CN116013072B - Parking lot vehicle entering and exiting counting method based on deep learning application - Google Patents
Parking lot vehicle entering and exiting counting method based on deep learning application Download PDFInfo
- Publication number
- CN116013072B CN116013072B CN202310003722.1A CN202310003722A CN116013072B CN 116013072 B CN116013072 B CN 116013072B CN 202310003722 A CN202310003722 A CN 202310003722A CN 116013072 B CN116013072 B CN 116013072B
- Authority
- CN
- China
- Prior art keywords
- parking lot
- matrix
- result
- vehicle
- video stream
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000013135 deep learning Methods 0.000 title claims abstract description 15
- 238000012545 processing Methods 0.000 claims abstract description 24
- 239000011159 matrix material Substances 0.000 claims description 83
- 238000001514 detection method Methods 0.000 claims description 25
- 238000012216 screening Methods 0.000 claims description 23
- 230000008859 change Effects 0.000 claims description 9
- 230000004927 fusion Effects 0.000 claims description 9
- 238000010586 diagram Methods 0.000 claims description 8
- 238000001914 filtration Methods 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 6
- 238000012937 correction Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000000638 solvent extraction Methods 0.000 claims description 3
- 238000013528 artificial neural network Methods 0.000 abstract description 3
- 230000008569 process Effects 0.000 description 6
- 230000006698 induction Effects 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000004888 barrier function Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000006378 damage Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention relates to the technical field of neural networks, in particular to a parking lot vehicle in-out counting method based on deep learning application. Step 1: acquiring video streams at the entrance and exit of a parking lot in real time, and judging whether vehicles enter the parking lot or whether vehicles leave the parking lot based on the acquired video streams; step 2: acquiring a video stream in a parking lot in real time, counting vehicles in the current parking lot based on the acquired video stream, and generating a third result; step 3: calculating a charge for the vehicle leaving the parking lot based on the second result; calculating the number of the remaining parking spaces in the parking lot based on the first result, the second result and the third result; step 4: and judging whether the new vehicle is allowed to enter the parking lot or not according to the calculated number of the remaining parking spaces in the parking lot. The invention realizes the automation of parking lot management through video image processing, combines three-dimensional video stream and morphology when video processing is carried out, and improves the accuracy of identification.
Description
Technical Field
The invention belongs to the technical field of neural networks, and particularly relates to a parking lot vehicle in-out counting method based on deep learning application.
Background
With the increasing improvement of people's life, vehicles have become one of the necessary tools for traveling. Parking lots are also an indispensable part of daily life. In the process of operating and maintaining a parking lot, a ground induction coil is often used at the present stage as a standard scheme for counting vehicles entering and exiting. However, this solution requires a large amount of construction and requires the destruction of the subsurface coil.
Meanwhile, in the operation and maintenance process, the ground induction coil cannot achieve a very good effect for vehicle counting, and the ground induction coil is mainly caused by the following reasons: 1. the subjective cause is: a. a personal condition vehicle remotely controlling the opening of the gate to come in and go out; b. cash or personal WeChat is collected, and the cash or personal WeChat is illegally paid to open the gate; 2. following the car to drive in and out the fee evasion; 3. parking charge management software defects; 4. a lane control system failure; 5. identifying the camera and external factors: a. the license plate is paint-dropped and dirty, so that the identification is wrong; b. the license plate recognition camera is aged, and the acquired image is unclear; c. the algorithm of the recognition camera is not updated, and the new license plate is not recognized; d. the rain and snow weather recognition effect is poor, and the vehicle has no entry record; 6. digital barrier failure: a. motor driving problem, mechanical jamming; b. the tension of the spring is too large, and the lever is automatically lifted illegally; c. the vehicle detector crashes, and the brake lever does not rise and fall; d. the barrier is arranged under the brake bar, and the brake bar rebounds when meeting the blockage.
Disclosure of Invention
The invention mainly aims to provide a parking lot vehicle in-out counting method based on deep learning application.
In order to achieve the above purpose, the technical scheme of the invention is realized as follows:
a parking lot in-out vehicle counting method based on a deep learning application, the method performing the steps of:
step 1: acquiring video streams at the entrance and the exit of a parking lot in real time, judging whether vehicles enter the parking lot or not based on the acquired video streams, and if yes, generating a first result; meanwhile, judging whether a vehicle leaves the parking lot or not based on the acquired video stream, and if so, generating a second result;
step 2: acquiring video streams in a parking lot in real time, counting vehicles in the current parking lot based on the acquired video streams, and generating a third result, wherein the method specifically comprises the following steps: acquiring video streams in real time through a full-field camera formed by a plurality of distributed sub-cameras, and performing three-dimensional video stream splicing on the video streams acquired by each sub-camera to acquire three-dimensional video streams; carrying out overall boundary detection on all vehicles in the parking lot based on the three-dimensional video stream to obtain a boundary detection result, and carrying out feature-based boundary deepening treatment on the boundary detection result to obtain a deepening treatment result; based on the deepened processing result, calculating the number of vehicles in the parking lot, and generating a third result;
step 3: calculating a charge for the vehicle leaving the parking lot based on the second result; based on the first result, the second result and the third result, calculating the number of the residual parking spaces in the parking lot by using a preset fusion counting algorithm;
step 4: and the lifting rod system judges whether the new vehicle is allowed to enter the parking lot or not according to the calculated number of the remaining parking spaces in the parking lot based on the first result.
Further, the method for determining whether the vehicle enters the parking lot based on the acquired video stream in the step 1 includes: randomly screening a plurality of different frames from the acquired video stream, and overlapping the different frames to obtain an overlapped frame image; performing image region screening on the overlapped frame images to obtain a screening region; and judging the vehicle direction of the screening area, so as to judge whether a vehicle enters the parking lot.
Further, the method for randomly screening out a plurality of different frames from the acquired video stream comprises the following steps: a plurality of time results are calculated using the following formula:and then respectively acquiring frames of the time results corresponding to the time in the video stream.
Further, the method for filtering the image areas of the overlapped frame images to obtain the filtering areas comprises the following steps: pre-partitioning the overlapping frames into a super-pixel set; calculating the pixel change rate of the super pixel set, and obtaining a horizontal movement velocity matrix and a vertical movement velocity matrix of the pixels according to the pixel change rate; obtaining a velocity amplitude matrix of each pixel point according to the horizontal motion velocity matrix and the vertical motion velocity matrix; generating a vehicle morphology matrix; performing matrix multiplication calculation on the vehicle morphology matrix and the horizontal movement speed matrix, the vertical movement speed matrix and the speed amplitude matrix respectively to obtain a horizontal movement speed morphology matrix, a vertical movement speed morphology matrix and a speed amplitude morphology matrix respectively; and (3) performing matrix connection on the horizontal movement speed morphology matrix, the vertical movement speed morphology matrix and the speed amplitude morphology matrix to obtain a connection matrix, and then solving the rank of the connection matrix, wherein the area corresponding to the part which is not 0 in the rank of the matrix is the screening area.
Further, the matrix expression of the vehicle morphology matrix is that
Wherein a is 1 Is a matrix element of a vehicle morphology matrix.
Further, the method for determining the vehicle direction of the screening area comprises the following steps: and calculating to obtain the gradient of the matrix pixel change rate, wherein the direction of the gradient is the vehicle direction.
Further, the method for determining whether the vehicle leaves the parking lot based on the acquired video stream in the step 1 includes: and judging whether the vehicle leaves the parking lot or not according to the direction of the vehicle.
Further, in the step 2, the method for performing three-dimensional video stream splicing on the video streams acquired by each sub-camera to obtain the three-dimensional video stream includes: acquiring a three-dimensional space structure diagram of a parking lot, and generating a three-dimensional space box with consistent form based on the three-dimensional space structure diagram; and converting the video stream shot by each sub-camera into a pixel video stream, projecting the pixel video stream to a position corresponding to the shooting area of the sub-camera in the three-dimensional space box, and then performing time synchronization to obtain the three-dimensional video stream.
Further, the method for obtaining the deepened result of the boundary based on the feature comprises the steps of: and carrying out boundary detection on vehicles in each frame in the three-dimensional video stream to obtain a boundary detection result, carrying out feature extraction and Kalman filtering on the boundary detection result respectively, and carrying out fusion of motion features and vehicle features on the feature extraction and Kalman filtering result to obtain a deepened processing result.
Further, the method for calculating the number of vehicles in the parking lot and generating the third result based on the deepened processing result comprises the following steps: and processing the deepened processing result by using a Hungary matching algorithm to obtain a third result.
The parking lot vehicle in-out counting method based on deep learning application has the following beneficial effects: compared with the prior art, the invention creatively uses the video stream processing based on the whole and the three-dimensional stereo, can obviously improve the accuracy rate for the three-dimensional stereo video stream processing, and can improve the efficiency for the whole processing, thereby reducing the efficiency reduction brought by the three-dimensional stereo video stream processing to a certain extent.
Drawings
Fig. 1 is a schematic flow diagram of a method for counting vehicles entering and exiting a parking lot based on deep learning application according to an embodiment of the present invention;
fig. 2 is a detailed flow chart of a parking lot vehicle in-out counting method based on deep learning application according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a third result of a parking lot vehicle entering and exiting counting method based on deep learning application according to an embodiment of the present invention
Fig. 4 is a schematic flow chart of intelligent counting of vehicle data in a parking lot based on a method for counting vehicles entering and exiting the parking lot by deep learning application according to an embodiment of the present invention.
Detailed Description
The method of the present invention will be described in further detail with reference to the accompanying drawings.
Example 1
As shown in fig. 1, 2, 3 and 4, a parking lot in-out vehicle counting method based on a deep learning application performs the steps of:
step 1: acquiring video streams at the entrance and the exit of a parking lot in real time, judging whether vehicles enter the parking lot or not based on the acquired video streams, and if yes, generating a first result; meanwhile, judging whether a vehicle leaves the parking lot or not based on the acquired video stream, and if so, generating a second result;
step 2: acquiring video streams in a parking lot in real time, counting vehicles in the current parking lot based on the acquired video streams, and generating a third result, wherein the method specifically comprises the following steps: acquiring video streams in real time through a full-field camera formed by a plurality of distributed sub-cameras, and performing three-dimensional video stream splicing on the video streams acquired by each sub-camera to acquire three-dimensional video streams; carrying out overall boundary detection on all vehicles in the parking lot based on the three-dimensional video stream to obtain a boundary detection result, and carrying out feature-based boundary deepening treatment on the boundary detection result to obtain a deepening treatment result; based on the deepened processing result, calculating the number of vehicles in the parking lot, and generating a third result;
step 3: calculating a charge for the vehicle leaving the parking lot based on the second result; based on the first result, the second result and the third result, calculating the number of the residual parking spaces in the parking lot by using a preset fusion counting algorithm;
step 4: and the lifting rod system judges whether the new vehicle is allowed to enter the parking lot or not according to the calculated number of the remaining parking spaces in the parking lot based on the first result.
Specifically, an approach camera is arranged at the entrance and exit of the parking lot, and a full-field camera is arranged in the parking lot. The method and the system can quickly and effectively calculate the number of vehicles entering and exiting the parking space based on video stream analysis and compare the number with the number of vehicles parked in the whole parking space; the full-field camera can also be used for predicting which vehicles want to leave or search for a parking space while calculating the number of the full-field vehicles; according to the number of the existing vehicles and the number of the parking spaces, the number of the vehicles which can be accommodated in the parking lot can be calculated; and according to comprehensive comparison information such as the characteristics of the approaching vehicles, the identification of the exiting license plates of the vehicles and the like, the charging accuracy of the vehicles is improved, and the cost is reduced.
Specifically, in the fusion counting algorithm, a fusion result is obtained by using a preset weighting coefficient and a preset weighting algorithm based on a first result, a second result and a third result, and then the fusion result is multiplied by a correction coefficient of the parking quantity in the parking lot to obtain the quantity of the remaining parking spaces in the parking lot.
In practice, since the number of parks in a parking lot varies in each period, some time points are peak periods and some time points are low peaks. By obtaining the history data, a correction factor can be obtained.
Example 2
On the basis of the above embodiment, the method for determining whether a vehicle enters the parking lot based on the acquired video stream in step 1 includes: randomly screening a plurality of different frames from the acquired video stream, and overlapping the different frames to obtain an overlapped frame image; performing image region screening on the overlapped frame images to obtain a screening region; and judging the vehicle direction of the screening area, so as to judge whether a vehicle enters the parking lot.
Specifically, the process of overlapping frames essentially synthesizes a plurality of images into one image, and not only can the screening area be well reached by one synthesized image, but also the traveling direction of the vehicle can be recognized. In the process of identifying the direction, the method is realized through pixel gradient, and compared with the prior art, the method for judging the image identification has the advantage that the efficiency is obviously higher.
Example 3
Based on the above embodiment, the method for randomly screening a plurality of different frames from the acquired video stream includes: a plurality of time results are calculated using the following formula: and then respectively acquiring frames of the time results corresponding to the time in the video stream.
Example 4
On the basis of the above embodiment, the method for performing image region screening on the overlapped frame image to obtain a screened region includes: pre-partitioning the overlapping frames into a super-pixel set; calculating the pixel change rate of the super pixel set, and obtaining a horizontal movement velocity matrix and a vertical movement velocity matrix of the pixels according to the pixel change rate; obtaining a velocity amplitude matrix of each pixel point according to the horizontal motion velocity matrix and the vertical motion velocity matrix; generating a vehicle morphology matrix; performing matrix multiplication calculation on the vehicle morphology matrix and the horizontal movement speed matrix, the vertical movement speed matrix and the speed amplitude matrix respectively to obtain a horizontal movement speed morphology matrix, a vertical movement speed morphology matrix and a speed amplitude morphology matrix respectively; and (3) performing matrix connection on the horizontal movement speed morphology matrix, the vertical movement speed morphology matrix and the speed amplitude morphology matrix to obtain a connection matrix, and then solving the rank of the connection matrix, wherein the area corresponding to the part which is not 0 in the rank of the matrix is the screening area.
Specifically, the invention can also use a neural network-based deep learning method for the video stream acquired at the approach camera, and can improve the accuracy and the effectiveness of the vehicle in and out counting of the parking lot and provide evidence for pursuit under the condition of no damage or large-scale construction; light-weight deployment of a complete scheme and algorithm implementation for accurately counting vehicles entering and exiting a parking lot based on a deep learning algorithm; effectively predicting how many vehicles can be parked in the parking lot; through video detection, which vehicles are to be out of the scene can be predicted; when the vehicle comes out, the vehicle is matched with the incoming data to predict the coming out.
Specifically, based on the invention, the vehicle can be effectively counted in and out of the parking lot. Not only can avoid the inaccurate problem of count of traditional based on ground induction coil, but also can avoid the erroneous problem of count that the human factor leads to simultaneously. The automatic counting can be achieved, meanwhile, the video stream images can be used as statistical basis and accordingly serve as a base. The scheme has the advantage of convenience, and the counting can be completed by adding a small number of cameras under the condition of not damaging the layout scheme of the original parking lot. And simultaneously, predicting which vehicles in the parking lot want to leave and how many vehicles are in the parking lot. When the vehicle comes out, the vehicle is matched with the coming-in recognition feature according to the coming-in recognition feature, so that accurate charging is realized, and erroneous recognition is avoided to a greater extent.
Example 5
On the basis of the above embodiment, the matrix expression of the vehicle morphology matrix is
Wherein a is 1 Is a matrix element of a vehicle morphology matrix.
In particular, the vehicle morphology matrix is based on a matrix constructed by vehicle morphology, and the matrix in the invention is not a matrix in the traditional sense, but is a constructed array, but the arrays follow the operation mode of the matrix.
Example 6
On the basis of the above embodiment, the method for determining the vehicle direction in the screening area includes: and calculating to obtain the gradient of the matrix pixel change rate, wherein the direction of the gradient is the vehicle direction.
Example 7
On the basis of the above embodiment, the method for determining whether the vehicle leaves the parking lot based on the acquired video stream in step 1 includes: and judging whether the vehicle leaves the parking lot or not according to the direction of the vehicle.
Example 8
On the basis of the above embodiment, the method for performing three-dimensional video stream splicing on the video streams acquired by each sub-camera in the step 2 to obtain a three-dimensional video stream includes: acquiring a three-dimensional space structure diagram of a parking lot, and generating a three-dimensional space box with consistent form based on the three-dimensional space structure diagram; and converting the video stream shot by each sub-camera into a pixel video stream, projecting the pixel video stream to a position corresponding to the shooting area of the sub-camera in the three-dimensional space box, and then performing time synchronization to obtain the three-dimensional video stream.
Example 9
On the basis of the above embodiment, the method for obtaining the deepened processing result of the boundary based on the three-dimensional video stream for carrying out the overall boundary detection on all vehicles in the parking lot to obtain the boundary detection result, and then carrying out the boundary deepening processing based on the characteristics on the boundary detection result comprises the following steps: and carrying out boundary detection on vehicles in each frame in the three-dimensional video stream to obtain a boundary detection result, carrying out feature extraction and Kalman filtering on the boundary detection result respectively, and carrying out fusion of motion features and vehicle features on the feature extraction and Kalman filtering result to obtain a deepened processing result.
Example 10
On the basis of the above embodiment, the method for calculating the number of vehicles in the parking lot and generating the third result based on the deepened processing result includes: and processing the deepened processing result by using a Hungary matching algorithm to obtain a third result.
In the several embodiments provided by the present invention, it should be understood that the disclosed apparatus and method may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed over a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in hardware plus software functional units.
The integrated units implemented in the form of software functional units described above may be stored in a computer readable storage medium. The software functional unit is stored in a storage medium, and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor (processor) to perform part of the steps of the methods according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional modules is illustrated, and in practical application, the above-described functional allocation may be performed by different functional modules according to needs, i.e. the internal structure of the apparatus is divided into different functional modules to perform all or part of the functions described above. The specific working process of the above-described device may refer to the corresponding process in the foregoing method embodiment, which is not described herein again.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.
Claims (7)
1. A method for counting vehicles entering and exiting a parking lot based on deep learning application, which is characterized by comprising the following steps:
step 1: acquiring video streams at the entrance and the exit of a parking lot in real time, judging whether vehicles enter the parking lot or not based on the acquired video streams, and if yes, generating a first result; judging whether a vehicle leaves the parking lot or not based on the acquired video stream, and if so, generating a second result;
step 2: acquiring video streams in a parking lot in real time, counting vehicles in the current parking lot based on the acquired video streams, and generating a third result, wherein the method specifically comprises the following steps: acquiring video streams in real time through a full-field camera formed by a plurality of distributed sub-cameras, and performing three-dimensional video stream splicing on the video streams acquired by each sub-camera to acquire three-dimensional video streams; carrying out overall boundary detection on all vehicles in the parking lot based on the three-dimensional video stream to obtain a boundary detection result, and carrying out feature-based boundary deepening treatment on the boundary detection result to obtain a deepening treatment result; based on the deepened processing result, calculating the number of vehicles in the parking lot, and generating a third result;
step 3: calculating a charge for the vehicle leaving the parking lot based on the second result; based on the first result, the second result and the third result, calculating the number of the residual parking spaces in the parking lot by using a preset fusion counting algorithm;
step 4: the lifting rod system judges whether a new vehicle is allowed to enter the parking lot or not according to the calculated number of the remaining parking spaces in the parking lot based on the first result;
the method for judging whether the vehicle enters the parking lot or not based on the acquired video stream in the step 1 comprises the following steps: randomly screening a plurality of different frames from the acquired video stream, and overlapping the different frames to obtain an overlapped frame image; performing image region screening on the overlapped frame images to obtain a screening region; carrying out vehicle direction judgment on the screening area, so as to judge whether a vehicle enters a parking lot or not;
the method for screening the image area of the overlapped frame image to obtain the screened area comprises the following steps: pre-partitioning the overlapping frames into a super-pixel set; calculating the pixel change rate of the super pixel set, and obtaining a horizontal movement velocity matrix and a vertical movement velocity matrix of the pixels according to the pixel change rate; obtaining a velocity amplitude matrix of each pixel point according to the horizontal motion velocity matrix and the vertical motion velocity matrix; generating a vehicle morphology matrix; performing matrix multiplication calculation on the vehicle morphology matrix and the horizontal movement speed matrix, the vertical movement speed matrix and the speed amplitude matrix respectively to obtain a horizontal movement speed morphology matrix, a vertical movement speed morphology matrix and a speed amplitude morphology matrix respectively; and (3) performing matrix connection on the horizontal movement speed morphology matrix, the vertical movement speed morphology matrix and the speed amplitude morphology matrix to obtain a connection matrix, and then solving the rank of the connection matrix, wherein the area corresponding to the part which is not 0 in the rank of the matrix is the screening area.
2. The method of claim 1, wherein the matrix expression of the vehicle morphology matrix is
；
Wherein,is a matrix element of a vehicle morphology matrix.
3. The method of claim 2, wherein the method of determining the direction of the vehicle in the screening area comprises: calculating to obtain a gradient of the matrix pixel change rate, wherein the direction of the gradient is the vehicle direction; the method for judging whether the vehicle leaves the parking lot or not based on the acquired video stream in the step 1 comprises the following steps: and judging whether the vehicle leaves the parking lot or not according to the direction of the vehicle.
4. The method of claim 3, wherein the step 3 further comprises generating a correction coefficient for the number of parks of the parking lot based on a time-to-number-of-parks relationship model, and the method specifically comprises: and acquiring historical data of the time and the parking number of the target parking lot, and generating correction coefficients based on the historical data.
5. The method of claim 4, wherein the method of performing three-dimensional video stream splicing on the video streams acquired by each sub-camera in step 2 to obtain the three-dimensional video stream comprises: acquiring a three-dimensional space structure diagram of a parking lot, and generating a three-dimensional space box with consistent form based on the three-dimensional space structure diagram; and converting the video stream shot by each sub-camera into a pixel video stream, projecting the pixel video stream to a position corresponding to the shooting area of the sub-camera in the three-dimensional space box, and then performing time synchronization to obtain the three-dimensional video stream.
6. The method of claim 5, wherein the method for performing overall boundary detection on all vehicles in the parking lot based on the three-dimensional video stream to obtain a boundary detection result, and performing feature-based boundary deepening on the boundary detection result to obtain a deepened result comprises: and carrying out boundary detection on vehicles in each frame in the three-dimensional video stream to obtain a boundary detection result, carrying out feature extraction and Kalman filtering on the boundary detection result respectively, and carrying out fusion of motion features and vehicle features on the feature extraction and Kalman filtering result to obtain a deepened processing result.
7. The method of claim 6, wherein the calculating the number of vehicles in the parking lot based on the deepened processing result, the generating a third result includes: and processing the deepened processing result by using a Hungary matching algorithm to obtain a third result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310003722.1A CN116013072B (en) | 2023-01-03 | 2023-01-03 | Parking lot vehicle entering and exiting counting method based on deep learning application |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310003722.1A CN116013072B (en) | 2023-01-03 | 2023-01-03 | Parking lot vehicle entering and exiting counting method based on deep learning application |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116013072A CN116013072A (en) | 2023-04-25 |
| CN116013072B true CN116013072B (en) | 2024-02-13 |
Family
ID=86031435
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310003722.1A Active CN116013072B (en) | 2023-01-03 | 2023-01-03 | Parking lot vehicle entering and exiting counting method based on deep learning application |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116013072B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103871076A (en) * | 2014-02-27 | 2014-06-18 | 西安电子科技大学 | Moving object extraction method based on optical flow method and superpixel division |
| CN108765975A (en) * | 2018-06-21 | 2018-11-06 | 智慧互通科技有限公司 | The vertical managing system of car parking of trackside and method |
| CN109712255A (en) * | 2019-02-27 | 2019-05-03 | 北京猎户智芯科技有限公司 | A kind of Car park payment evidence-obtaining system and method based on dynamic picture synthesis |
| CN110717380A (en) * | 2019-08-28 | 2020-01-21 | 武汉理工大学 | Parking space intelligent monitoring method and system based on deep learning |
| CN110866939A (en) * | 2019-10-17 | 2020-03-06 | 南京师范大学 | Robot motion state identification method based on camera pose estimation and deep learning |
| CN111932934A (en) * | 2020-08-10 | 2020-11-13 | 广州立信电子科技有限公司 | Intelligent parking lot in-and-out vehicle management system based on big data |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8831101B2 (en) * | 2008-08-02 | 2014-09-09 | Ecole De Technologie Superieure | Method and system for determining a metric for comparing image blocks in motion compensated video coding |
-
2023
- 2023-01-03 CN CN202310003722.1A patent/CN116013072B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103871076A (en) * | 2014-02-27 | 2014-06-18 | 西安电子科技大学 | Moving object extraction method based on optical flow method and superpixel division |
| CN108765975A (en) * | 2018-06-21 | 2018-11-06 | 智慧互通科技有限公司 | The vertical managing system of car parking of trackside and method |
| CN109712255A (en) * | 2019-02-27 | 2019-05-03 | 北京猎户智芯科技有限公司 | A kind of Car park payment evidence-obtaining system and method based on dynamic picture synthesis |
| CN110717380A (en) * | 2019-08-28 | 2020-01-21 | 武汉理工大学 | Parking space intelligent monitoring method and system based on deep learning |
| CN110866939A (en) * | 2019-10-17 | 2020-03-06 | 南京师范大学 | Robot motion state identification method based on camera pose estimation and deep learning |
| CN111932934A (en) * | 2020-08-10 | 2020-11-13 | 广州立信电子科技有限公司 | Intelligent parking lot in-and-out vehicle management system based on big data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116013072A (en) | 2023-04-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107301776A (en) | Track road conditions processing and dissemination method based on video detection technology | |
| TW200905575A (en) | Method for finding paths in video | |
| CN108765975B (en) | Roadside vertical parking lot management system and method | |
| CN102129785A (en) | Intelligent management system for large-scene parking lot | |
| CN108710827B (en) | A kind of micro- police service inspection in community and information automatic analysis system and method | |
| CN113160575A (en) | Traffic violation detection method and system for non-motor vehicles and drivers | |
| CN114023062B (en) | Traffic flow information monitoring method based on deep learning and edge calculation | |
| CN111047723B (en) | City wisdom behavior analysis system based on image processing | |
| CN110111565A (en) | A kind of people's vehicle flowrate System and method for flowed down based on real-time video | |
| CN107248296B (en) | Video traffic flow statistical method based on unmanned aerial vehicle and time sequence characteristics | |
| CN111739335A (en) | Parking detection method and device based on visual difference | |
| WO2023179416A1 (en) | Method and apparatus for determining entry and exit of vehicle into and out of parking space, device, and storage medium | |
| CN111311766A (en) | Roadside parking intelligent charging system and method based on license plate recognition and tracking technology | |
| KR100820952B1 (en) | Detecting method at automatic police enforcement system of illegal-stopping and parking vehicle using single camera and system thereof | |
| CN111898485A (en) | Parking space vehicle detection processing method and device | |
| CN111223303B (en) | Charging pile parking space management method, device and system | |
| CN114049612A (en) | Highway vehicle charge auditing system based on graph searching technology and dual-obtaining and checking method for driving path | |
| CN116092206A (en) | Identification and charging system and method for unattended vehicle entering and exiting parking lot | |
| CN110880205B (en) | Parking charging method and device | |
| CN112836683A (en) | License plate recognition method, device, equipment and medium for portable camera equipment | |
| CN114255428A (en) | Vehicle parking identification method based on edge intelligence and roadside high-level monitoring video | |
| CN115909223A (en) | Method and system for matching WIM system information with monitoring video data | |
| CN108510744B (en) | System and method for realizing semi-closed parking lot management based on long and short focal length cameras | |
| CN116013072B (en) | Parking lot vehicle entering and exiting counting method based on deep learning application | |
| WO2023246720A1 (en) | Roadside parking detection method, roadside parking system, and electronic device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 518000, Building 502, Nanhai Yiku, Xinghua Road, Shekou, Shuiwan Community, Merchants Street, Nanshan District, Shenzhen, Guangdong Province Applicant after: Shenzhen easy Alliance Technology Co.,Ltd. Address before: 413, Building 2, Nanhai Yiku, Xinghua Road, Shekou, Shuiwan Community, Merchants Street, Nanshan District, Shenzhen City, Guangdong Province, 518000 Applicant before: Shenzhen easy Alliance Technology Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |