CN111696135A - Intersection ratio-based forbidden parking detection method - Google Patents
Intersection ratio-based forbidden parking detection method Download PDFInfo
- Publication number
- CN111696135A CN111696135A CN202010504028.4A CN202010504028A CN111696135A CN 111696135 A CN111696135 A CN 111696135A CN 202010504028 A CN202010504028 A CN 202010504028A CN 111696135 A CN111696135 A CN 111696135A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- frame
- target
- max
- ratio
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/08—Detecting or categorising vehicles
Abstract
The invention discloses a contraband parking detection method based on intersection-parallel ratio, which comprises the following steps: step 1: taking 1 frame of picture every N frames, detecting targets such as vehicles by using a supervised learning mechanism through a deep learning neural network detection algorithm, and extracting coordinates of the lower left corner and the upper right corner of a rectangular frame of the vehicle target; step 2: calculating the intersection ratio of all target rectangular frames in the current frame and the previous frame of picture and sequencing; and step 3: and when the detected intersection ratio is larger than the threshold value and is continuously above the threshold value for multiple times, returning the coordinates of the target object and the rectangular frame, and judging that the vehicle is parked. The method can accurately judge the abnormal parking event and provides a more accurate detection method for detecting the abnormal parking event through the video on the urban expressway.
Description
Technical Field
The invention belongs to the field of artificial intelligence and intelligent traffic, and provides a contraband parking detection method based on intersection-parallel ratio.
Background
The traditional forbidden parking method comprises the steps of detecting a foreground target through difference between a video image and a background image, judging a static target through pixel-level time sequence analysis, and identifying forbidden vehicles through feature extraction.
With the development of the deep neural network technology, it has appeared that the vehicle is first detected by the neural network, and then the vehicle is determined to be stopped if the position changes little or not by the change of the center position of the vehicle target between successive frames. However, when the scheme is actually applied, the shaking phenomenon of the frame of the detected vehicle can cause the shaking of the center of the vehicle, and the parking missing detection rate is high.
Disclosure of Invention
1. Objects of the invention
The invention provides a detection method for detecting vehicles through a supervised learning neural network and detecting illegal parking based on a cross-merge ratio, aiming at solving the problems of jitter phenomenon and high omission ratio.
2. The technical scheme adopted by the invention
The invention discloses a contraband parking detection method based on intersection-parallel ratio, which comprises the following steps:
step 1: taking 1 frame of picture every N frames, detecting a vehicle target by using a supervised learning mechanism through a deep learning neural network detection algorithm, and extracting coordinates of the lower left corner and the upper right corner of a rectangular frame of the vehicle target;
step 2: calculating the intersection ratio of all target rectangular frames in the current frame and the previous frame of picture and sequencing;
and step 3: and when the detected intersection ratio is larger than the threshold value and is continuously above the threshold value for multiple times, returning the coordinates of the target object and the rectangular frame, and judging that the vehicle is parked.
The detailed process of the steps is as follows:
in the step 1:
for monitoring video stream, every Nc frame, wherein Nc > is 1, taking 1 frame of picture, detecting a vehicle target by a deep learning neural network detection algorithm including YOLO and SSD, and executing a target detection algorithm once;
the k detection has M vehicles, and the M (1 ═ M)<m<M) coordinates of the lower left and upper right corners of the rectangular frame of the vehicle are noted as (x, y) mkld, (x，y) mkrt，The k-1 th detection has N vehicles, and the N (1 ═ th)<n<N) coordinates of the lower left corner and the upper right corner of the rectangular frame of the vehicle are (x, y) n(k-1)ld (x，y) n(k-1)rt 。
Step 2, calculating the intersection ratio of the two vehicles:
firstly, a calculation formula of the cross-to-parallel ratio of two vehicles is defined, and pixel coordinates (x) of the upper left corner and the lower right corner of a rectangular box of the vehicle Aalt,yalt) And (x)ard,yard) Pixel coordinates (x) of the upper left corner and the lower right corner of the rectangular frame of the vehicle Bblt,yblt) And (x)brd,ybrd) Then, the intersection-to-parallel ratio of the vehicle a and the vehicle B is calculated as follows:
and calculating the coordinates of the intersection area of the vehicle A and the vehicle B as follows:
xi1＝max(xalt,xblt)
yi1＝max(yalt,yblt)
xi2＝min(xard,xbrd)
yi2＝min(yard,ybrd)
the areas of A, B, A ^ B and AjB are calculated
AA＝(xard-xalt+1)*(yard-yalt+1)
AB＝(xbrd-xblt+1)*(ybrd-yblt+1)
AA∩B＝max(xi2-xi1+1,0)*max(yi2-yi1+1,0)
AA∪B＝AA+AB-AA∩B
The intersection ratio of A and B is:
IOUA,B＝AA∩B/AA∪B。
further, in step 2, the intersection ratio of all target vehicles in the current frame and all target vehicles in the previous frame is calculated: detecting M target vehicles at the kth time in the step 1, and detecting N vehicles at the k-1 time;calculating the intersection ratio IOU of the m-th vehicle detected at the k time and the n-th vehicle detected at the k-1 timemk,n(k-1)(ii) a The maximum intersection ratio of the mth vehicle detected at the kth time and all the target vehicles detected at k-1 is as follows:
IOUmk,N(k-1),max＝max(IOUmk,n(k-1))1<＝m<＝M；1<＝n<＝N
the number i of the vehicle detected k times corresponding to the maximum value k-1n(k-1),imk。
Further, defining a cross-over ratio threshold σ, 0< σ < 1;
if the IOU is not in the normal statemk,N(k-1),max>If σ, the count value C _ i of the vehicle is obtainedmk＝C_in(k-1)+ 1; if the IOU is not in the normal statemk,N(k-1),max<σ, then calculate IOUmk,N(k-2),max；
If the IOU is not in the normal statemk,N(k-2),max>If σ, the count value C _ i of the vehicle is obtainedmk＝C_in(k-2)+ 1; if the IOU is not in the normal statemk,N(k-2),max<σ, then C _ imk＝0。
Still further, a preferred empirical threshold σ is set to 0.8.
The contraband parking detection method based on the intersection ratio is characterized by comprising the following steps of 3:
when the m-th vehicle detected at the k-th time is counted as a count value C _ imk>＝CthresJudging that the vehicle is parked; wherein C isthres＝int(Tthres*fps/Nc)；TthresAnd setting a vehicle stop time threshold value according to actual needs, wherein fps is the frame rate of the video, and Nc is the frame extraction rate, namely, detecting once every Nc frames.
3. Advantageous effects adopted by the present invention
(1) The invention adopts a deep neural network to detect the vehicle target, and then judges the possible parking vehicle by calculating the intersection ratio of the vehicles among different frames.
(2) The invention sets the area range of each target in the picture and solves the jitter phenomenon.
(3) The invention adopts the comparison between two pictures of adjacent frames of time sequence to calculate the cross-over ratio, thereby solving the problem of high missing rate.
(4) The invention can obtain the difference degree of the same vehicle target at different moments, thereby judging whether the target is in a moving state or not and obviously distinguishing a moving vehicle from a static vehicle.
(5) The invention introduces redundancy factors to reduce interference: in the same target, the ROI detection values of adjacent multiple frames are close to 1, and the vehicle is judged to be a stationary vehicle only when the state lasts for a set time (for example, 3 minutes), so that false alarm of scenes such as short-time parking, stop-and-go and stop-and-go of the vehicle is avoided.
(6) The invention introduces a forgetting factor to improve the precision, and the ROI is calculated for multiple times by setting adjacent multiframes.
In conclusion, the method can accurately judge the abnormal parking event and provides a more accurate detection method for detecting the abnormal parking event through the video on the urban expressway.
Drawings
Fig. 1 is a picture of 10 minutes 19 seconds frames from 13 points of a video based on a deep neural network.
Fig. 2 is a picture of a frame for detecting 13 points for 10 minutes and 20 seconds.
Fig. 3 is a 11 min 34 sec real frame picture from time 13.
Fig. 4 is a 13-point 11-minute 42-second frame picture.
Detailed Description
The technical solutions in the examples of the present invention are clearly and completely described below with reference to the drawings in the examples of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without inventive step, are within the scope of the present invention.
The present invention will be described in further detail with reference to the accompanying drawings.
Example 1
Step 1, detecting a vehicle:
for monitoring video stream, every Nc frame, wherein Nc > is 1, taking 1 frame of picture, detecting a vehicle target by a deep learning neural network detection algorithm including YOLO and SSD, and executing a target detection algorithm once;
the k detection has M vehicles, and the M (1 ═ M)<m<M) coordinates of the lower left and upper right corners of the rectangular frame of the vehicle are noted as (x, y) mkld ，(x，y) mkrt The k-1 th detection has N vehicles, and the N (1 ═ th)<n<N) coordinates of the lower left corner and the upper right corner of the rectangular frame of the vehicle are (x, y) n(k-1)ld ，(x，y) n(k-1)rt 。
Step 2, calculating the intersection ratio:
firstly, a calculation formula of the cross-to-parallel ratio of two vehicles is defined, and pixel coordinates (x) of the upper left corner and the lower right corner of a rectangular box of the vehicle Aalt,yalt) And (x)ard,yard) Pixel coordinates (x) of the upper left corner and the lower right corner of the rectangular frame of the vehicle Bblt,yblt) And (x)brd,ybrd) Then, the intersection-to-parallel ratio of the vehicle a and the vehicle B is calculated as follows:
the coordinates of the intersection area of the A vehicle and the B vehicle are calculated as
xi1＝max(xalt,xblt)
yi1＝max(yalt,yblt)
xi2＝min(xard,xbrd)
yi2＝min(yard,ybrd)
The areas of A, B, A ^ B and AjB are calculated
AA＝(xard-xalt+1)*(yard-yalt+1)
AB＝(xbrd-xblt+1)*(ybrd-yblt+1)
AA∩B＝max(xi2-xi1+1,0)*max(yi2-yi1+1,0)
AA∪B＝AA+AB-AA∩B
The intersection ratio of A and B is:
IOUA,B＝AA∩B/AA∪B。
calculating the intersection ratio of all target vehicles of the current frame and all target vehicles of the previous frame: detecting M target vehicles at the kth time in the step 1, and detecting N vehicles at the k-1 time; calculating the intersection ratio IOU of the m-th vehicle detected at the k time and the n-th vehicle detected at the k-1 timemk,n(k-1)(ii) a The maximum intersection ratio of the m vehicle detected at the k time and all the target vehicles detected at k-1 is
IOUmk,N(k-1),max＝max(IOUmk,n(k-1))1<＝m<＝M；1<＝n<N-1 number of detected vehicles k times corresponding to the maximum value thereofn(k-1),imk。
Manually setting and defining an intersection ratio threshold value sigma by software, wherein 0< sigma < 1;
if the IOU is not in the normal statemk,N(k-1),max>If σ, the count value C _ i of the vehicle is obtainedmk＝C_in(k-1)+ 1; if the IOU is not in the normal statemk,N(k-1),max<σ, then calculate IOUmk,N(k-2),max；
If the IOU is not in the normal statemk,N(k-2),max>If σ, the count value C _ i of the vehicle is obtainedmk＝C_in(k-2)+ 1; if the IOU is not in the normal statemk,N(k-2),max<σ, then C _ imk＝0。
A preferred empirical threshold sigma is set to 0.8.
And step 3:
when the m-th vehicle detected at the k-th time is counted as a count value C _ imk>＝CthresJudging that the vehicle is parked; wherein C isthres＝int(Tthres*fps/Nc)；TthresAnd setting a vehicle stop time threshold value according to actual needs, wherein fps is the frame rate of the video, and Nc is the frame extraction rate, namely, detecting once every Nc frames.
Through the algorithm from step 1 to step 3, the two vehicles below the rightmost lane in fig. 1-4 can be determined as forbidden parking.
The method carries out target identification on each frame of picture, adopts deep learning convolutional neural network detection to realize target detection, targets such as vehicles, trucks, dangerous goods vehicles and the like, and provides the area range of each target in the picture (preferably expressed by a rectangular frame, and the upper left end point (x) of the rectangular frame is adoptedlt,y lt ) Lower right endpoint (x) rd ,y rd ) Represents); the cross-over ratio is calculated by comparing two pictures between time sequence adjacent frames.
According to the invention, the difference degree of the same vehicle target at different moments can be obtained, so that whether the target is in a motion state or not can be judged. The method can obviously distinguish a moving vehicle from a static vehicle: the positions of the static vehicle targets in a plurality of adjacent frame pictures in time sequence are basically kept unchanged, so the ROI intersection ratio is close to 1; the higher the moving speed of the moving object, the position in several frames of pictures adjacent in time sequence can be changed, so that the ROI intersection ratio is far away from 1.
The invention introduces redundancy factors to reduce interference: in the same target, the ROI detection values of adjacent multiple frames are close to 1, and the vehicle is judged to be a stationary vehicle only when the state lasts for a set time (for example, 3 minutes), so that false alarm of scenes such as short-time parking, stop-and-go and stop-and-go of the vehicle is avoided.
The invention introduces a forgetting factor to improve the precision: when the object a is still, the passing of the object B beside the object a may cause the area range of the object a detected at the corresponding time to change. The method is arranged in adjacent multiframes for calculating the ROI for multiple times, and also judges abnormal parking when the distribution of the ROI is mostly close to 1 (for example, 4 ROIs of adjacent 5 frames are close to 1, namely, the abnormal parking event can be judged)
The position of each vehicle in the picture is shown in fig. 1-4.
In fig. 1, a vehicle or pedestrian target is detected from a certain frame of picture of 13 points, 10 minutes and 19 seconds of a video based on a deep neural network, and the region and the type of the target are displayed by using a box and a label.
In fig. 2, frame extraction detection is performed in real time, after 1 second, 13 points of a certain frame of picture of 10 minutes and 20 seconds are detected again, a merging ratio calculation is performed, and a target (a lower right corner region of the picture) with the merging ratio larger than a threshold value is detected.
In fig. 3, the real-time detection lasts for a long time to 13 points, 11 minutes and 34 seconds, the intersection ratio of the vehicle is repeatedly tracked for a plurality of times and is greater than the threshold value, the algorithm judges that the target continuous stationary time is greater than 60 seconds, and the target continuous stationary time is judged to be an illegal parking event.
In fig. 4, 13 points 11 minutes 42 seconds: for a moving white vehicle, the intersection ratio is far away from 1, so that the vehicle is determined to be a moving vehicle and is not identified as an illegal event.
Based on the characteristics, the abnormal parking event can be accurately judged, and a relatively accurate detection method is provided for detecting the abnormal parking event through videos on the urban expressway.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (7)
1. A contraband parking detection method based on intersection ratio is characterized by comprising the following steps:
step 1: taking 1 frame of picture every N frames, detecting a vehicle target by using a supervised learning mechanism through a deep learning neural network detection algorithm, and extracting coordinates of the lower left corner and the upper right corner of a rectangular frame of the vehicle target;
step 2: calculating the intersection ratio of all target rectangular frames in the current frame and the previous frame of picture and sequencing;
and step 3: and when the detected intersection ratio is larger than the threshold value and is continuously above the threshold value for multiple times, returning the coordinates of the target object and the rectangular frame, and judging that the vehicle is parked.
2. The intersection ratio-based contraband parking detection method according to claim 1, wherein in step 1:
for monitoring video stream, every Nc frame, wherein Nc > is 1, taking 1 frame of picture, detecting a vehicle target by a deep learning neural network detection algorithm including YOLO and SSD, and executing a target detection algorithm once;
the k detection has M vehicles, and the M (1 ═ M)<m<M) coordinates of the upper left corner and the lower right corner of the rectangular frame of the vehicle are (x, y) mklt、 (x，y) mkrd The k-1 st detection has N vehicles, and the coordinates of the upper left corner and the lower right corner of the rectangular frame of the nth vehicle are recorded as (x, y) n(k-1)lt (x，y) n(k-1)rd Wherein 1 ═<n<＝N。
3. The intersection ratio-based contraband parking detection method according to claim 1, wherein in step 2:
firstly, a calculation formula of the cross-to-parallel ratio of two vehicles is defined, and pixel coordinates (x) of the upper left corner and the lower right corner of a rectangular box of the vehicle Aalt,yalt) And (x)ard,yard) Pixel coordinates (x) of the upper left corner and the lower right corner of the rectangular frame of the vehicle Bblt,yblt) And (x)brd,ybrd) Then, the intersection-to-parallel ratio of the vehicle a and the vehicle B is calculated as follows:
the coordinates of the intersection area of the A vehicle and the B vehicle are calculated as
xi1＝max(xalt,xblt)
yi1＝max(yalt,yblt)
xi2＝min(xard,xbrd)
yi2＝min(yard,ybrd)
The areas of A, B, A ^ B and AjB are calculated
AA＝(xard-xalt+1)*(yard-yalt+1)
AB＝(xbrd-xblt+1)*(ybrd-yblt+1)
AA∩B＝max(xi2-xi1+1,0)*max(yi2-yi1+1,0)
AA∪B＝AA+AB-AA∩B
The intersection ratio of A and B is:
IOUA,B＝AA∩B/AA∪B。
4. the prohibited parking detection method based on intersection ratio as claimed in claim 1, characterized in that in step 2, the intersection ratio of all target vehicles in the current frame and all target vehicles in the previous frame is calculated: detecting M target vehicles at the kth time in the step 1, and detecting N vehicles at the k-1 time; calculating the intersection ratio IOU of the m-th vehicle detected at the k time and the n-th vehicle detected at the k-1 timemk,n(k-1)(ii) a The maximum intersection ratio of the mth vehicle detected at the kth time and all the target vehicles detected at k-1 is as follows:
IOUmk,N(k-1),max＝max(IOUmk,n(k-1))1<＝m<＝M；1<＝n<＝N；
the number i of the vehicle detected k times corresponding to the maximum value k-1n(k-1),imk。
5. The contraband parking detection method based on the cross-over ratio as claimed in claim 4, wherein in step 2, a cross-over ratio threshold σ is defined, 0< σ < 1;
if the IOU is not in the normal statemk,N(k-1),max>If σ, the count value C _ i of the vehicle is obtainedmk＝C_in(k-1)+ 1; if the IOU is not in the normal statemk,N(k-1),max<σ, then calculate IOUmk,N(k-2),max；
If the IOU is not in the normal statemk,N(k-2),max>If σ, the count value C _ i of the vehicle is obtainedmk＝C_in(k-2)+ 1; if the IOU is not in the normal statemk,N(k-2),max<σ, then C _ imk＝0。
6. The intersection ratio-based contraband parking detection method according to claim 5, wherein: the empirical threshold σ is set to 0.8.
7. The intersection ratio-based contraband parking detection method according to claim 1, wherein in step 3:
when the m-th vehicle detected at the k-th time is counted as a count value C _ imk>＝CthresWhen it is determined thatThe vehicle is parked; wherein C isthres＝int(Tthres*fps/Nc)；TthresAnd setting a vehicle stop time threshold value according to actual needs, wherein fps is the frame rate of the video, and Nc is the frame extraction rate, namely, detecting once every Nc frames.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010504028.4A CN111696135A (en) | 2020-06-05 | 2020-06-05 | Intersection ratio-based forbidden parking detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010504028.4A CN111696135A (en) | 2020-06-05 | 2020-06-05 | Intersection ratio-based forbidden parking detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111696135A true CN111696135A (en) | 2020-09-22 |
Family
ID=72479465
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010504028.4A Pending CN111696135A (en) | 2020-06-05 | 2020-06-05 | Intersection ratio-based forbidden parking detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111696135A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112597917A (en) * | 2020-12-25 | 2021-04-02 | 太原理工大学 | Vehicle parking detection method based on deep learning |
| CN112735163A (en) * | 2020-12-25 | 2021-04-30 | 北京百度网讯科技有限公司 | Method for determining static state of target object, road side equipment and cloud control platform |
| CN113409587A (en) * | 2021-06-16 | 2021-09-17 | 北京字跳网络技术有限公司 | Abnormal vehicle detection method, device, equipment and storage medium |
| CN115116012A (en) * | 2022-07-20 | 2022-09-27 | 广州英码信息科技有限公司 | Method and system for detecting parking state of vehicle parking space based on target detection algorithm |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109118523A (en) * | 2018-09-20 | 2019-01-01 | 电子科技大学 | A kind of tracking image target method based on YOLO |
| CN109919072A (en) * | 2019-02-28 | 2019-06-21 | 桂林电子科技大学 | Fine vehicle type recognition and flow statistics method based on deep learning and trajectory tracking |
| CN110647852A (en) * | 2019-09-27 | 2020-01-03 | 集美大学 | Traffic flow statistical method, terminal equipment and storage medium |
| CN110910655A (en) * | 2019-12-11 | 2020-03-24 | 深圳市捷顺科技实业股份有限公司 | Parking management method, device and equipment |
| CN111179581A (en) * | 2018-11-13 | 2020-05-19 | 千寻位置网络有限公司 | System and method for sharing bicycle ID matching based on vision |
-
2020
- 2020-06-05 CN CN202010504028.4A patent/CN111696135A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109118523A (en) * | 2018-09-20 | 2019-01-01 | 电子科技大学 | A kind of tracking image target method based on YOLO |
| CN111179581A (en) * | 2018-11-13 | 2020-05-19 | 千寻位置网络有限公司 | System and method for sharing bicycle ID matching based on vision |
| CN109919072A (en) * | 2019-02-28 | 2019-06-21 | 桂林电子科技大学 | Fine vehicle type recognition and flow statistics method based on deep learning and trajectory tracking |
| CN110647852A (en) * | 2019-09-27 | 2020-01-03 | 集美大学 | Traffic flow statistical method, terminal equipment and storage medium |
| CN110910655A (en) * | 2019-12-11 | 2020-03-24 | 深圳市捷顺科技实业股份有限公司 | Parking management method, device and equipment |
Non-Patent Citations (2)
| Title |
|---|
| 赵逸如；刘正熙；熊运余；严广宇: "基于目标检测和语义分割的人行道违规停车检测", 《现代计算机》 * |
| 赵逸如；刘正熙；熊运余；严广宇: "基于目标检测和语义分割的人行道违规停车检测", 《现代计算机》, 7 May 2020 (2020-05-07), pages 82 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112597917A (en) * | 2020-12-25 | 2021-04-02 | 太原理工大学 | Vehicle parking detection method based on deep learning |
| CN112735163A (en) * | 2020-12-25 | 2021-04-30 | 北京百度网讯科技有限公司 | Method for determining static state of target object, road side equipment and cloud control platform |
| CN112735163B (en) * | 2020-12-25 | 2022-08-02 | 阿波罗智联(北京)科技有限公司 | Method for determining static state of target object, road side equipment and cloud control platform |
| CN112597917B (en) * | 2020-12-25 | 2022-09-30 | 太原理工大学 | Vehicle parking detection method based on deep learning |
| CN113409587A (en) * | 2021-06-16 | 2021-09-17 | 北京字跳网络技术有限公司 | Abnormal vehicle detection method, device, equipment and storage medium |
| CN113409587B (en) * | 2021-06-16 | 2022-11-22 | 北京字跳网络技术有限公司 | Abnormal vehicle detection method, device, equipment and storage medium |
| CN115116012A (en) * | 2022-07-20 | 2022-09-27 | 广州英码信息科技有限公司 | Method and system for detecting parking state of vehicle parking space based on target detection algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110136449B (en) | Deep learning-based traffic video vehicle illegal parking automatic identification snapshot method | |
| Cucchiara et al. | Statistic and knowledge-based moving object detection in traffic scenes | |
| CN111696135A (en) | Intersection ratio-based forbidden parking detection method | |
| US9501701B2 (en) | Systems and methods for detecting and tracking objects in a video stream | |
| KR101735365B1 (en) | The robust object tracking method for environment change and detecting an object of interest in images based on learning | |
| US20080181499A1 (en) | System and method for feature level foreground segmentation | |
| EP1811457A1 (en) | Video signal analysis | |
| Bedruz et al. | Real-time vehicle detection and tracking using a mean-shift based blob analysis and tracking approach | |
| Makhmutova et al. | Object tracking method for videomonitoring in intelligent transport systems | |
| Zang et al. | Object classification and tracking in video surveillance | |
| Huang et al. | A real-time and color-based computer vision for traffic monitoring system | |
| Srinivas et al. | Image processing edge detection technique used for traffic control problem | |
| Fuentes et al. | From tracking to advanced surveillance | |
| Xia et al. | Automatic multi-vehicle tracking using video cameras: An improved CAMShift approach | |
| JP7125843B2 (en) | Fault detection system | |
| KR20200060868A (en) | multi-view monitoring system using object-oriented auto-tracking function | |
| Lin et al. | A real-time multiple-vehicle detection and tracking system with prior occlusion detection and resolution, and prior queue detection and resolution | |
| Sridevi et al. | Automatic generation of traffic signal based on traffic volume | |
| Vijverberg et al. | High-level traffic-violation detection for embedded traffic analysis | |
| Sutjiadi et al. | Adaptive background extraction for video based traffic counter application using Gaussian mixture models algorithm | |
| Fuentes et al. | Tracking people for automatic surveillance applications | |
| Lin et al. | Crossroad traffic surveillance using superpixel tracking and vehicle trajectory analysis | |
| Huang | Video-based traffic analysis system using a hierarchical feature point grouping approach | |
| Lin et al. | A street scene surveillance system for moving object detection, tracking and classification | |
| Utasi et al. | Anomaly detection with low-level processes in videos |
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 |