Data perception methods, devices and media based on artificial intelligence technology

Abstract

The present invention discloses a data perception method, device, and medium based on artificial intelligence technology, and relates to the field of data perception processing technology. The data perception method, device, and medium based on artificial intelligence technology first complete the initial form of dynamic perception model establishment operation by introducing data from a data set, determine whether the characteristic position of the vehicle before parking needs to be adjusted, then introduce real-time dynamic data during parking into the dynamic perception model to form dynamic changes, record the situation during the real-time dynamic changes and compare them with standard threshold data, confirm the deviation of the parking operation, and transmit the resulting data. In this way, while establishing the model, not only can the original scene be simulated, but the actual situation can be predicted in advance to generate adjustment instruction results in advance to complete the adjustment operation, thereby improving the efficiency and accuracy of data perception processing and reducing human intervention operations.

Description

Technical Field

[0001] The present invention relates to the field of data perception and processing technology, and specifically to a data perception method, device, and medium based on artificial intelligence technology. Background Art

[0002] The reference patent name is: An entity recognition method, device, equipment and medium based on artificial intelligence (patent publication number: CN116992879A, patent publication date: 2023-11-03). The method includes: obtaining text information of the text to be recognized and target category information of the entity to be recognized; determining target prompt information based on the target category information; splicing the target prompt information and text information and inputting them into a pre-trained language model to obtain semantic representation information of the text information under the target prompt information; inputting the semantic representation information into a preset feedforward neural network to obtain the first position information of the head entity and the second position information of the tail entity in the semantic representation information; obtaining entity representations of the head entity and the tail entity based on the first position information, the second position information and the target category information; inputting the entity representation into a preset linear classification layer to obtain the probability that the head entity and the tail entity correspond to the target category information, and completing entity recognition of the text information based on the probability.

[0003] Based on the description in the above-mentioned document, although existing data perception operations can realize data perception and processing, they can only be known, and cannot make adjustments for unused data situations, so that in the subsequent actual operation process, the entity after the operation is different from the expected one. At the same time, the perception and analysis of surrounding features cannot be accurately completed, which will also cause operational errors. For this reason, the present invention provides a data perception method, device and medium based on artificial intelligence technology. Summary of the Invention

[0004] In response to the shortcomings of the existing technology, the present invention provides a data perception method, equipment and medium based on artificial intelligence technology, which solves the problem that in existing data perception operations, although data perception and processing can be achieved, only knowledge can be obtained, and adjustments cannot be made for unused data situations. As a result, in the subsequent actual operation process, the entity after the operation is different from the expected one. At the same time, the perception and analysis of surrounding features cannot be accurately completed, which will also cause operational errors.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a data perception method based on artificial intelligence technology, specifically comprising the following steps:

[0006] A1. Collect parking behavior data of vehicles using multiple video acquisition sensors and transmit it to a processing terminal via wireless communication for data analysis and processing.

[0007] A2. Receive video data and segment the data into multiple segments of image data. Filter and classify the image data to form the required multi-category dataset.

[0008] A3. Introducing the data set, first complete the initial dynamic perception model establishment operation to determine whether the vehicle's feature position needs to be adjusted before parking. Then, the real-time dynamic data during parking is introduced into the dynamic perception model to generate dynamic changes. The real-time dynamic changes are recorded and compared with the standard threshold data to identify any deviations in the parking operation. The resulting data is then transmitted.

[0009] A4. The result data is output to the vehicle's visual touch screen interface and displayed in a combination of charts, text and voice explanations.

[0010] Preferably, the operation of dividing the video data into multiple segments of image data in A2 is:

[0011] a21. Extracting video data when parking is required, and performing a video image segmentation operation according to the number of frames of the same cycle;

[0012] a22, sorting the segmented image data according to the segmentation time node, and comparing the adjacent image data, removing the data with the same feature position and dynamic operation and retaining only one;

[0013] a23. Finally, multiple segments of image data are generated without duplicate data and arranged in order of time nodes.

[0014] Preferably, the operations of filtering and classifying the image data in A2 to form the required multi-category data set are:

[0015] b1. Grayscale all image data segments. Based on the grayscale values, identify feature boundaries and segment them into multiple features. Based on feature matching in the historical database, extract image data containing both parking space features and vehicle features, while filtering out other image data.

[0016] b2. Classifying the extracted image data into driving image data and parking image data according to feature changes;

[0017] b3. Aggregate the driving image data to form a driving dataset marked as M and aggregate the parking image data to form a parking dataset marked as N.

[0018] Preferably, the operation of establishing the dynamic perception model in A3 is:

[0019] a31. Select the parking space feature to be used, extract the driving data set M and parking data N related to the selected parking space feature, and select the last one in the current driving data subset;

[0020] a32. Determine the feature portion located around the parking space feature based on the perceived data, including road features, curb features, and other vehicle features, and aggregate all features to establish a dynamic perception model similar to the driving data subset.

[0021] Preferably, the operation of determining whether the characteristic position of the vehicle before parking needs to be adjusted in A3 is:

[0022] C1. Determine based on the features in the driving data subset the vehicle's characteristic width L1, the characteristic distance between the right front wheel and the right parking entrance L2, and the characteristic distance between the left wheel and the left curb L3. Extract the threshold values ​​for the distance between the right front wheel and the right parking entrance H1 and H2, respectively, and the threshold value for the road width required for parking H3 from the stored database. Compare the measured distances with the threshold values.

[0023] C2, and the result of the comparison is:

[0024] Result 1: L1+L2+L3<H3, then the vehicle cannot perform parking maneuvers;

[0025] Result 2: L1+L2+L3≥H3, and L2≥H1, L3≥H2, then the vehicle can be parked without adjustment;

[0026] Result three: L1+L2+L3≥H3, and L2

[0027] Result 4: L1+L2+L3≥H3, and L2≥H1, L3<H2, and L2-H1≥H2-L3, then the vehicle can perform parking operations after adjustment. Conversely, when L2-H1<H2-L3, the vehicle cannot perform parking operations.

[0028] Preferably, the adjustment operations corresponding to the third and fourth results in the C2 operation are:

[0029] c21. Based on the driving data subset selected in a31, a processing operation is performed to establish a coordinate axis with the right front wheel of the vehicle feature or the left wheel of the vehicle feature as the center point marked as O, and an X-axis is established with the direction of travel of the vehicle feature, and a Y-axis is established in a direction perpendicular to the direction of travel of the vehicle feature;

[0030] ​c22. Calculate the coordinates of the starting point O' to which the center point O should move based on the required moving distance and the angle of rotation. The calculation formula is:

[0031] ;

[0032] y is the vertical distance from the center point O to the starting point O' on the Y axis, [(|L3-H2|+|L2-H1|) / 2] is the vertical distance from the center point O to the starting point O' on the X axis, and tanα is the tangent value of the wheel steering angle;

[0033] That is, the coordinates of the starting point O' are (x, y).

[0034] Preferably, after the real-time dynamic data of parking in A3 is introduced into the dynamic perception model, the operation of comparing the situation during the dynamic change process with the standard threshold data is as follows:

[0035] D1. By extracting operational data on vehicle characteristics and parking from the data repository, the dynamic perception model is introduced to simulate and determine whether there is a deviation between vehicle characteristics and parking space characteristics.

[0036] D2. If the vehicle features are completely within the parking space features, no adjustment is required. Otherwise, the position of the vehicle features should be adjusted before parking.

[0037] D3. The calculated results are transmitted to generate control instructions for actual operation.

[0038] Preferably, the operation of adjusting the position of the vehicle feature in D2 is:

[0039] d21. Determine the location of the parking space feature. Draw an extension line from the long edge of the parking space feature, closest to the vehicle, and mark the width of the inner edge of the parking space feature as R1. The vertical distance from the rear right wheel feature of the vehicle feature to the extension line is R2. The distance between the rear right wheel feature of the vehicle feature and the right parking space entrance is [(|L3 - H2| + |L2 - H1|) / 2].

[0040] d22, the moving distance of the vehicle feature in the direction of travel is obtained based on the calculation, and an instruction is generated to control the vehicle to adjust;

[0041] d23. The calculation formula for the moving distance of the vehicle feature in the direction of travel is:

[0042] ;

[0043] Wherein S is the distance that the vehicle feature needs to move, and when S<0, the vehicle feature needs to perform a reverse processing operation along the direction of travel, and when S>0, the vehicle feature performs a moving operation along the direction of travel.

[0044] The present invention also discloses a device for a data perception method based on artificial intelligence technology, comprising:

[0045] The data acquisition module has multiple acquisition sensors installed in different locations of the vehicle to sense and collect driving data and parameter data around the vehicle;

[0046] The data asset quantum security anchor host is installed on the vehicle and is used for the integrated management of data collection, analysis, processing, storage and result output;

[0047] The visual touch screen interface is used to display real-time dynamic result data, and personnel can perform intelligent control operations on the touch screen.

[0048] The present invention also discloses a computer-readable storage medium, on which a computer program is stored. When the computer program is run by a processor, the steps of the data perception method based on artificial intelligence technology are executed.

[0049] The present invention provides a data perception method, device, and medium based on artificial intelligence technology. Compared with the existing technology, it has the following advantages:

[0050] (1) The data perception method, device and medium based on artificial intelligence technology first complete the initial form of dynamic perception model establishment operation by introducing the data of the data set, determine whether the vehicle feature position needs to be adjusted before parking, and then introduce the real-time dynamic data during parking into the dynamic perception model to form dynamic changes, record the situation during the real-time dynamic change process and compare it with the standard threshold data, confirm the deviation of the parking operation, and transmit the result data. In this way, while establishing the model, not only the original scene simulation is realized, but also the actual situation is predicted in advance to generate the adjustment instruction result in advance to complete the adjustment operation, thereby improving the efficiency and accuracy of data perception processing and reducing human intervention operations.

[0051] (2) The data perception method, device and medium based on artificial intelligence technology determines whether the characteristic position of the vehicle needs to be adjusted before parking, determines various data according to the characteristics in the driving data subset, and thus calculates the distance between the vehicle and the parking space, and compares the measured distance with the threshold. In this way, it can be perceived whether parking is possible before parking and whether adjustment operations are needed, so as to avoid parking collision problems and adjust to a tolerance distance to make parking operations more accurate.

[0052] (3) The data perception method, device and medium based on artificial intelligence technology extracts the operational data about the vehicle characteristics and parking from the data repository, introduces the dynamic perception model simulation to determine whether there is a deviation between the vehicle characteristics and the parking space characteristics. If the vehicle characteristics are completely within the inclusion range of the parking space characteristics, no adjustment is required. Otherwise, the position of the vehicle characteristics is adjusted before the parking operation is performed. In this way, the position is adjusted before parking, thereby determining the final parking posture, improving the efficiency of data perception and processing, achieving fast parking and making the parking posture more standardized. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is an operational flow chart of the data perception method of the present invention;

[0054] Figure 2 This is a flowchart of the operation of image data filtering and classification of the present invention;

[0055] Figure 3 This is a principle block diagram of the data sensing device of the present invention. DETAILED DESCRIPTION

[0056] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0057] See also Figure 1-Figure 3 , the present invention provides four technical solutions:

[0058] Embodiment 1: A data perception method based on artificial intelligence technology specifically includes the following steps:

[0059] A1. Collect parking behavior data of vehicles using multiple video acquisition sensors and transmit it to a processing terminal via wireless communication for data analysis and processing.

[0060] A2. Receive video data and segment the data into multiple segments of image data. Filter and classify the image data to form the required multi-category dataset.

[0061] A3. Introducing the data set, first complete the initial dynamic perception model establishment operation to determine whether the vehicle's feature position needs to be adjusted before parking. Then, the real-time dynamic data during parking is introduced into the dynamic perception model to generate dynamic changes. The real-time dynamic changes are recorded and compared with the standard threshold data to identify any deviations in the parking operation. The resulting data is then transmitted.

[0062] A4. The result data is output to the vehicle's visual touch screen interface and displayed in a combination of charts, text and voice explanations.

[0063] Among them, by introducing the data of the data set, the initial form of the dynamic perception model establishment operation is first completed to determine whether the vehicle feature position needs to be adjusted before parking, and then the real-time dynamic data during parking is introduced into the dynamic perception model to form dynamic changes. The situation during the real-time dynamic change process is recorded and compared with the standard threshold data to confirm the deviation of the parking operation, and the result data is transmitted. In this way, while establishing the model, not only the simulation of the original scene is realized, but also the actual situation is predicted in advance to generate the adjustment instruction results in advance to complete the adjustment operation, thereby improving the efficiency and accuracy of data perception processing and reducing human intervention operations.

[0064] In the embodiment of the present invention, the operation of dividing the video data into multiple segments of image data in A2 is:

[0065] a21. Extracting video data when parking is required, and performing a video image segmentation operation according to the number of frames of the same cycle;

[0066] a22, sorting the segmented image data according to the segmentation time node, and comparing the adjacent image data, removing the data with the same feature position and dynamic operation and retaining only one;

[0067] a23. Finally, multiple segments of image data are generated without duplicate data and arranged in order of time nodes.

[0068] In the embodiment of the present invention, the operations of filtering and classifying the image data in A2 to form the required multi-category data set are:

[0069] b1. Grayscale all image data segments. Based on the grayscale values, identify feature boundaries and segment them into multiple features. Based on feature matching in the historical database, extract image data containing both parking space features and vehicle features, while filtering out other image data.

[0070] b2. Classifying the extracted image data into driving image data and parking image data according to feature changes;

[0071] b3. Aggregate the driving image data to form a driving dataset marked as M and aggregate the parking image data to form a parking dataset marked as N.

[0072] In the embodiment of the present invention, the operation of establishing the dynamic perception model in A3 is:

[0073] a31. Select the parking space feature to be used, extract the driving data set M and parking data N related to the selected parking space feature, and select the last one in the current driving data subset;

[0074] a32. Determine the feature portion located around the parking space feature based on the perceived data, including road features, curb features, and other vehicle features, and aggregate all features to establish a dynamic perception model similar to the driving data subset.

[0075] In the embodiment of the present invention, the operation of determining whether the characteristic position of the vehicle before parking needs to be adjusted in A3 is:

[0076] C1. Determine based on the features in the driving data subset the vehicle's characteristic width L1, the characteristic distance between the right front wheel and the right parking entrance L2, and the characteristic distance between the left wheel and the left curb L3. Extract the threshold values ​​for the distance between the right front wheel and the right parking entrance H1 and H2, respectively, and the threshold value for the road width required for parking H3 from the stored database. Compare the measured distances with the threshold values.

[0077] C2, and the result of the comparison is:

[0078] Result 1: L1+L2+L3<H3, then the vehicle cannot perform parking maneuvers;

[0079] Result 2: L1+L2+L3≥H3, and L2≥H1, L3≥H2, then the vehicle can be parked without adjustment;

[0080] Result three: L1+L2+L3≥H3, and L2

[0081] Result 4: L1+L2+L3≥H3, and L2≥H1, L3<H2, and L2-H1≥H2-L3, then the vehicle can perform parking operations after adjustment. Conversely, when L2-H1<H2-L3, the vehicle cannot perform parking operations.

[0082] In the embodiment of the present invention, the adjustment operations corresponding to the third and fourth results in the C2 operation are:

[0083] c21. Based on the driving data subset selected in a31, a processing operation is performed to establish a coordinate axis with the right front wheel of the vehicle feature or the left wheel of the vehicle feature as the center point marked as O, and an X-axis is established with the direction of travel of the vehicle feature, and a Y-axis is established in a direction perpendicular to the direction of travel of the vehicle feature;

[0084] ​c22. Calculate the coordinates of the starting point O' to which the center point O should move based on the required moving distance and the angle of rotation. The calculation formula is:

[0085] ;

[0086] y is the vertical distance from the center point O to the starting point O' on the Y axis, [(|L3-H2|+|L2-H1|) / 2] is the vertical distance from the center point O to the starting point O' on the X axis, and tanα is the tangent value of the wheel steering angle;

[0087] That is, the coordinates of the starting point O' are (x, y).

[0088] Among them, by determining whether the characteristic position of the vehicle needs to be adjusted before parking, various data are determined according to the characteristics in the driving data subset, so as to calculate the distance between the vehicle and the parking space, and compare the measured distance with the threshold. In this way, it is possible to perceive whether parking is possible before parking and whether adjustment operations are needed, so as to avoid parking collision problems and adjust to a fault-tolerant distance to make parking operations more accurate.

[0089] In the embodiment of the present invention, after the real-time dynamic data of parking in A3 is introduced into the dynamic perception model, the operation of comparing the dynamic change process with the standard threshold data is as follows:

[0090] D1. By extracting operational data on vehicle characteristics and parking from the data repository, the dynamic perception model is introduced to simulate and determine whether there is a deviation between vehicle characteristics and parking space characteristics.

[0091] D2. If the vehicle features are completely within the parking space features, no adjustment is required. Otherwise, the position of the vehicle features should be adjusted before parking.

[0092] D3. The calculated results are transmitted to generate control instructions for actual operation.

[0093] In the embodiment of the present invention, the operation of adjusting the position of the vehicle feature in D2 is:

[0094] d21. Determine the location of the parking space feature. Draw an extension line from the long edge of the parking space feature, closest to the vehicle, and mark the width of the inner edge of the parking space feature as R1. The vertical distance from the rear right wheel feature of the vehicle feature to the extension line is R2. The distance between the rear right wheel feature of the vehicle feature and the right parking space entrance is [(|L3 - H2| + |L2 - H1|) / 2].

[0095] d22, the moving distance of the vehicle feature in the direction of travel is obtained based on the calculation, and an instruction is generated to control the vehicle to adjust;

[0096] d23. The calculation formula for the moving distance of the vehicle feature in the direction of travel is:

[0097] ;

[0098] Wherein S is the distance that the vehicle feature needs to move, and when S<0, the vehicle feature needs to perform a reverse processing operation along the direction of travel, and when S>0, the vehicle feature performs a moving operation along the direction of travel.

[0099] Among them, by extracting the operational data about vehicle feature parking from the data repository, the dynamic perception model simulation is introduced to determine whether there is a deviation between the vehicle features and the parking space features. If the vehicle features are completely within the inclusion range of the parking space features, no adjustment is required. Otherwise, the position of the vehicle features is adjusted before the parking operation is performed. In this way, the position is adjusted before parking, thereby determining the final parking posture, improving the efficiency of data perception and processing, achieving fast parking, and making the parking posture more standardized.

[0100] The difference between the second embodiment and the first embodiment is that the present invention further discloses a device for a data perception method based on artificial intelligence technology, including:

[0101] The data acquisition module has multiple acquisition sensors installed in different locations of the vehicle to sense and collect driving data and parameter data around the vehicle;

[0102] The data asset quantum security anchor host is installed on the vehicle and is used for the integrated management of data collection, analysis, processing, storage and result output;

[0103] The visual touch screen interface is used to display real-time dynamic result data, and personnel can perform intelligent control operations on the touch screen.

[0104] Among them, the Data Asset Quantum Security Anchor Host is a 2U dual-core rack-mounted server developed based on the Hygon 5000 series processor platform. It is an existing equipment product, mainly used for integrated management operations of data collection, analysis, processing, storage and result output.

[0105] Example 3: The difference compared with Example 1 and Example 2 is that: the present invention also discloses a computer-readable storage medium, on which a computer program is stored, and when the computer program is run by a processor, the steps of the data perception method based on artificial intelligence technology are executed.

[0106] The fourth embodiment differs from the first and second embodiments in that the data perception method of the existing artificial intelligence technology and the data perception method of the artificial intelligence technology of the present invention are used to test various types of vehicles, and the parking completion time and parking deviation are recorded during the test. The specific results are shown in Table 1:

[0107] Table 1 Parking situation table

[0108]

[0109] In summary, by adopting the data perception method of the artificial intelligence technology of the present invention to perform parking operations on a vehicle, the parking time is shorter and the parking success rate is higher, and it can be better applied in actual vehicle parking operations.

[0110] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0111] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0112] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A data perception method based on artificial intelligence technology, characterized by: The specific steps include: A1. Collect parking behavior data of vehicles using multiple video acquisition sensors and transmit it to a processing terminal via wireless communication for data analysis and processing. A2. Receive video data and segment the data into multiple segments of image data. Filter and classify the image data to form the required multi-category dataset. A3. Introducing the data set, first complete the initial dynamic perception model establishment operation to determine whether the vehicle's feature position needs to be adjusted before parking. Then, the real-time dynamic data during parking is introduced into the dynamic perception model to generate dynamic changes. The real-time dynamic changes are recorded and compared with the standard threshold data to identify any deviations in the parking operation. The resulting data is then transmitted. A4. The result data is output to the vehicle's visual touch screen interface and displayed in a combination of charts, text and voice explanations; The operation of establishing the dynamic perception model in A3 is: a31. Select the parking space feature to be used, extract the driving data set M and parking data N related to the selected parking space feature, and select the last one in the current driving data subset; a32. Determine the feature portion surrounding the parking space feature based on the sensed data, including road features, curb features, and other vehicle features, and aggregate all features to establish a dynamic perception model similar to the driving data subset; The operation of determining whether the characteristic position of the vehicle before parking needs to be adjusted in A3 is: C1. Determine based on the features in the driving data subset the vehicle's characteristic width L1, the characteristic distance between the right front wheel and the right parking entrance L2, and the characteristic distance between the left wheel and the left curb L3. Extract the threshold values ​​for the distance between the right front wheel and the right parking entrance H1 and H2, respectively, and the threshold value for the road width required for parking H3 from the stored database. Compare the measured distances with the threshold values. C2, and the result of the comparison is: Result 1: L1+L2+L3<H3, then the vehicle cannot perform parking maneuvers; Result 2: L1+L2+L3≥H3, and L2≥H1, L3≥H2, then the vehicle can be parked without adjustment; Result three: L1+L2+L3≥H3, and L2<H1, L3≥H2, and L3-H2≥H1-L2, then the vehicle can perform parking after adjustment. Conversely, if L3-H2<H1-L2, the vehicle cannot perform parking; Result 4: If L1+L2+L3≥H3, L2≥H1, L3<H2, and L2-H1≥H2-L3, the vehicle can be parked after adjustment. Conversely, if L2-H1<H2-L3, the vehicle cannot be parked. The adjustment operations corresponding to the third and fourth results in the C2 operation are: c21. Based on the driving data subset selected in a31, a processing operation is performed to establish a coordinate axis with the right front wheel of the vehicle feature or the left wheel of the vehicle feature as the center point marked as O, and an X-axis is established with the direction of travel of the vehicle feature, and a Y-axis is established in a direction perpendicular to the direction of travel of the vehicle feature; c22. Calculate the coordinates of the starting point O' to which the center point O should move based on the required moving distance and the angle of rotation. The calculation formula is: ； y is the vertical distance from the center point O to the starting point O' on the Y axis, [(|L3-H2|+|L2-H1|) / 2] is the vertical distance from the center point O to the starting point O' on the X axis, and tanα is the tangent value of the wheel steering angle; That is, the coordinates of the starting point O' are (x, y); After the real-time dynamic data of parking in A3 is introduced into the dynamic perception model, the operation of comparing the situation during the dynamic change process with the standard threshold data is as follows: D1. By extracting operational data on vehicle characteristics and parking from the data repository, the dynamic perception model is introduced to simulate and determine whether there is a deviation between vehicle characteristics and parking space characteristics. D2. If the vehicle features are completely within the parking space features, no adjustment is required. Otherwise, the position of the vehicle features should be adjusted before parking. D3. The calculated results are transmitted to generate control instructions for actual operation.

2. The data perception method based on artificial intelligence technology according to claim 1, characterized in that: The operation of dividing the video data into multiple segments of image data in A2 is: a21. Extracting video data when parking is required, and performing a video image segmentation operation according to the number of frames of the same cycle; a22, sorting the segmented image data according to the segmentation time node, and comparing the adjacent image data, removing the data with the same feature position and dynamic operation and retaining only one; a23. Finally, multiple segments of image data are generated without duplicate data and arranged in order of time nodes.

3. The data perception method based on artificial intelligence technology according to claim 2, characterized in that: The operations for filtering and classifying the image data in A2 to form the required multi-category dataset are as follows: b1. Grayscale all image data segments. Based on the grayscale values, identify feature boundaries and segment them into multiple features. Based on feature matching in the historical database, extract image data containing both parking space features and vehicle features, while filtering out other image data. b2. Classifying the extracted image data into driving image data and parking image data according to feature changes; b3. Aggregate the driving image data to form a driving dataset marked as M and aggregate the parking image data to form a parking dataset marked as N.

4. The data perception method based on artificial intelligence technology according to claim 1, characterized in that: The operation of adjusting the position of the vehicle feature in D2 is: d21. Determine the location of the parking space feature. Draw an extension line from the long edge of the parking space feature, closest to the vehicle, and mark the width of the inner edge of the parking space feature as R1. The vertical distance from the rear right wheel feature of the vehicle feature to the extension line is R2. The distance between the rear right wheel feature of the vehicle feature and the right parking space entrance is [(|L3 - H2| + |L2 - H1|) / 2]. d22, the moving distance of the vehicle feature in the direction of travel is obtained based on the calculation, and an instruction is generated to control the vehicle to adjust; d23. The calculation formula for the moving distance of the vehicle feature in the direction of travel is: ； Wherein S is the distance that the vehicle feature needs to move, and when S<0, the vehicle feature needs to perform a reverse processing operation along the direction of travel, and when S>0, the vehicle feature performs a moving operation along the direction of travel.

5. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, executes the steps of the data perception method based on artificial intelligence technology as described in any one of claims 1 to 4.

Images