A sports racing timing system based on multi-sensor fusion
Through the multi-sensor fusion system, the main visual and specular reflection sensing array modules are used to reconstruct the spatial position and time point of the player's line blinking moment, solving the problem of misjudgment in traditional timing methods under occlusion and environmental conditions, and achieving higher accuracy and fair racing timing.
Patent Information
- Application Number
- CN202510434146.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-08
- Publication Date
- 2025-07-18
- Estimated Expiration
- 2045-04-08
AI Technical Summary
Traditional sports racing timing methods are prone to visual blind spots when players intensively cross the line, block the finish line or poor environmental conditions, resulting in inaccurate timing and affecting the fairness of the competition.
The multi-sensor fusion system is adopted, combined with the main visual acquisition module and the specular reflection sensing array module, through the interference pulse positioning mechanism and the multi-path image fusion strategy, the spatial position and time point of the player's line bounce moment is reconstructed, and the lateral perception path is constructed using the specular reflection sensing array module to supplement the optical information of the occlusion area, and the space-time mapping and decision-making logic judgment are performed through the data integration module.
It effectively eliminates the risk of misjudgment in blind spots in traditional equipment perspectives, improves the accuracy and consistency of timing results, and ensures the fairness of the competition.
Smart Images

Figure CN120126230B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of electronic timing, and more specifically, to a sports racing timing system based on multi-sensor fusion. Background Art
[0002] In the field of sports racing, especially in high-precision timing events such as sprints and skiing, traditional timing methods mainly rely on a single visual acquisition device to directly capture the finish line image, and then manually or semi-automatically identify the time when the athlete reaches the finish line. However, in the case of dense finishes, end point occlusion, or poor environmental conditions, such as rainy or snowy weather and insufficient light, this traditional method is prone to visual blind spots, resulting in the inability to accurately capture the moment when the athlete crosses the finish line, seriously affecting the accuracy and fairness of the competition results. Summary of the Invention
[0003] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a sports racing timing system based on multi-sensor fusion. By jointly collecting image information of the finish line area through the main visual acquisition module and the specular reflection sensor array module, and combining the interference pulse positioning mechanism and the multi-path image fusion strategy, the spatial position and time point of the athlete's finish moment are reconstructed to solve the problems raised in the above background art.
[0004] To achieve the above object, the present invention provides the following technical solution: A sports racing timing system based on multi-sensor fusion, including a main visual acquisition module, a time synchronization control module, a data integration module, a specular reflection sensor array module, and a reflection path reconstruction module;
[0005] The main visual acquisition module is used to record the image sequence directly in front of the athlete's finish line area at a fixed frame rate, and extract key frames from the image sequence to generate a time judgment basis for the athlete's preliminary finish.
[0006] The time synchronization control module distributes a unified clock pulse signal to all modules depending on time stamps through the crystal oscillator time base, so that the timing data recorded by each module is based on the same time reference frame.
[0007] The data integration module is used to perform time domain alignment, spatial mapping unification, and final finish action decision logic judgment on the image information generated by the main visual acquisition module and the reflection path reconstruction module.
[0008] The specular reflection sensor array module constructs a reflection path from the finish line area to the lateral sensing path by arranging multiple reflection mirror units at fixed angles, and is used to guide the light that is not captured by the main visual acquisition module due to occlusion to the visible area, and perform optical completion of the occlusion area.
[0009] The reflection path reconstruction module is used to reconstruct the reflection mapping position of the finish line action in space according to the corresponding time difference information in the reflection path formed by the specular reflection sensing array module and the layout geometric parameters of the specular units, and transmit this position to the data integration module as an auxiliary judgment basis.
[0010] In a preferred embodiment, it further includes an end region calibration module and a result output module;
[0011] The end region calibration module sets the position of the end recognition plane in the three-dimensional space coordinate system based on laser ranging or manual calibration means, and uses this position as the layout reference basis for the specular reflection sensing array module and the main vision acquisition module to align the reflection path and the main view path in space;
[0012] The result output module is used to convert the finish line time judgment result generated by the data integration module into a score data format and transmit it to an external display terminal or a competition control platform through an interface.
[0013] In a preferred embodiment, it further includes an interference pulse emission module, an interference pulse reception module, and a sub-time domain correction module;
[0014] The interference pulse emission module is used to drive a high-frequency laser diode to emit an interference pulse signal with a coding structure. The interference pulse signals are emitted at a constant interval and used as an optical trigger source for time domain positioning for the time difference backtracking process of the reflection path;
[0015] The interference pulse reception module consists of an optoelectronic sensor array and is used to receive each interference pulse signal returned from the reflection path and record its return timestamp. The return timestamp is time-aligned with the unified clock pulse provided by the time synchronization control module;
[0016] The sub-time domain correction module is used to perform dynamic calibration according to the image time point data output by the reflection path reconstruction module and the time reference frame provided by the time synchronization control module, and map the reconstructed image time points back to the unified main vision frame sequence timeline to compensate for the time domain perturbation error caused by the difference in the propagation length of the reflection path.
[0017] In a preferred embodiment, the reflection path reconstruction module is used to reconstruct the path length between the contestant's body position and the reflection mirror to deduce the propagation time of the interference pulse signal; a reflection path perturbation integral model is constructed by the reflection path reconstruction module;
[0018] The reflection path perturbation integral model performs path mapping fitting on the reflection paths formed by each specular unit through integration, and finally outputs the effective propagation length of the reflection path based on the local reflection angle perturbation and the material-induced asymmetric response, as the input basis for the sub-time domain correction module and the data integration module;
[0019] The reflection path perturbation integral model is expressed as: ;in For the The effective reflection path length value corresponding to each reflective mirror unit; Represents the spatial position variable on the reflection path; Represents the actual propagation direction vector on the reflection path; represents the linear propagation direction vector of the reflection path under ideal conditions; is the square of the norm; Indicated in On the reflection path corresponding to the reflective mirror unit, the spatial position is The reflection disturbance gain factor at the location point; represents the starting point coordinates of the reflection path integral; Represents the coordinates of the end point of the reflection path integral.
[0020] In a preferred embodiment, the sub-time domain correction module is used to dynamically correct the image time point generated by the reflection path so that it is synchronized with the time line of the main visual frame. A nonlinear function model composed of error factors is constructed by the sub-time domain correction module, which is combined with the main clock reference to perform mapping correction on the original time; the time correction mapping function is expressed as:
[0021] ;
[0022] in For the The final synchronization time point corresponding to the reflection path; is the reflection timestamp originally obtained by the interference pulse receiving module; represents a nonlinear error compensation function consisting of three error quantities; is the delay error caused by the direction deviation of the i-th path; is the system-level clock drift error of the i-th path; represents the quantization error caused by the temporal sampling resolution; Represents the unpredictable disturbance term caused by comprehensive thermal noise, device response jitter, and reflection difference;
[0023] ;
[0024] in is the hyperbolic tangent function; It represents the square of the delay error caused by the path direction deviation angle in the i-th reflection path; It represents the cube of the nonlinear sampling error caused by the quantization accuracy limitation in the time sampling process of the i-th path; is the natural logarithm function; is the square of the drift deviation between the i-th path and the master clock.
[0025] In a preferred embodiment, the interference pulse receiving module receives and processes the continuous interference pulse signal returned by the reflection path, and extracts the echo return time point with stable characteristics by detecting its time-domain change structure; this time point is used as timestamp information and transmitted to the reflection path reconstruction module to establish a time difference model of the optical propagation path and construct a corresponding spatial reflection mapping relationship;
[0026] The extraction formula of the echo return time point based on the interference pulse received signal is expressed as:
[0027] ;
[0028] where is the time point when the -th pulse signal in the interference pulse sequence is received; represents an arbitrary time variable on the continuous time axis; is the continuous voltage signal stream output by the interference pulse pair receiving module; is the modulation recognition kernel function used for the -th pulse recognition; represents the inflection point of the detection integral curve.
[0029] In a preferred embodiment, the data integration module is used to fuse the main visual image frame and the reflection path image frame. By comparing the gradient information of the image boundary space, it is judged whether the player has completed the finish line action. The objective function is constructed by the normal difference of the boundary tensor field, and the extreme value search method is used to determine the finish line moment. The objective function is the boundary transition matching extreme value function, which is expressed as:
[0030] ;
[0031] where represents the image frame number judged to have the finish line action occurred; is the -th frame image recorded by the main visual acquisition module; represents the image frame corrected by the reflection path reconstruction module; is the image edge scene quantity extraction function; is the normal gradient operator of the tensor; represents the square of the norm; is the image time series confidence evaluation function.
[0032] In a preferred embodiment, the end - region calibration module establishes a standard space coordinate system through preset space calibration points in the end region to define the accurate position of the end plane; based on three space calibration points with known positions, a spatial unit normal vector of the end plane is constructed and mapped into the respective spatial viewing angles of the mirror - reflection sensing array module and the main - vision acquisition module, thereby achieving the spatial geometric alignment between the reflection path and the main - vision path; this spatial geometric alignment relationship provides a unified spatial - coordinate reference benchmark for the reflection - path reconstruction module and the data - integration module;
[0033] ;
[0034] where is the first space calibration point selected by the end - region calibration module in the end region; is the second space calibration point selected by the end - region calibration module in the end region; is the third space calibration point selected by the end - region calibration module in the end region; represents the direction vector formed from the space calibration point pointing to the space calibration point and is used to construct one side of the end plane; represents the direction vector formed from the space calibration point pointing to the space calibration point and is used to construct the other side of the end plane; represents the vector cross - product operation; is the spatial unit normal vector of the end plane; is the position of the th mirror - reflection unit; represents the affine - transformation function that projects the plane onto the image space along the mirror path, and the output is the projection image coordinate system under the reflection path; is the projection of the end plane under the viewing angle of the
[0035] Technical effects and advantages of the present invention:
[0036] Through the mirror - reflection sensing array module and the reflection - path reconstruction module, when the viewing angle of the main - vision acquisition module is blocked, the present invention effectively supplements the blocked area through the lateral reflection path, and reconstructs the position and time of the athlete crossing the finish line through time - domain and spatial mapping, fundamentally eliminating the misjudgment risk caused by the viewing - angle blind area of traditional devices and improving the accuracy of the timing result;
[0037] Through the non-linear error compensation model of the sub-time domain correction module, the time sampling error and device drift error between different paths are eliminated, realizing accurate time domain synchronization of multiple paths and ensuring the timing consistency of path information;
[0038] Through the data integration module, spatial gradient difference matching analysis is performed on the two image information of the main vision and the reflection path, and the boundary transition tensor objective function is constructed, improving the objectivity of the finish line moment judgment;
[0039] By combining the interference pulse emission module and the interference pulse reception module, the measurement of the path propagation time is realized with high-frequency laser interference pulse signals, providing a stable time reference for the reflection path reconstruction;
[0040] By presetting spatial calibration points through the end area calibration module and establishing a unified end space coordinate reference system, the spatial geometry consistency between the reflection path and the main vision path is effectively realized, ensuring the accuracy and stability of the overall spatial mapping of the system. Description of the Drawings
[0041] Figure 1 It is a schematic diagram of the system module of the present invention. Detailed Embodiments
[0042] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.
[0043] Refer to the attached drawings of the specification Figure 1 A sports racing timing system based on multi-sensor fusion according to an embodiment of the present invention includes a main vision acquisition module, a time synchronization control module, a data integration module, a specular reflection sensor array module, and a reflection path reconstruction module;
[0044] The main vision acquisition module is used to record the image sequence directly in front of the finish line area of the athlete at a fixed frame rate and extract key frames from the image sequence to generate the time judgment basis for the athlete's preliminary finish line;
[0045] The time synchronization control module distributes a unified clock pulse signal to all modules that depend on time stamps through the crystal oscillator time base, so that the timing data recorded by each module is based on the same time reference frame; in this solution, the modules that depend on time stamps include but are not limited to: the main vision acquisition module, the reflection path reconstruction module, the interference pulse emission module, the interference pulse reception module, and the sub-time domain correction module;
[0046] The data integration module is used to perform time-domain alignment, spatial mapping unification, and final finish-line action decision logic judgment on the image information generated by the main vision acquisition module and the reflection path reconstruction module;
[0047] The specular reflection sensing array module constructs a reflection path from the finish area to the lateral sensing path by arranging multiple reflection mirror units at fixed angles, and is used to guide the light that is not captured by the main vision acquisition module due to occlusion to the visible area, and perform optical completion of the occlusion area; In practical applications, the reflection mirror unit includes an optical glass lens or a metal mirror assembly with a high-reflectivity coating, a fixed bracket for maintaining the stability of the incident angle and the reflection angle, and an angle adjustment structure for adjusting the mirror orientation. The reflection mirror unit is used to construct a stable and controllable reflection path in the finish area to guide the incident light in the occlusion direction to the lateral sensing path;
[0048] The reflection path reconstruction module is used to reconstruct the reflection mapping position of the finish-line action in space according to the corresponding time difference information in the reflection path formed by the specular reflection sensing array module and the layout geometric parameters of the mirror unit, and transmit this position to the data integration module as an auxiliary judgment basis.
[0049] It also includes a finish area calibration module and a result output module;
[0050] The finish area calibration module sets the position of the finish recognition plane in the three-dimensional space coordinate system based on laser ranging or manual calibration means, and uses this position as the layout reference basis for the specular reflection sensing array module and the main vision acquisition module to align the space of the reflection path and the main view path;
[0051] The result output module is used to convert the finish time judgment result generated by the data integration module into a score data format and transmit it to an external display terminal or a competition control platform through an interface.
[0052] It also includes an interference pulse emission module, an interference pulse reception module, and a sub-time domain correction module;
[0053] The interference pulse emission module is used to drive a high-frequency laser diode to emit an interference pulse signal with a coding structure. The interference pulse signal is emitted at a constant interval and is used as an optical trigger source for time-domain positioning for use in the time difference inverse derivation process of the reflection path;
[0054] The interference pulse reception module is composed of an optoelectronic sensor array and is used to receive each interference pulse signal returned from the reflection path and record its return timestamp. The return timestamp is time-aligned with the unified clock pulse provided by the time synchronization control module;
[0055] The sub-time domain calibration module is used to perform dynamic calibration based on the image time point data output by the reflection path reconstruction module and the time reference frame provided by the time synchronization control module, and map the reconstructed image time points back to the unified main visual frame sequence timeline to compensate for the time domain perturbation error caused by the difference in the propagation length of the reflection path.
[0056] The reflection path reconstruction module is used to reconstruct the path length between the player's body position and the reflection mirror surface, and is used to deduce the propagation time of the interference pulse signal; a reflection path perturbation integral model is constructed through the reflection path reconstruction module;
[0057] Traditional geometric modeling does not consider the influence of the reflection surface perturbation and the mirror microstructure. The reflection path perturbation integral model performs path mapping fitting on the reflection paths formed by each reflection mirror unit through integration, and at the same time, based on the local reflection angle perturbation and the material-induced asymmetric response, finally outputs the effective propagation length of the reflection path as the input basis for the sub-time domain calibration module and the data integration module;
[0058] The reflection path perturbation integral model is expressed as: ; where is the effective reflection path length value corresponding to the th reflection mirror unit. In practical applications, the effective reflection path length value represents the light propagation path that can still be measured after the reflection path has perturbations in reality; represents the spatial position variable on the reflection path, the value of represents the path point of the reflection path from the player's position (denoted as ) to the position of the interference pulse receiving module (denoted as ), which is used to describe the continuous change of the reflection perturbation with the spatial position on this path segment, represents a small spatial position increment in the reflection path, which is used for the integration operation; represents the starting coordinate of the reflection path integration; represents the ending coordinate of the reflection path integration; represents the actual propagation direction vector on the reflection path. Due to factors such as mirror micro-perturbation, incident angle drift, and uneven material reflectivity, so will have curvature; represents the linear propagation direction vector of the reflection path under ideal conditions, which is used as the reference direction; In the formula, is the square of the norm, which represents the energy of the direction error; represents the reflection perturbation gain factor at the position point with the spatial position of on the reflection path corresponding to the th reflection mirror unit, It is used to measure the contribution of the path point to the overall disturbance degree of the reflection direction. It is used as a weight term in the integral model and has spatial variability and nonlinear amplitude adjustment capabilities.
[0059] ;
[0060] in is the structural control coefficient in the disturbance function; Used to adjust the reflection direction deviation caused by the overall geometric non-uniformity of the path, Used to adjust the periodic reflection disturbance caused by the micro-texture structure on the mirror surface. Used to adjust the directional jump effect caused by the sudden change of local reflection phase. Used to adjust the attenuation response of the mirror center energy focusing and edge interference suppression; Indicates The excitation frequency factor of the periodic microstructure on the surface of the reflective mirror unit, It is used to model the periodic modulation effect of the mirror surface texture on the reflection direction. The higher the frequency of the spatial periodic change of the mirror surface micro-texture, the more drastic the change of the reflection direction along the path. Indicates The reflective mirror unit is located at the path position The phase perturbation function at , Reflects the reflection phase changes caused by factors such as mirror material, thickness fluctuations, and temperature fields; is the phase perturbation function at the path position The first-order derivative at represents the local reflection phase mutation rate, that is, the phase jump speed of the reflected light at this point in the path direction. In addition, the square term of this derivative is used to construct the energy contribution of the perturbation at this point to the directional stability; represents an exponential boundary suppression function, which is used to reduce the dominant influence of the two ends of the reflection path, i.e., the start and end points, on the overall disturbance evaluation; the coefficient is the mirror energy focusing factor, coefficient It is used to determine the concentration of the contribution of the center area of the mirror to the reflection. Its physical meaning is: the center of the mirror has the greatest influence on the reflection direction, and the edge influence decays rapidly.
[0061] The sub-time domain correction module is used to dynamically correct the image time point generated by the reflection path to keep it synchronized with the timeline of the main visual frame. Due to different paths, different sampling precisions and hardware delays, the two paths will produce time domain offsets. Therefore, a nonlinear function model composed of error factors is constructed through the sub-time domain correction module, which is combined with the main clock reference to perform mapping correction on the original time; the time correction mapping function is expressed as:
[0062] ;
[0063] wherein is the final synchronization time point corresponding to the th reflection path, and is used to output to the data integration module; is the reflection timestamp obtained by the original interference pulse receiving module, which is uncorrected; represents a non - linear error compensation function composed of three error amounts; is the delay error caused by the direction deviation of the th path; is the system - level clock drift error of the th path;
[0064] ;
[0065] wherein is the hyperbolic tangent function, which is used to non - linearly compress the input composite disturbance term so that the high - amplitude error will not get out of control during the correction process; represents the square of the time - delay error caused by the path direction deviation angle in the th reflection path, which is used to reflect the influence intensity of the path attitude change on time synchronization; represents the cube of the non - linear sampling error generated by the quantization accuracy limitation during the time sampling process of the th path, which is used to reflect its progressive enhancement effect on time drift; in the formula, the sum of and is used as the input term, representing the composite coupling influence of two disturbance sources, and then the tanh function is used to limit its non - linear change amplitude; is the natural logarithm function, and the natural logarithm function in the formula is used to control the system - level clock drift error within a limited range to make its influence on the final compensation smoother; is the square of the drift deviation between the th path and the master clock,
[0066] The interference pulse receiving module receives and processes the continuous interference pulse signal returned by the reflection path, and extracts the echo return time point with stable characteristics by detecting its time - domain change structure; this time point is used as the timestamp information and transmitted to the reflection path reconstruction module to establish the time - difference model of the optical propagation path and construct the corresponding spatial reflection mapping relationship;
[0067] The formula for extracting the echo return time point based on the interference pulse receiving signal is expressed as:
[0068] ;
[0069] where is the time point when the -th pulse signal in the interference pulse sequence is received; represents an arbitrary moment variable on the continuous time axis; is the continuous voltage signal stream output by the interference pulse pair receiving module, which represents the real echo; is used for the -th pulse identification modulation identification kernel function, which is determined according to the transmission structure; the time differential element is used to integrate the historical echo and smooth the structure waveform; represents the inflection point of the detection integration curve, which corresponds to the change in the echo slope. The inflection point is considered the main response of the reflected echo; in addition, the condition "=0" in the formula means to find the extreme point of the slope curve.
[0070] The data integration module is used to fuse the main visual image frame and the reflected path image frame. By comparing the gradient information of the image boundary space, it judges whether the player has completed the finish line crossing action. It constructs an objective function through the normal difference of the boundary tensor field and uses the extreme value search method to determine the finish line crossing moment. The objective function is the boundary transition matching extreme value function, which is expressed as:
[0071]
[0072] where represents the image frame number judged to have the finish line crossing action occurred; is the -th frame image recorded by the main visual acquisition module; represents the image frame corrected by the reflected path reconstruction module; is the image edge scene quantity extraction function, which is used to extract the object boundary in the image; is the normal gradient operator of the tensor. The normal gradient operator of the tensor represents the boundary change direction; In the formula, used as the square of the norm represents the square difference between two tensor fields, which is used to measure the sudden change of the image structure; is the image time series confidence evaluation function, used to filter out visual errors; represents the variable that solves to make the expression reach the maximum value.
[0073] The end - point area calibration module establishes a standard spatial coordinate system through preset spatial calibration points in the end - point area, defining the accurate position of the end - point plane. Based on three spatial calibration points with known positions, a spatial unit normal vector of the end - point plane is constructed, and this normal vector is respectively mapped into the corresponding spatial viewing angles of the mirror - reflection sensing array module and the main vision acquisition module, thereby achieving the spatial geometric alignment between the reflection path and the main vision path. This spatial geometric alignment relationship provides a unified spatial coordinate reference benchmark for the reflection path reconstruction module and the data integration module, ensuring the accuracy of the spatial coordinate information when finally judging the athlete's crossing - the - line action.
[0074]
[0075] Among them is the first spatial calibration point selected by the end - point area calibration module in the end - point area. Its position coordinates are determined by a laser rangefinder or manual measurement method, and are used to construct the initial position reference of the end - point recognition plane. is the second spatial calibration point selected by the end - point area calibration module in the end - point area, which is also obtained by laser ranging or manual calibration. Together with point they jointly constitute the first spatial direction reference of the plane. is the third spatial calibration point selected by the end - point area calibration module in the end - point area. Through this point and points and the spatial orientation of the end - point recognition plane is determined, and a spatial normal vector is determined by the three points together, forming the geometric spatial benchmark for end - point determination. represents the direction vector formed from the spatial calibration point pointing to the spatial calibration point and is used to construct one side of the end - point plane. represents the direction vector formed from the spatial calibration point pointing to the spatial calibration point and is used to construct the other side of the end - point plane. represents the vector cross - product operation to generate the normal vector. In the formula, the part represents the vector norm. is the spatial unit normal vector of the end - point plane. is the position of the th mirror - reflection unit. represents the affine transformation function that projects the plane onto the mirror path into the image space, and the output is the projection image coordinate system under the reflection path. is the projection of the end - point plane under the viewing angle corresponding to the th mirror unit in the mirror - reflection sensing array module.
[0076] For the present invention, it should be further noted that the main vision acquisition module first captures the image directly in front of the finish line area of the athlete at a fixed frame rate and extracts the key frames as the basis for preliminary finish line judgment. The time synchronization control module synchronizes a unified clock signal to all modules related to time stamps through the crystal oscillator time base to ensure the consistency of the data acquisition timing;
[0077] When the view angle of the main vision acquisition module is blocked, multiple reflection mirror units at fixed angles in the specular reflection sensing array module guide the light in the blocked area to the lateral path to form a reflection path; the reflection path reconstruction module reconstructs the reflection mapping position of the athlete at the finish line moment according to the time difference of the specular reflection path and the geometric parameters of the mirror unit and provides it to the data integration module;
[0078] The interference pulse emission module generates high-frequency coded laser interference pulses. The interference pulse receiving module receives and records the returned pulse signals. The sub-time domain correction module dynamically corrects the time difference generated by the propagation of the reflection path to synchronize the image time point of the reflection path with the main vision timeline; the data integration module fuses the main vision and reflection path information and determines the final finish line frame of the athlete through the extreme value function of the image space boundary feature difference; the end area calibration module establishes a space coordinate system through the preset space calibration points to ensure the spatial alignment of the reflection path and the main vision path, and finally outputs the result, converts the determined finish line time into a standard format, and outputs it to the external display terminal or the competition control platform;
[0079] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.
Claims
1. A sports racing timing system based on multi-sensor fusion, comprising a main vision acquisition module, a time synchronization control module, a data integration module, a specular reflection sensing array module, and a reflection path reconstruction module, characterized in that: The main vision acquisition module is used to record the image sequence directly in front of the finish line area of the contestants at a fixed frame rate, and extract key frames from the image sequence to generate a time judgment basis for the contestants' preliminary crossing of the finish line; The time synchronization control module distributes a unified clock pulse signal to all modules relying on timestamps through a crystal oscillator time base, so that the timing data recorded by each module is based on the same time reference frame; The data integration module is used to perform time domain alignment, spatial mapping unification, and final finish line action decision logic judgment on the image information generated by the main vision acquisition module and the reflection path reconstruction module; The specular reflection sensing array module constructs a reflection path from the finish line area to the lateral sensing path by arranging multiple specular reflection units at fixed angles, and is used to guide the light that is not captured by the main vision acquisition module due to occlusion to the visible area to perform optical complementation of the occlusion area; The reflection path reconstruction module is used to reconstruct the reflection mapping position of the finish line action in space according to the time difference information corresponding to the reflection path formed by the specular reflection sensing array module and the layout geometric parameters of the specular units, and transmit this position to the data integration module as an auxiliary judgment basis.
2. A sports racing timing system based on multi-sensor fusion according to claim 1, characterized in that: It further includes a finish line area calibration module and a result output module; The finish line area calibration module sets the position of the finish line recognition plane in the three-dimensional space coordinate system based on laser ranging or manual calibration means, and uses this position as the layout reference basis for the specular reflection sensing array module and the main vision acquisition module to align the space of the reflection path and the main view path; The result output module is used to convert the finish line time judgment result generated by the data integration module into a score data format and transmit it to an external display terminal or a competition control platform through an interface.
3. A sports racing timing system based on multi-sensor fusion according to claim 2, characterized in that: It further includes an interference pulse emission module, an interference pulse reception module, and a sub-time domain correction module; The interference pulse emission module is used to drive a high-frequency laser diode to emit an interference pulse signal with a coding structure. The interference pulse signal is emitted at a constant interval and is used as an optical trigger source for time domain positioning for the time difference inversion process of the reflection path; The interference pulse reception module is composed of a photoelectric sensor array, and is used to receive each interference pulse signal returned from the reflection path and record its return timestamp. The return timestamp is time-aligned with the unified clock pulse provided by the time synchronization control module; The sub-time domain correction module is used to perform dynamic calibration according to the image time point data output by the reflection path reconstruction module and the time reference frame provided by the time synchronization control module, and map the reconstructed image time point back to the unified main vision frame sequence timeline to compensate for the time domain perturbation error caused by the difference in the propagation length of the reflection path.
4. A sports racing timing system based on multi-sensor fusion according to claim 3, characterized in that: The reflection path reconstruction module is used to reconstruct the path length between the contestant's body position and the reflection mirror surface, for deriving the propagation time of the interference pulse signal; a reflection path perturbation integration model is constructed by the reflection path reconstruction module; The reflection path perturbation integration model performs path mapping fitting on the reflection paths formed by each reflection mirror unit through integration, and at the same time, based on the local reflection angle perturbation and the material-induced asymmetric response, finally outputs the effective propagation length of the reflection path, as the input basis for the sub-time domain correction module and the data integration module; The reflection path perturbation integral model is expressed as: ; where is the effective reflection path length value corresponding to the th reflection mirror unit; represents the spatial position variable on the reflection path; represents the actual propagation direction vector on the reflection path; represents the linear propagation direction vector of the reflection path under ideal conditions; is the square of the norm; represents the reflection perturbation gain factor at the position point with spatial position on the reflection path corresponding to the th reflection mirror unit; represents the starting coordinate of the reflection path integral; represents the ending coordinate of the reflection path integral.
5. A sports racing timing system based on multi-sensor fusion according to claim 4, characterized in that: The sub-time domain correction module is used to dynamically correct the image time points generated by the reflection path to keep it synchronized with the time line of the main visual frame. A non-linear function model composed of error factors is constructed by the sub-time domain correction module, which combines the main clock reference to perform mapping correction on the original time; the time correction mapping function is expressed as: ; where is the final synchronization time point corresponding to the th reflection path; is the reflection timestamp originally obtained by the interference pulse receiving module; represents a non - linear error compensation function composed of three error quantities; is the delay error caused by the direction deviation of the ith path; is the system - level clock drift error of the ith path; represents the quantization error caused by the time sampling resolution; represents the unpredictable disturbance term caused by comprehensive thermal noise, device response jitter, and reflection differential ; wherein is the hyperbolic tangent function; represents the square of the time delay error caused by the path direction deviation angle in the i-th reflection path; represents the cubic power of the non-linear sampling error generated by the quantization precision limitation during the time sampling process of the i-th path; is the natural logarithm function; is the square of the drift deviation between the i-th path and the master clock.
6. A sports racing timing system based on multi-sensor fusion according to claim 5, characterized in that: The continuous interference pulse signal returned by the reflection path is received and processed by the interference pulse receiving module. By detecting its time domain change structure, the echo return time point with stable characteristics is extracted; this time point is used as timestamp information and transmitted to the reflection path reconstruction module for establishing a time difference model of the optical propagation path and constructing the corresponding spatial reflection mapping relationship; The echo return time point extraction formula based on the interference pulse received signal is expressed as: ; wherein is the time point when the th pulse signal in the interference pulse sequence is received; represents an arbitrary time variable on the continuous time axis; is the continuous voltage signal stream output by the interference pulse pair receiving module; is the modulation recognition kernel function for the th pulse recognition; represents the inflection point of the detection integration curve.
7. A sports racing timing system based on multi-sensor fusion according to claim 6, characterized in that: The data integration module is used to fuse the main visual image frame and the reflection path image frame. By comparing the gradient information of the image boundary space, it is judged whether the contestant has completed the crossing line action. The objective function is constructed by the normal difference of the boundary tensor field, and the extreme value search method is used to determine the crossing line moment. The objective function is the boundary transition matching extreme value function, which is expressed as: ; Among them represents the image frame number determined as the finish line crossing action occurs; is the th frame image recorded by the main vision acquisition module; represents the image frame corrected by the reflection path reconstruction module; is the image edge scene quantity extraction function; is the normal gradient operator of the tensor; represents the squared norm; is the image time series confidence evaluation function.
8. A sports racing timing system based on multi-sensor fusion according to claim 7, characterized in that: The finish area calibration module establishes a standard space coordinate system through the preset space calibration points in the finish area, and defines the accurate position of the finish plane; based on three space calibration points with known positions, a spatial unit normal vector of the finish plane is constructed, and this normal vector is respectively mapped into the corresponding spatial perspectives of the mirror reflection sensing array module and the main visual acquisition module, so as to realize the spatial geometric alignment between the reflection path and the main visual path; This spatial geometric alignment relationship provides a unified spatial coordinate reference basis for the reflection path reconstruction module and the data integration module; ; Among them is the first spatial calibration point selected by the end region calibration module in the end region; is the second spatial calibration point selected by the end region calibration module in the end region; is the third spatial calibration point selected by the end region calibration module in the end region; represents the direction vector formed by pointing from the spatial calibration point to the spatial calibration point and is used to construct one side of the end plane; represents the direction vector formed by pointing from the spatial calibration point to the spatial calibration point and is used to construct the other side of the end plane; represents the vector cross product operation; is the spatial unit normal vector of the end plane; is the position of the th mirror reflection unit; represents the affine transformation function that projects the plane onto the image space along the mirror path, and the output is the projection image coordinate system under the reflection path; is the projection of the end plane under the viewing angle corresponding to the th mirror unit in the mirror reflection sensing array module.
Citation Information
Patent Citations
Straightway running detection and timing system and method based on face and human body recognition
CN111275030A
Long-distance running timing method based on dynamic visual identification
CN116524576A