermined reference area is the bottom bracket position of the vehicle body, and the calibrated object is the crank.

Images