Collision detection method, device and system and storage medium

Abstract

The invention provides a collision detection method, device and system and a storage medium, and belongs to the technical field of collision detection.The method comprises the steps that a collision signal and multiple pieces of original three-axis acceleration data are obtained from an acceleration sensor arranged on an anti-collision beam; if the time for generating the collision signal is within the preset suppression time, whether the anti-collision beam is collided or not is analyzed based on all the original three-axis acceleration data, and an initial collision result is obtained after analysis; and determining whether to acquire data from an angle sensor arranged on the anti-collision beam or not according to the initial collision result, analyzing whether the anti-collision beam is knocked down or not according to the acquired data, and sending an analysis result to a preset terminal. According to the invention, the collision condition of the anti-collision beam can be accurately detected, the accuracy of collision detection is improved, and the false alarm rate and the missing report rate of collision detection are reduced.

Description

Technical Field

[0001] The present invention mainly relates to the technical field of collision detection, and in particular to a collision detection method, device, system and storage medium. Background Art

[0002] At railway bridges and culverts, height limit frames are installed to prevent vehicles from directly colliding with the railway bridges and culverts due to their excessive height. However, in many cases, there are still many vehicles with excessive height that want to force their way through the railway bridges and culverts, which results in the anti-collision beams in the height limit frames being frequently collided with by excessive height vehicles. The existing method is to install sensors in the anti-collision beams. When the anti-collision beams are collided, the sensors will send a signal to the background, thereby letting the background know that the anti-collision beams have been collided. However, because the sensitivities of the sensors are different, some are too sensitive and some are not sensitive enough, resulting in high false alarm and missed alarm rates of collision detection. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a collision detection method, device, system and storage medium in view of the deficiencies in the prior art.

[0004] The technical solution of the present invention to solve the above technical problem is as follows: A collision detection method comprises the following steps:

[0005] S1: Obtaining a collision signal and a plurality of original three-axis acceleration data from an acceleration sensor disposed on the anti-collision beam;

[0006] S2: If the time when the collision signal is generated is within the preset suppression time, analyzing whether there is a collision with the anti-collision beam based on all the original three-axis acceleration data, and obtaining an initial collision result after analysis;

[0007] S3: Determine whether to obtain data from the angle sensor provided on the anti-collision beam according to the initial collision result, analyze whether the anti-collision beam is knocked down according to the obtained data, and send the analysis result to a preset terminal.

[0008] Another technical solution of the present invention to solve the above technical problem is as follows: a collision detection device, comprising:

[0009] A data acquisition module, used to obtain a collision signal and a plurality of original three-axis acceleration data from an acceleration sensor arranged on the anti-collision beam;

[0010] A collision analysis module, configured to analyze whether the collision occurs on the anti-collision beam based on all the original three-axis acceleration data if the collision signal is generated within a preset suppression time, and obtain an initial collision result after analysis;

[0011] The knock-down analysis module is used to determine whether to obtain data from the angle sensor arranged on the anti-collision beam according to the initial collision result, and analyze whether the anti-collision beam is knocked down according to the acquired data, and send the analysis result to a preset terminal.

[0012] Based on the above-mentioned collision detection method, the present invention also provides a collision detection system.

[0013] Another technical solution of the present invention to solve the above technical problem is as follows: a collision detection system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the collision detection method described above is implemented.

[0014] Based on the above-mentioned collision detection method, the present invention also provides a computer-readable storage medium.

[0015] Another technical solution of the present invention to solve the above technical problem is as follows: a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the collision detection method as described above is implemented.

[0016] The beneficial effects of the present invention are as follows: by obtaining a collision signal and original three-axis acceleration data from an acceleration sensor arranged on the anti-collision beam, if the time when the collision signal is generated is within a preset suppression time, an analysis is performed based on the original three-axis acceleration data to determine whether the anti-collision beam is collided, and an initial collision result is obtained after analysis. According to the initial collision result, it is determined whether to obtain data from an angle sensor arranged on the anti-collision beam, and an analysis is performed based on the obtained data to determine whether the anti-collision beam is knocked down, and the analysis result is sent to a preset terminal, which can accurately detect the situation where the anti-collision beam is hit, thereby improving the accuracy of collision detection and reducing the false alarm rate and missed alarm rate of collision detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic diagram of a collision detection method provided by an embodiment of the present invention Figure 2

[0018] Figure 2 A module block diagram of a collision detection device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The principles and features of the present invention are described below in conjunction with the accompanying drawings. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.

[0020] Figure 1 A schematic flow chart of a collision detection method provided by an embodiment of the present invention.

[0021] like Figure 1 As shown, a collision detection method comprises the following steps:

[0022] S1: Obtaining a collision signal and a plurality of original three-axis acceleration data from an acceleration sensor disposed on the anti-collision beam;

[0023] S2: If the time when the collision signal is generated is within the preset suppression time, analyzing whether there is a collision with the anti-collision beam based on all the original three-axis acceleration data, and obtaining an initial collision result after analysis;

[0024] S3: Determine whether to obtain data from the angle sensor provided on the anti-collision beam according to the initial collision result, analyze whether the anti-collision beam is knocked down according to the obtained data, and send the analysis result to a preset terminal.

[0025] It should be understood that the preset terminal may be a client.

[0026] It should be understood that if the collision signal is not within the preset suppression time, it means that the collision signal of the acceleration sensor is not generated by a collision, so no collision is sent to the preset terminal.

[0027] In the above embodiment, by obtaining a collision signal and original three-axis acceleration data from an acceleration sensor arranged on the anti-collision beam, if the time when the collision signal is generated is within the preset suppression time, an analysis is performed based on the original three-axis acceleration data to determine whether the anti-collision beam is collided. After analysis, an initial collision result is obtained, and according to the initial collision result, it is determined whether to obtain data from the angle sensor arranged on the anti-collision beam, and an analysis is performed based on the obtained data to determine whether the anti-collision beam is knocked down. The analysis result is sent to a preset terminal, which can accurately detect the situation where the anti-collision beam is hit, thereby improving the accuracy of collision detection and reducing the false alarm rate and missed alarm rate of collision detection.

[0028] Optionally, as an embodiment of the present invention, in S2, the process of analyzing whether the anti-collision beam is collided based on all the original three-axis acceleration data and obtaining the initial collision result after analysis includes:

[0029] S21: preprocessing all the original three-axis acceleration data to obtain a plurality of preprocessed three-axis acceleration data;

[0030] S22: Calculating target three-axis acceleration data for all the pre-processed three-axis acceleration data to obtain a target X-axis acceleration value, a target Y-axis acceleration value, and a target Z-axis acceleration value;

[0031] S23: Determine whether the target X-axis acceleration value, the target Y-axis acceleration value, and the target Z-axis acceleration value meet the determination condition, if not, return to S1; if yes, execute S24, the determination condition is that the target X-axis acceleration value is greater than or equal to the preset X-axis acceleration threshold, or the target Y-axis acceleration value is greater than or equal to the preset Y-axis acceleration threshold, or the target Z-axis acceleration value is greater than or equal to the preset Z-axis acceleration threshold;

[0032] S24: performing labeling processing on each of the preprocessed three-axis acceleration data respectively to obtain a labeled three-axis acceleration value corresponding to each of the preprocessed three-axis acceleration data, and combining all the labeled three-axis acceleration values ​​to obtain a labeled three-axis acceleration data set;

[0033] S25: constructing a training model by using a K-nearest neighbor algorithm, and performing model analysis on the training model according to the annotated three-axis acceleration data set to obtain a collision detection model;

[0034] S26: Detecting all the pre-processed three-axis acceleration data using the collision detection model to obtain an initial collision result.

[0035] It should be understood that the absolute values ​​Dx, Dy, and Dz of the differences between the maximum and minimum accelerations of the three axes X, Y, and Z (i.e., the target X-axis acceleration value, the target Y-axis acceleration value, and the target Z-axis acceleration value) are compared with the preset X, Y, and Z three-axis thresholds Tx (i.e., the preset X-axis acceleration threshold), Ty (i.e., the preset X-axis acceleration threshold), and Tz (i.e., the preset Z-axis acceleration threshold). If the D values ​​of the three axes X, Y, and Z are all less than the corresponding thresholds, it is judged as non-collision; otherwise, it proceeds to the next step. The algorithm formula of this step is as follows, where S is the judgment result of this step:

[0036]

[0037] S=S x VS y VS z .

[0038] It should be understood that the KNN machine learning algorithm is used to process the acceleration data of the three axes of X, Y, and Z (i.e., the three-axis acceleration data set after annotation) to determine whether it is false alarm data.

[0039] Specifically, the K-nearest neighbor algorithm, namely KNN (K-Nearest Neighbors), is a basic classification and regression algorithm, and its working principle is very intuitive: prediction is made by measuring the distance between different feature values.

[0040] Specifically, the KNN algorithm (i.e., K nearest neighbor algorithm) is first judged based on a known data set. The acceleration data of the collision sensor (i.e., the pre-processed three-axis acceleration data) is defined as three sets of vector data according to the three axes of X, Y, and Z. The acceleration vector data of each collision is collected, and the data (i.e., the pre-processed three-axis acceleration data) is manually labeled to generate a training data set for the KNN algorithm (i.e., the labeled three-axis acceleration data set). The vector data of the three axes of X, Y, and Z are defined as follows, where x, y, and z are normalized acceleration vector data:

[0041] x=(x0,x1…x n ),

[0042] y=(y0,y1…y n ),

[0043] z=(z0,z1…z n ).

[0044] In the above embodiment, whether the anti-collision beam is collided is analyzed based on all the original three-axis acceleration data, and the initial collision result is obtained after analysis, which confirms the initial collision result, lays the foundation for further precise collision detection in the future, improves the accuracy of collision detection, and reduces the false alarm rate and missed alarm rate of collision detection.

[0045] Optionally, as an embodiment of the present invention, in S21, the process of preprocessing all the original three-axis acceleration data to obtain a plurality of preprocessed three-axis acceleration data includes:

[0046] S211: Counting the data length of each of the original three-axis acceleration data respectively to obtain the data length corresponding to each of the original three-axis acceleration data;

[0047] S212: Determine whether the data length is greater than a preset data length threshold. If not, send the non-collision anti-collision beam to the preset terminal and return to S1; if yes, execute S213;

[0048] S213: performing data filtering processing on the original three-axis acceleration data corresponding to the data length to obtain pre-processed three-axis acceleration data.

[0049] It should be understood that the validity of the acceleration data in the three axes of X, Y and Z is judged. When the valid data length (i.e., data length) is greater than a preset threshold (i.e., the preset data length threshold), the data is judged to be valid; when it is less than the preset threshold (i.e., the preset data length threshold), it is judged to be invalid data and it is judged to be a non-collision (i.e., no collision with the anti-collision beam).

[0050] Specifically, a data edge filtering algorithm is used to perform data filtering processing on the original three-axis acceleration data corresponding to the data length, wherein the data edge filtering algorithm removes the head and tail data of the data and filters out misjudgment caused by edge data fluctuations.

[0051] It should be understood that the data edge filtering algorithm generally refers to an algorithm used to remove or reduce noise and details in image processing or signal processing, especially to remove high-frequency noise components in an image. The data edge filtering algorithm can smooth an image while preserving edge information as much as possible.

[0052] In the above embodiment, all original three-axis acceleration data are preprocessed to obtain multiple preprocessed three-axis acceleration data, which can filter out misjudgments caused by edge data fluctuations, improve the accuracy of collision detection, and reduce the false alarm rate and missed alarm rate of collision detection.

[0053] Optionally, as an embodiment of the present invention, the preprocessed three-axis acceleration data includes a preprocessed X-axis acceleration value, a preprocessed Y-axis acceleration value, and a preprocessed Z-axis acceleration value.

[0054] The process of S22 includes:

[0055] All the pre-processed X-axis acceleration values ​​are calculated by the first formula to obtain the target X-axis acceleration value, and the first formula is:

[0056] D x =|max(x1, x2, ..., x n )-min(x1, x2, …, x n )|,

[0057] Among them, D x is the target X-axis acceleration value, x n is the nth preprocessed X-axis acceleration value;

[0058] All the pre-processed Y-axis acceleration values ​​are calculated by the second formula to obtain the target Y-axis acceleration value, and the second formula is:

[0059] D y =|max(y1, y2, ..., y n )-min(y1,y2,…,y n )|,

[0060] Among them, D y is the target Y-axis acceleration value, y n is the Y-axis acceleration value after the nth preprocessing;

[0061] All the pre-processed Z-axis acceleration values ​​are calculated by the third formula to obtain the target Z-axis acceleration value, and the third formula is:

[0062] D z =|max(z1, z2, ..., z n )-min(z1,z2,…,z n )|,

[0063] Among them, D z is the target Z-axis acceleration value, z n It is the nth Z-axis acceleration value after preprocessing.

[0064] Specifically, the absolute values ​​Dx, Dy, and Dz of the differences between the maximum and minimum accelerations in the three axes of X, Y, and Z (i.e., the target X-axis acceleration value, the target Y-axis acceleration value, and the target Z-axis acceleration value) are calculated respectively, and the formulas are as follows:

[0065] D x =|max(x1, x2, ..., x n )-min(x1, x2, …, x n )|,

[0066] D y =|max(y1, y2, ..., y n )-min(y1,y2,…,y n )|,

[0067] D z =|max(z1, z2, ..., z n )-min(z1,z2,…,z n )|.

[0068] In the above embodiment, the target three-axis acceleration data is calculated for all pre-processed three-axis acceleration data to obtain the target X-axis acceleration value, the target Y-axis acceleration value and the target Z-axis acceleration value, which improves the accuracy of collision detection and reduces the false alarm rate and missed alarm rate of collision detection.

[0069] Optionally, as an embodiment of the present invention, in S25, the process of performing model analysis on the training model according to the annotated three-axis acceleration data set to obtain the collision detection model includes:

[0070] Dividing the annotated three-axis acceleration data set into an annotated three-axis acceleration training set and an annotated three-axis acceleration verification set according to a preset ratio;

[0071] The training model is trained using the labeled three-axis acceleration training set to obtain a trained model;

[0072] The trained model is verified by using the labeled three-axis acceleration verification set to obtain a collision detection model.

[0073] It should be understood that the preset ratio may be 8:2.

[0074] Specifically, the data (i.e., the annotated three-axis acceleration data set) is split into a training set (i.e., the annotated three-axis acceleration training set) and a validation set (i.e., the annotated three-axis acceleration validation set) according to a ratio of 8:2. The training set (i.e., the annotated three-axis acceleration training set) is used to train the KNN algorithm (i.e., the training model), and the validation set (i.e., the annotated three-axis acceleration validation set) is used to test the algorithm effect. After multiple training and testing, the K value is finally set to 5, and the Euclidean distance is used to measure the distance. The measurement formula is as follows:

[0075]

[0076] In the above embodiment, a collision detection model is obtained by performing model analysis on the training model according to the labeled three-axis acceleration data set, which can accurately detect the situation where the anti-collision beam is hit, improve the accuracy of collision detection, and reduce the false alarm rate and missed alarm rate of collision detection.

[0077] Optionally, as an embodiment of the present invention, the process of S3 includes:

[0078] S31: Determine whether the initial collision result is a collision with an anti-collision beam. If not, send the initial collision result as an analysis result to the preset terminal; if yes, execute S32;

[0079] S32: Obtaining an angle change value from an angle sensor disposed on the anti-collision beam;

[0080] S33: Determine whether the angle change value is greater than or equal to a first angle change threshold value. If so, send the knocked-down anti-collision beam as an analysis result to the preset terminal; if not, execute S34;

[0081] S34: Determine whether the angle change value is greater than or equal to a second angle change threshold value. If so, send the knocked-down anti-collision beam as an analysis result to the preset terminal; if not, execute S35;

[0082] S35: Determine whether angle abnormality information is obtained from the angle sensor. If so, send the knocked-down anti-collision beam as an analysis result to the preset terminal; if not, send the angle abnormality information as an analysis result to the preset terminal.

[0083] It should be understood that the first angle change threshold is greater than the second angle change threshold.

[0084] It should be understood that whether the anti-collision beam has collapsed is determined by a logic algorithm based on the collision detection algorithm result (ie, the initial collision result) and the angle data of the collision sensor (ie, the angle change value).

[0085] In the above embodiment, whether to obtain data from the angle sensor provided on the anti-collision beam is determined according to the initial collision result, and an analysis is performed based on the obtained data as to whether the anti-collision beam is knocked down. This can accurately detect the situation in which the anti-collision beam is hit, thereby improving the accuracy of collision detection and reducing the false alarm rate and missed alarm rate of collision detection.

[0086] Optionally, as an embodiment of the present invention, it further includes:

[0087] When the preset judgment time is reached, it is judged whether all the initial collision results within the preset judgment time meet the judgment conditions. If so, the collision anti-collision beam is sent to the preset terminal; if not, the non-collision anti-collision beam is sent to the preset terminal, and the judgment condition is that any initial collision result among all the initial collision results is a collision.

[0088] In the above embodiment, when the preset judgment time is reached, it is judged that all initial collision results within the preset judgment time meet the judgment conditions, and the situation of the anti-collision beam being hit can be accurately detected, thereby improving the accuracy of collision detection and reducing the false alarm rate and missed alarm rate of collision detection.

[0089] Optionally, as another embodiment of the present invention, the present invention analyzes and processes the acceleration data and angle data of the collision alarm module to reduce the false alarm rate and improve the accuracy of collision detection. The collision detection algorithm includes two parts. The first part is the anti-collision beam collision discrimination algorithm. The goal of this algorithm is to distinguish whether the anti-collision beam is actually knocked down by a passing vehicle; the second part is the anti-collision beam knockdown discrimination algorithm. The goal of this algorithm is to judge whether a serious collision occurs and causes the anti-collision beam to be knocked down, so as to improve the accuracy of the height limit frame collision monitoring and reduce the false alarm rate and missed alarm rate.

[0090] Optionally, as another embodiment of the present invention, the collision sensor of the present invention has a built-in three-axis acceleration sensor chip, which can output acceleration values ​​of the three axes of X, Y, and Z (XYZ as a group) when the device itself is vibrated, and send the three-axis acceleration values ​​to the information processing device. The collision discrimination algorithm built into the information processing device will analyze and process the acceleration data of the collision sensor, and output a judgment result on whether a collision has occurred. The present invention is divided into a combination of an acceleration data preprocessing algorithm and a collision discrimination algorithm based on a machine learning algorithm, which is specifically aimed at anti-collision beam collision product scenarios.

[0091] Figure 2 A module block diagram of a collision detection device provided by an embodiment of the present invention.

[0092] Optionally, as another embodiment of the present invention, Figure 2 As shown, a collision detection device comprises:

[0093] A data acquisition module, used to obtain a collision signal and a plurality of original three-axis acceleration data from an acceleration sensor arranged on the anti-collision beam;

[0094] A collision analysis module, configured to analyze whether the collision occurs on the anti-collision beam based on all the original three-axis acceleration data if the collision signal is generated within a preset suppression time, and obtain an initial collision result after analysis;

[0095] The knock-down analysis module is used to determine whether to obtain data from the angle sensor arranged on the anti-collision beam according to the initial collision result, and analyze whether the anti-collision beam is knocked down according to the acquired data, and send the analysis result to a preset terminal.

[0096] Optionally, another embodiment of the present invention provides a collision detection system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the collision detection method described above is implemented. The system may be a computer or other system.

[0097] Optionally, another embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the collision detection method as described above is implemented.

[0098] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0099] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0100] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0101] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present invention.

[0102] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0103] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0104] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A collision detection method, characterized in that: The steps include: S1: Obtaining a collision signal and a plurality of original three-axis acceleration data from an acceleration sensor disposed on the anti-collision beam; S2: If the time when the collision signal is generated is within the preset suppression time, analyzing whether there is a collision with the anti-collision beam based on all the original three-axis acceleration data, and obtaining an initial collision result after analysis; S3: Determine whether to obtain data from the angle sensor provided on the anti-collision beam according to the initial collision result, analyze whether the anti-collision beam is knocked down according to the obtained data, and send the analysis result to a preset terminal.

2. The collision detection method according to claim 1, characterized in that: In S2, the process of analyzing whether the collision occurs on the anti-collision beam based on all the original three-axis acceleration data and obtaining the initial collision result after analysis includes: S21: preprocessing all the original three-axis acceleration data to obtain a plurality of preprocessed three-axis acceleration data; S22: Calculating target three-axis acceleration data for all the pre-processed three-axis acceleration data to obtain a target X-axis acceleration value, a target Y-axis acceleration value, and a target Z-axis acceleration value; S23: Determine whether the target X-axis acceleration value, the target Y-axis acceleration value, and the target Z-axis acceleration value meet the determination condition, if not, return to S1; if yes, execute S24, the determination condition is that the target X-axis acceleration value is greater than or equal to the preset X-axis acceleration threshold, or the target Y-axis acceleration value is greater than or equal to the preset Y-axis acceleration threshold, or the target Z-axis acceleration value is greater than or equal to the preset Z-axis acceleration threshold; S24: performing labeling processing on each of the preprocessed three-axis acceleration data respectively to obtain a labeled three-axis acceleration value corresponding to each of the preprocessed three-axis acceleration data, and combining all the labeled three-axis acceleration values ​​to obtain a labeled three-axis acceleration data set; S25: constructing a training model by using a K-nearest neighbor algorithm, and performing model analysis on the training model according to the annotated three-axis acceleration data set to obtain a collision detection model; S26: Detecting all the pre-processed three-axis acceleration data using the collision detection model to obtain an initial collision result.

3. The collision detection method according to claim 2, characterized in that: In S21, the process of preprocessing all the original three-axis acceleration data to obtain a plurality of preprocessed three-axis acceleration data includes: S211: Counting the data length of each of the original three-axis acceleration data respectively to obtain the data length corresponding to each of the original three-axis acceleration data; S212: Determine whether the data length is greater than a preset data length threshold. If not, send the non-collision anti-collision beam to the preset terminal and return to S1; if yes, execute S213; S213: performing data filtering processing on the original three-axis acceleration data corresponding to the data length to obtain pre-processed three-axis acceleration data.

4. The collision detection method according to claim 2, characterized in that: The pre-processed three-axis acceleration data includes a pre-processed X-axis acceleration value, a pre-processed Y-axis acceleration value, and a pre-processed Z-axis acceleration value. The process of S22 includes: All the pre-processed X-axis acceleration values ​​are calculated by the first formula to obtain the target X-axis acceleration value, and the first formula is: D x =|max(x1,x2,…,x n )-min(x1, x2,…, x n )|, Among them, D x is the target X-axis acceleration value, x n is the nth preprocessed X-axis acceleration value; All the pre-processed Y-axis acceleration values ​​are calculated by the second formula to obtain the target Y-axis acceleration value, and the second formula is: D y =|max(y1,y2,…,y n )-min(y1,y2,…,y n )|, Among them, D y is the target Y-axis acceleration value, y n is the Y-axis acceleration value after the nth preprocessing; All the pre-processed Z-axis acceleration values ​​are calculated by the third formula to obtain the target Z-axis acceleration value, and the third formula is: D z =|max(z1,z2,…,z n )-min(z1,z2,…,z n )|, Among them, D z is the target Z-axis acceleration value, z n It is the nth Z-axis acceleration value after preprocessing.

5. The collision detection method according to claim 2, characterized in that: In S25, the process of performing model analysis on the training model according to the annotated three-axis acceleration data set to obtain a collision detection model includes: Dividing the annotated three-axis acceleration data set into an annotated three-axis acceleration training set and an annotated three-axis acceleration verification set according to a preset ratio; The training model is trained using the labeled three-axis acceleration training set to obtain a trained model; The trained model is verified by using the labeled three-axis acceleration verification set to obtain a collision detection model.

6. The collision detection method according to claim 2, characterized in that: The S3 process includes: S31: Determine whether the initial collision result is a collision with an anti-collision beam. If not, send the initial collision result as an analysis result to the preset terminal; if yes, execute S32; S32: Obtaining an angle change value from an angle sensor disposed on the anti-collision beam; S33: Determine whether the angle change value is greater than or equal to a first angle change threshold value. If so, send the knocked-down anti-collision beam as an analysis result to the preset terminal; if not, execute S34; S34: Determine whether the angle change value is greater than or equal to a second angle change threshold value. If so, send the knocked-down anti-collision beam as an analysis result to the preset terminal; if not, execute S35; S35: Determine whether angle abnormality information is obtained from the angle sensor. If so, send the knocked-down anti-collision beam as an analysis result to the preset terminal; if not, send the angle abnormality information as an analysis result to the preset terminal.

7. The collision detection method according to claim 1, characterized in that: Also includes: When the preset judgment time is reached, it is judged whether all the initial collision results within the preset judgment time meet the judgment condition, and if so, the collision anti-collision beam is sent to the preset terminal; If not, the non-collision anti-collision beam is sent to the preset terminal, and the judgment condition is that any initial collision result among all the initial collision results is a collision.

8. A collision detection device, characterized in that: include: A data acquisition module, used for acquiring a collision signal and a plurality of original three-axis acceleration data from an acceleration sensor arranged on the anti-collision beam; A collision analysis module, configured to analyze whether the collision occurs on the anti-collision beam based on all the original three-axis acceleration data if the collision signal is generated within a preset suppression time, and obtain an initial collision result after analysis; The knock-down analysis module is used to determine whether to obtain data from the angle sensor arranged on the anti-collision beam according to the initial collision result, and analyze whether the anti-collision beam is knocked down according to the acquired data, and send the analysis result to a preset terminal.

9. A collision detection system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the collision detection method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the collision detection method according to any one of claims 1 to 7 is implemented.

Images