A method and device for correcting entry and exit orders based on deep learning algorithm
Through deep learning algorithms, the vehicle feature values and similarity are extracted, and the entry and exit time points in the smart parking system are automatically corrected, which solves the problem of inefficient order correction in the prior art and improves the operational efficiency of vehicle parking management.
Patent Information
- Application Number
- CN202210264605.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-17
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2042-03-17
AI Technical Summary
The existing smart parking management system has inefficient problems when handling vehicle entry and exit orders, and requires manual intervention to repair wrong orders, resulting in large workload and low efficiency.
Using a deep learning algorithm method, by obtaining the characteristic value of the rectangular frame area and the characteristic value of the intermediate time point of the vehicle's unblocked part, the similarity is calculated to judge the time interval of the vehicle's movement event, and automatically correct the entry and exit time point. Until the set threshold is reached.
It has achieved automated correction of vehicle entry and exit orders, improved parking management efficiency, reduced the need for manual intervention, and improved operational efficiency.
Smart Images

Figure CN114648380B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of intelligent vehicle identification, and in particular to a method and device for correcting entry and exit orders based on a deep learning algorithm. Background Art
[0002] In recent years, with the continuous development of social economy and the improvement of people's living standards, the number of motor vehicles in my country has increased year by year. As the scale of cities expands, managing parking issues is particularly important for the healthy development of urban transportation. Using traditional manual methods to manage parking issues has high operating costs and low efficiency. Therefore, smart parking management methods based on Internet + artificial intelligence technology have begun to develop rapidly.
[0003] Currently, the smart parking management method of Internet + artificial intelligence has replaced the traditional manual recording of vehicle entering and exiting parking spaces. The equipment monitors parking events to automatically generate parking orders. Although this intelligent processing method is efficient, it cannot be completely accurate. When an erroneous order occurs, in the face of complaints from parking users, manual customer service intervention is still required to repair the erroneous order. This order-making measure is inefficient and labor-intensive. Summary of the invention
[0004] The purpose of the present invention is to solve the problem of low efficiency caused by manually processing user entry and exit complaint orders in the existing vehicle parking management.
[0005] To achieve the above object, the present invention provides a method for correcting vehicle entry and exit order complaints, the method comprising:
[0006] S1: Obtain the rectangular frame area of the unobstructed part of the complaint vehicle as the basic area, and extract the feature value of the basic area as the initial feature value of the vehicle;
[0007] S2: Obtain the middle time point between the entry time and the exit time of the complaint order, and retrieve the characteristic value of the middle time point in the basic area as the characteristic value of the middle time point;
[0008] S3: Calculate the similarity between the basic area feature value and the intermediate point feature value; determine the time interval of the vehicle movement event according to the similarity;
[0009] S4: If the time interval of the vehicle movement event is less than the set threshold, the intermediate time point is corrected to be the correct entry / exit time point.
[0010] As a further improvement to the above scheme, if the time interval of the vehicle movement event is greater than the set threshold, the similarity between the intermediate point feature value and the initial feature value is re-obtained to narrow the time interval of the vehicle movement time until the time interval of the vehicle movement event is less than the set threshold.
[0011] As a further improvement to the above solution, the steps of obtaining the rectangular frame area of the unobstructed part of the complaint vehicle as the basic area and extracting the feature value of the basic area include:
[0012] Obtain basic close-up images based on the type of entry and exit complaints and the start and end time of the complaint order;
[0013] If the complaint entry time is wrong, a close-up image of the parking space where the vehicle has parked steadily at a certain time point before the exit time is obtained as the basic close-up image, and the basic area feature value is extracted as the initial feature value;
[0014] If the complaint about the exit time is wrong, a close-up image of the parking space where the vehicle has parked steadily at a certain time point after the entry time is obtained as the basic close-up image, and the basic area feature values are extracted as the initial feature values.
[0015] As a further improvement to the above scheme, the similarity between the basic area feature value and the intermediate point feature value is calculated; and the step of determining the time interval of the vehicle movement event according to the similarity includes:
[0016] If the similarity is greater than a certain threshold, the vehicle has not moved;
[0017] If the complaint type is entry error, the time from the vehicle entry time to the middle time point in the order is the time interval when the vehicle movement event occurs;
[0018] If the complaint type is departure error, the time from the intermediate time point to the vehicle departure time in the order is the time interval in which the vehicle movement event occurs.
[0019] As a further improvement to the above solution, the method further includes:
[0020] If the similarity is less than a certain threshold, the vehicle moves;
[0021] If the complaint type is entry error, the time between the intermediate time point and the vehicle exit time in the order is the time interval in which the vehicle movement event occurs;
[0022] If the complaint type is exit error, the time from the vehicle entry time to the middle time point in the order is the time interval when the vehicle movement event occurs.
[0023] The present invention also discloses a device for correcting complaints about vehicle entry and exit orders, the device comprising:
[0024] An acquisition module is used to acquire the rectangular frame area of the unobstructed part of the complaint vehicle as a basic area, and extract the feature value of the basic area as an initial feature value;
[0025] It is used to obtain the middle time point between the entry time and exit time of the complaint order, retrieve the close-up image of the middle time point, and extract the feature value in the basic area of the close-up image as the feature value of the middle time point;
[0026] A judgment module is used to calculate the similarity between the initial eigenvalue and the intermediate point eigenvalue;
[0027] A selection module is used to determine the time interval of the vehicle movement event according to the similarity;
[0028] The determination module is used to determine that the intermediate time point is the correct entry / exit time point if the time interval of the vehicle movement event is less than a set threshold.
[0029] As a further improvement to the above solution, the acquisition module is further used to:
[0030] Obtain basic close-up images based on the type of entry and exit complaints and the start and end time of the complaint order;
[0031] If the complaint entry time is wrong, a close-up image of the parking space where the vehicle has parked steadily at a certain time point before the exit time is obtained as the basic close-up image, and the basic area feature value is extracted as the initial feature value;
[0032] If the complaint about the exit time is wrong, a close-up image of the parking space where the vehicle has parked steadily at a certain time point after the entry time is obtained as the basic close-up image, and the basic area feature values are extracted as the initial feature values.
[0033] As a further improvement to the above solution, the selection module is further used to calculate the similarity between the initial feature value and the intermediate point feature value; and the step of determining the time interval of the vehicle movement event according to the similarity includes:
[0034] If the similarity is greater than a certain threshold, the vehicle has not moved;
[0035] If the complaint type is entry error, the time from the vehicle entry time to the middle time point in the order is the time interval when the vehicle movement event occurs;
[0036] If the complaint type is departure error, the time from the intermediate time point to the vehicle departure time in the order is the time interval in which the vehicle movement event occurs.
[0037] As a further improvement to the above solution, the selection module is further used for:
[0038] If the similarity is less than a certain threshold, the vehicle moves;
[0039] If the complaint type is entry error, the time between the intermediate time point and the vehicle exit time in the order is the time interval in which the vehicle movement event occurs;
[0040] If the complaint type is exit error, the time from the vehicle entry time to the middle time point in the order is the time interval when the vehicle movement event occurs.
[0041] By adopting the entry and exit order correction method and device based on deep learning algorithm provided by the present invention, the features of the complaint vehicle image can be extracted through the feature extraction algorithm based on deep learning, and then the basic area features of other intermediate time points can be calculated. It can be determined that a vehicle movement event occurred at a certain intermediate time point, thereby confirming the correct vehicle entry and exit time of the complaint order, and automatically repairing the complaint order of the parking user. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] Figure 1 It is a flow chart of a method for correcting parking space entry and exit complaint orders based on a deep learning algorithm provided by the present invention;
[0043] Figure 2 It is a structural schematic diagram of a parking space entry and exit complaint order correction device based on a deep learning algorithm provided by the present invention;
[0044] Figure 3 It is a schematic diagram of a close-up image of a vehicle complained about in a selected parking space provided by the present invention;
[0045] Figure 4 It is a schematic diagram of the basic area of the complaint vehicle on the selected parking space provided by the present invention;
[0046] Figure 5 It is a schematic diagram of selecting the time interval for the next feature comparison provided by the present invention. DETAILED DESCRIPTION
[0047] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments.
[0048] The present invention provides a method and device for correcting parking space entry and exit complaint orders based on a deep learning algorithm. According to the type of entry and exit complaints and the start and end time of the complaint order, basic close-up images of vehicles in the order are obtained and processed accordingly. Specifically, Figure 1 As shown, the method includes:
[0049] S1: Obtain the rectangular frame area of the unobstructed part of the complaint vehicle in the basic close-up image as the basic area, and extract the feature value of the basic area as the initial feature value of the vehicle;
[0050] Specifically, the steps of obtaining the unobstructed rectangular frame area of the complained vehicle as the basic area and extracting the basic area feature value include: first, obtaining the basic close-up image according to the type of entry and exit complaints and the start and end time of the complaint order; if the complained entry time is wrong, then obtaining a close-up image of the parking space where the vehicle has parked steadily at a certain time point before the exit time point as the basic close-up image, and extracting the basic area feature value as the initial feature value; if the complained exit time is wrong, then obtaining a close-up image of the parking space where the vehicle has parked steadily at a certain time point after the entry time point as the basic close-up image, and extracting the basic area feature value as the initial feature value.
[0051] S2: Select the middle time point between the entry time and the exit time of the complaint order, obtain the close-up image of the middle time point, and extract the feature value of the middle time point in the basic area of the close-up image as the feature value of the middle time point;
[0052] S3: Calculate the similarity between the basic area feature value and the intermediate point feature value; determine the time interval of the vehicle movement event according to the similarity;
[0053] S4: If the time interval of the vehicle movement event is less than the set threshold, the intermediate time point is corrected to be the correct entry / exit time point.
[0054] If the time interval of the vehicle movement event is greater than the set threshold, the similarity between the intermediate point feature value and the initial feature value is re-obtained to reduce the time interval of the vehicle movement time until the time interval of the vehicle movement event is less than the set threshold.
[0055] Specifically, Figure 3-5 As shown in the figure, taking the complaint exit error as an example, the initial feature is compared with the feature similarity of the midpoint 1. If the features are similar, it means that the vehicle did not move in the time interval from the start time of the order to the midpoint 1, and it can be inferred that the vehicle movement event occurred in the time interval from the midpoint 1 to the end of the order.
[0056] Further, the time interval from the intermediate time point 1 to the end of the order is taken as the time interval for the occurrence of the vehicle movement event, and the time point between the intermediate time point 1 and the end of the order is selected as the intermediate time, that is, the intermediate time point 2; the feature value in the basic area of the intermediate time point 2 is extracted, and the similarity is compared with the initial feature. At this time, the features are not similar, and it is inferred that the vehicle movement event occurred in the time interval from the intermediate time point 1 to the intermediate time point 2, and the time point between the intermediate time point 1 and the intermediate time point 2 is continued to be selected as the intermediate time 3; the feature value in the basic area of the intermediate time point 3 is continued to be extracted, and the similarity is compared with the initial feature. At this time, the features are not similar, and it is inferred that the vehicle movement event occurred in the time interval from the intermediate time point 1 to the intermediate time point 3, and the time point between the intermediate time point 1 and the intermediate time point 3 is selected as the intermediate time point 4; the feature value in the basic area of the intermediate time point 4 is continued to be extracted, and the similarity is compared with the basic feature. At this time, the features are not similar, and the time interval at this time is less than a certain threshold, and the intermediate time point 4 is the correct exit time point.
[0057] Similarly, a similar approach is used when making complaints.
[0058] like Figure 2 As shown, the present invention also discloses a parking space entry and exit complaint order correction device based on a deep learning algorithm, the device comprising:
[0059] An acquisition module is used to acquire the rectangular frame area of the unobstructed part of the complaint vehicle as a basic area, and extract the feature value of the basic area as an initial feature value;
[0060] It is used to obtain the middle time point between the entry time and exit time of the complaint order, retrieve the close-up image of the middle time point, and extract the feature value in the basic area of the close-up image as the feature value of the middle time point;
[0061] A judgment module is used to calculate the similarity between the initial eigenvalue and the intermediate point eigenvalue;
[0062] A selection module is used to determine the time interval of the vehicle movement event according to the similarity;
[0063] The determination module is used to determine that the intermediate time point is the correct entry / exit time point if the time interval of the vehicle movement event is less than a set threshold.
[0064] As a further improvement to the above solution, the acquisition module is further used to:
[0065] The acquisition module is also used to, if the time interval of the vehicle movement event is greater than a set threshold, re-acquire the similarity between the intermediate point feature value and the initial feature value, and reduce the time interval of the vehicle movement time until the time interval of the vehicle movement event is less than the set threshold;
[0066] Obtain basic close-up images based on the type of entry and exit complaints and the start and end time of the complaint order;
[0067] If the complaint entry time is wrong, a close-up image of the parking space where the vehicle has parked steadily at a certain time point before the exit time is obtained as the basic close-up image, and the basic area feature value is extracted as the initial feature value;
[0068] If the complaint about the exit time is wrong, a close-up image of the parking space where the vehicle has parked steadily at a certain time point after the entry time is obtained as the basic close-up image, and the basic area feature values are extracted as the initial feature values.
[0069] As a further improvement to the above solution, the selection module is further used to calculate the similarity between the initial feature value and the intermediate point feature value; and the step of determining the time interval of the vehicle movement event according to the similarity includes:
[0070] If the similarity is greater than a certain threshold, the vehicle has not moved;
[0071] If the complaint type is entry error, the time from the vehicle entry time to the middle time point in the order is the time interval when the vehicle movement event occurs;
[0072] If the complaint type is departure error, the time from the intermediate time point to the vehicle departure time in the order is the time interval in which the vehicle movement event occurs.
[0073] As a further improvement to the above solution, the selection module is further used for:
[0074] If the similarity is less than a certain threshold, the vehicle moves;
[0075] If the complaint type is entry error, the time between the intermediate time point and the vehicle exit time in the order is the time interval in which the vehicle movement event occurs;
[0076] If the complaint type is exit error, the time from the vehicle entry time to the middle time point in the order is the time interval when the vehicle movement event occurs.
[0077] As a further improvement to the above scheme, the selection module is also used to re-acquire the intermediate time point if the time interval of the vehicle movement event is greater than a set threshold, extract the new intermediate time point feature value, and compare the intermediate time point feature value with the initial feature value again to calculate the similarity, and determine the new time interval of the vehicle movement based on the similarity.
[0078] As a further improvement to the above solution, the determining module is further used to:
[0079] Repeatedly obtain the time interval of the vehicle movement time until the time interval of the vehicle movement event is less than a certain threshold, and determine whether the vehicle movement event is found based on the threshold. If so, correct the intermediate time point to the correct exit / entry time.
[0080] The parking space entry and exit order revision method and device based on deep learning algorithm provided by the present invention can solve the problem of user entry and exit complaint orders in vehicle parking management and the low efficiency of manual order filling.
[0081] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A method for correcting complaints about vehicle entry and exit orders, characterized in that: The method comprises: S1: Obtain the rectangular frame area of the unobstructed part of the complaint vehicle as the basic area, and extract the feature value of the basic area as the initial feature value of the vehicle; S2: Select the middle time point between the entry time and the exit time of the complaint order, and obtain the characteristic value of the middle time point in the basic area as the characteristic value of the middle time point; S3: Calculate the similarity between the initial eigenvalue and the intermediate point eigenvalue; determine the time interval of the vehicle movement event based on the similarity; The step of calculating the similarity between the initial feature value and the intermediate feature value; judging the time interval of the vehicle movement event according to the similarity includes: if the similarity is greater than a certain threshold, the vehicle has not moved; if the complaint type is an entry error, the time interval from the vehicle entry time in the order to the intermediate time point is the time interval of the vehicle movement event; if the complaint type is an exit error, the time interval from the intermediate time point to the vehicle exit time in the order is the time interval of the vehicle movement event; if the similarity is less than a certain threshold, the vehicle has moved; if the complaint type is an entry error, the time interval from the intermediate time point to the vehicle exit time in the order is the time interval of the vehicle movement event; if the complaint type is an exit error, the time interval from the vehicle entry time in the order to the intermediate time point is the time interval of the vehicle movement event; S4: If the time interval of the vehicle movement event is less than the set threshold, the intermediate time point is corrected to be the correct entry / exit time point; If the time interval of the vehicle movement event is greater than the set threshold, the similarity between the intermediate point feature value and the initial feature value is re-obtained to reduce the time interval of the vehicle movement time until the time interval of the vehicle movement event is less than the set threshold.
2. The method for correcting vehicle entry and exit order complaints according to claim 1, characterized in that: The steps of obtaining the rectangular frame area of the unobstructed part of the complaint vehicle as the basic area, and extracting the basic area feature value as the initial feature value of the vehicle include: obtaining a basic close-up image according to the type of entry and exit complaints and the start and end time of the complaint order; if the complaint entry time is wrong, obtaining a close-up image of the parking space where the vehicle has stopped steadily at a certain time point before the exit time point as the basic close-up image, and extracting the basic area feature value as the initial feature value; if the complaint exit time is wrong, obtaining a close-up image of the parking space where the vehicle has stopped steadily at a certain time point after the entry time point as the basic close-up image, and extracting the basic area feature value as the initial feature value.
3. A device for correcting complaints about vehicle entry and exit orders, characterized in that: The device includes: an acquisition module, which is used to acquire the rectangular frame area of the unobstructed part of the complaint vehicle as the basic area, and extract the basic area feature value as the initial feature value, which is used as the initial feature value of the vehicle; it is used to acquire the middle time point between the entry time and the exit time of the complaint order, retrieve the close-up image of the middle time point, and extract the feature value in the basic area in the close-up image as the feature value of the middle time point; a judgment module, which is used to calculate the similarity between the initial feature value and the middle point feature value; a selection module, which is used to judge the time interval of the vehicle movement event according to the similarity; a determination module, which is used to determine that the middle time point is the correct entry / exit time point if the time interval of the vehicle movement event is less than a set threshold; The acquisition module is also used to, if the time interval of the vehicle movement event is greater than a set threshold, re-acquire the similarity between the intermediate point feature value and the initial feature value, and reduce the time interval of the vehicle movement time until the time interval of the vehicle movement event is less than the set threshold; The selection module is also used to calculate the similarity between the initial feature value and the intermediate feature value; the step of judging the time interval of the vehicle movement event according to the similarity includes: if the similarity is greater than a certain threshold, the vehicle has not moved; if the complaint type is an entry error, the time interval from the vehicle entry time in the order to the intermediate time point is the time interval of the vehicle movement event; if the complaint type is an exit error, the time interval from the intermediate time point to the vehicle exit time in the order is the time interval of the vehicle movement event; The selection module is also used for: if the similarity is less than a certain threshold, the vehicle moves; if the complaint type is an entry error, the time from the middle time point to the vehicle exit time in the order is the time interval for the vehicle movement event; if the complaint type is an exit error, the time from the vehicle entry time in the order to the middle time point is the time interval for the vehicle movement event.
4. The device for correcting vehicle entry and exit order complaints according to claim 3; characterized in that: The acquisition module is also used to acquire a basic close-up image according to the type of entry and exit complaints and the start and end time of the complaint order; if the complaint entry time is wrong, a close-up image of the parking space where the vehicle has parked steadily at a certain time point before the exit time is acquired as the basic close-up image, and basic area feature values are extracted as initial feature values; If the complaint about the exit time is wrong, a close-up image of the parking space where the vehicle has parked steadily at a certain time point after the entry time is obtained as the basic close-up image, and the basic area feature values are extracted as the initial feature values.
Citation Information
Patent Citations
Mobile platform and method for mobile platform long-time tracking
CN109398533A
Geomagnetic positioning and intelligent terminal construction method based on block chain traceable technology
CN112710324A
Human-vehicle integration algorithm based on hidden Markov model within time threshold
CN112882073A
Method, system and platform for assisting in correcting license plate
CN113205089A
Intelligent parking management implementation system and method
CN113706926A
