CN110674703A - Video tripwire alarm counting method and flow in intelligent monitoring - Google Patents
Video tripwire alarm counting method and flow in intelligent monitoring Download PDFInfo
- Publication number
- CN110674703A CN110674703A CN201910836375.4A CN201910836375A CN110674703A CN 110674703 A CN110674703 A CN 110674703A CN 201910836375 A CN201910836375 A CN 201910836375A CN 110674703 A CN110674703 A CN 110674703A
- Authority
- CN
- China
- Prior art keywords
- detection
- vector
- line
- pedestrian
- tripwire
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30232—Surveillance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30242—Counting objects in image
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Signal Processing (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
- Closed-Circuit Television Systems (AREA)
Abstract
The invention discloses a video tripwire alarm counting method and a flow in intelligent monitoring, S1, obtaining a real-time monitoring image of a camera, and marking a warning line on the real-time monitoring image, wherein the warning line can be set according to requirements, and different warning lines can be set for a plurality of cameras simultaneously; s2, detecting and tracking pedestrians and vehicles on the real-time monitoring picture, and outputting a detection image comprising a pedestrian frame or a vehicle frame; s3, determining whether a pedestrian or a vehicle is in a detection image to trip a line, and outputting information about whether the pedestrian or the vehicle trips the line; and S4, counting the number of tripwires in the current monitoring picture, and transmitting the number to the client in real time to remind and save the current monitoring picture accompanied by the warning wire and the object frame. The invention realizes automatic alarm for important areas or gates according to set rules by drawing detection areas and warning lines for monitoring pictures and monitoring the images by using pedestrian and vehicle detection and tracking algorithms.
Description
Technical Field
The invention relates to the field of computer vision and security intelligent monitoring, in particular to a video tripwire alarm counting method and a video tripwire alarm counting process in intelligent monitoring.
Background
Video monitoring is taken as an important component of a safety precaution system, is a comprehensive system with stronger precaution capacity, along with the rapid development of social economy and the progress of scientific technology, a monitoring system is widely applied to various industries with intuitive, accurate, timely and rich contents, however, the large-scale application of video monitoring causes that people can hardly realize all-round identification, the traditional video system generally realizes safety protection by arranging specific personnel to watch monitoring pictures through shifts, the safety can not be actively and effectively ensured in practice, and especially when monitoring points are more and widely distributed, the personnel monitoring can not estimate all monitoring scenes at all; meanwhile, the attention of monitoring personnel is difficult to be paid for all monitoring scenes in twenty-four hours, time and labor are wasted, the monitoring personnel are easily interfered by human factors, and serious missing detection and false detection are generated on some important checkpoints or regions; passive recording is furthermore usually only possible after an event has occurred, with the resulting losses being irretrievable, and with the playback forensic mode being particularly inefficient.
In summary, the defects of the traditional monitoring at the important checkpoints or areas are reflected, and an improvement mode needs to be provided urgently, while the intelligent video analysis can effectively solve the defects of the traditional monitoring at the important checkpoints or areas, the main characteristic of the intelligent monitoring is that a computer vision mode is adopted, under the condition that the human intervention is hardly needed, the positioning, the identification and the tracking of the specific object in the real-time video frame sequence shot and recorded by the camera are realized, along with the continuous development of scientific technology and computer vision, the advanced technology is applied to the traditional monitoring, not only the passive monitoring is realized as the active monitoring, the advance early warning is achieved, but also the computer is used for replacing the manual real-time monitoring task, the in-fact processing is achieved, finally the pixel target/time can be quickly searched in mass videos, and the convenience condition is provided for the later evidence obtaining, therefore, for a heavy spot area or an important gate, it is urgent to design a monitoring system for actively alarming when a pedestrian or a vehicle crosses a warning line in violation of a certain rule by using an advanced technology.
Disclosure of Invention
The invention aims to provide a tripwire alarm counting method and a tripwire alarm counting process based on video analysis, which mainly aim at the problem that pedestrians or vehicles enter an important area or an important entrance to carry out active alarm, monitor the pedestrians and the vehicles in a camera picture by defining the important area or a warning line and simultaneously utilizing the detection and tracking services of the pedestrians and the vehicles, immediately send an alarm to a terminal if the pedestrians and the vehicles break into the important area or cross the warning line once, and remind security personnel of abnormal conditions appearing in the camera picture in real time, thereby solving the defects of low intelligent degree and poor safety of the traditional monitoring system.
In order to achieve the purpose, the invention provides the following technical scheme: a video tripwire alarm counting method and flow in intelligent monitoring comprises the following steps:
s1, acquiring a real-time monitoring image of the camera, and defining a warning line on the real-time monitoring image, wherein a single warning line or a plurality of warning lines can be set according to requirements, and different warning lines can be set for a plurality of cameras simultaneously;
s2, detecting and tracking pedestrians or vehicles on the real-time monitoring picture, and outputting a detection image comprising a pedestrian frame or a vehicle frame;
s3, carrying out wire tripping judgment on the pedestrian frame or the vehicle frame in the detection image, and outputting relevant information whether the pedestrian or the vehicle trips the wire;
and S4, counting the number of tripwires in the current monitoring picture according to the information output in S3, and transmitting the current monitoring picture accompanied with the warning wire and the object frame to the client in real time for reminding and storing.
Further optimizing the technical solution, step S1 specifically includes:
s11, opening a video alarm service configuration program, capturing an image of a camera monitoring picture, firstly drawing a detection area on the image according to a Z-shaped sequence, then drawing a warning line in the detection area after selecting a related warning line configuration scheme, and finally storing a configuration file.
Further optimizing the technical solution, step S2 specifically includes:
s21, training a pedestrian and vehicle detection model based on a YOLOv3 network framework; further, model pruning and compression are implemented to improve the detection speed of one image of the model under the condition of ensuring that the detection precision is almost unchanged;
s22, detecting vehicles or pedestrians on the real-time monitoring picture of the camera by using the trained vehicle and pedestrian detection model, and outputting a detection image comprising a pedestrian detection frame or a vehicle detection frame;
s23, tracking the result output by the S22 by adopting an FDSST (fast discrete Scale Space tracker) tracking algorithm, and simultaneously recording the motion trail of the target;
further optimizing the technical scheme, the pedestrian and vehicle training data set used in the invention is MS COCODataset + Imagenet dataset.
Further optimizing the technical solution, step S3 specifically includes:
s31, a rule of crossing the alert line is established, as shown in fig. 2 and fig. 3 (line segment L1L2 represents the alert line, and rectangular box ABCD represents the pedestrian or vehicle detection box in one image).
Further optimizing the technical scheme, the step S31 is specifically:
s311, coordinates L1(a1, b1) and L2(a2, b2) at two ends of the alarm ring line; 4 vertex coordinates A (x1, y1), B (x2, y2), C (x3, y3) and D (x4, y4) of the detection frame are determined by utilizing the directivity of vector outer products and judging the relation between 4 sides of the detection frame and the warning line;
s312, further, the specific judgment rule is illustrated by taking two endpoints L1 and L2 of the detection frame vertices A and C and the warning line as an example, namely the judgment lineThe relationship between segment AC and segment L1L2, calculating a vector
And vector
Sum vector
And vector
Outer product of, in vectors
And vector
The outer product of (a) is an example to illustrate the calculation process:
identical vector
And vector
The outer product of (c) is also performed in this manner;
s313, calculating a vectorSum vector
Outer product and vector of
Sum vector
While computing the vector
Sum vectorOuter product and vector of
Sum vector
If both of the above products are less than 0, it means that the line segment L1L2 intersects the line segment AC, otherwise, it means that the line segment L1L2 does not intersect the line segment AC;
s314, repeating the steps S312 and S313, sequentially calculating the relation between the line segment AB, the line segment CD and the line segment BD and the line segment L1L2, and determining that the pedestrian or the vehicle in the detection frame has the behavior of crossing the warning line as long as a certain line segment is intersected with the line segment L1L2 in the detection frame;
s315, FIG. 3 is a class of extensions of FIG. 2, and other similar extensions are included in the present invention.
S32, establishing tripwire detection categories, wherein the tripwire detection categories are divided into two categories: one is a unidirectional tripwire alarm, the other is a bidirectional tripwire alarm, the unidirectional tripwire class is shown in the attached figures 4 and 5, and the bidirectional tripwire class is shown in the attached figures 6 and 7: (the direction of the line segment L1L2 represents the guard line direction, and the direction of the line segment BA represents the direction of the travel locus of the pedestrian or the vehicle).
FIGS. 4 and 5 show two kinds of one-way tripwire diagrams, and the extension of the two kinds of diagrams belongs to the content of the invention, and FIGS. 6 and 7 show two kinds of two-way tripwire diagrams, and the extension of the two kinds of diagrams belongs to the content of the invention.
Further optimizing the technical scheme, the step S32 is specifically:
s321, judging the rule of the bidirectional tripwire: the same as the implementation step of S31, searching from the result returned in S31, if a pedestrian or a vehicle crosses a warning line and the tripwire type in the configuration file is selected as a bidirectional tripwire alarm, directly alarming and transmitting the result back to the client;
s322, judging a one-way tripwire alarm rule: in the present invention, as shown in fig. 4 and 5, fig. 4 is taken as an example to explain, two end points L1(a1, B1) and L2(a2, B2) of the alarm line and the travel direction of the object at a predetermined time are set, B (x1, y1) and a (x2, y2) are set, whether the tripwire rule is met is judged by using the vector outer product with directivity, and a vector is calculated
And
outer product of (i.e.)
If vector
And
if the outer product of the alarm is larger than zero, the pedestrian or the vehicle is judged to have the behavior of crossing the alarm line in one direction, and then the result is transmitted back to the client side to give an alarm.
Further optimize this technical scheme, the intelligent monitoring tripwire warning still includes:
and S4, counting the number of tripwires based on the result of S3, setting a counter, counting the number of pedestrian or vehicle detection frames crossing the warning line at the current moment, and simultaneously transmitting the number back to the client to remind and early warn monitoring personnel in real time.
Advantageous effects
Compared with the prior art, the invention provides a video tripwire alarm counting method and a video tripwire alarm counting process in intelligent monitoring, which have the following beneficial effects:
1. according to the video tripwire alarm counting method and the video tripwire alarm counting process in intelligent monitoring, the key area and the warning line are defined for the obtained real-time monitoring image, so that on one hand, the searching range is narrowed, the accuracy and the speed of an intelligent detection algorithm are improved, on the other hand, the warning line is set, and once pedestrians or vehicles violate the set warning line rule, the system provides an automatic alarm function and directly returns the result to a client.
2. The invention is based on a YOLOv3 network framework, and trains a detection model of people and vehicles by using MS COCO dataset and Imagenet dataset as a training set, then prunes and compresses the model by adopting a pruning algorithm, and finally finely adjusts the model, thereby not only reducing the space occupied by a depth model, but also improving the prediction speed and accuracy of the model, and further improving the performance of an algorithm system due to the excellent prediction capability of the YOLOv3 network on images with different scales, and finally, the detection algorithm is combined with a tracking algorithm to ensure that the robustness of the system is higher.
3. According to the video tripwire alarm counting method and the video tripwire alarm counting process in intelligent monitoring, different rules are set for key areas or key ports, and alarm modes configured for areas with different key degrees in multiple modes are more flexibly realized.
4. The video tripwire alarm counting method and the video tripwire alarm counting process in the intelligent monitoring are suitable for various practical environments, can effectively judge whether pedestrians or vehicles break into an early warning area, realize early warning in advance and treatment in the event, and provide key evidence for later consultation and evidence collection.
Drawings
FIG. 1 is a block diagram of a video tripwire alarm counting method and process for intelligent monitoring according to the present invention;
FIG. 2 is a schematic view of a pedestrian or vehicle crossing a fence according to the present invention;
FIG. 3 is an extension of the pedestrian or vehicle crossing cordline schematic of FIG. 2 of the present invention;
FIG. 4 is a schematic diagram of a pedestrian or a vehicle crossing a warning line in one direction according to the present invention;
FIG. 5 is an expanded view of the pedestrian or vehicle crossing the warning line in one direction of FIG. 4 according to the present invention;
FIG. 6 is a schematic diagram of a pedestrian or vehicle crossing a fence in both directions according to the present invention;
FIG. 7 is an extension of the pedestrian or vehicle bi-directional crossing cordline schematic of FIG. 4 in accordance with the present invention;
FIG. 8 is a schematic diagram of a software interface for real-time monitoring of image rendering, configuration of detection zones and warning lines according to the present invention;
FIG. 9 is a schematic diagram illustrating the effect of monitoring the detection area and the warning line drawn by the image in real time according to the present invention;
FIG. 10 is a schematic diagram illustrating the warning effect of a pedestrian crossing the warning line in both directions according to a real-time monitoring image of the present invention;
fig. 11 is a schematic diagram of the warning effect of pedestrian crossing the warning line in one direction according to the real-time monitoring image of the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the embodiments 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 given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1-11, the present invention discloses a video tripwire alarm counting method and process in intelligent monitoring, which is characterized in that the method comprises the following steps:
s1, acquiring a real-time monitoring image of the camera, and defining a warning line on the real-time monitoring image, wherein a single warning line or a plurality of warning lines can be set according to requirements, and different warning lines can be set for a plurality of cameras simultaneously;
s2, detecting and tracking pedestrians and vehicles in the real-time monitoring picture, and outputting a detection image comprising a pedestrian frame and a vehicle frame;
s3, carrying out trip line judgment on the pedestrian frame and the vehicle frame in the detection image, and outputting the relevant information whether the pedestrian or the vehicle trips the line;
and S4, counting the number of tripwires in the current monitoring picture according to the information output in S3, and transmitting the current monitoring picture accompanied with the warning wire and the object frame to the client in real time for reminding and storing.
The invention sets a detection area and a warning line for the obtained monitoring image, monitors a key checkpoint or an important area by using a pedestrian and vehicle detection and tracking algorithm, judges whether abnormal pedestrians or vehicles exist in the detection area of the monitoring image, tracks and judges whether a tripwire rule is violated if the abnormal pedestrians or vehicles exist in the detection area of the monitoring image, and gives corresponding alarm.
As a specific optimization scheme of this embodiment: step S1 specifically includes: s11, opening a video alarm service configuration program, capturing an image of a camera monitoring picture, firstly drawing a detection area on the image according to a Z-shaped sequence, then drawing a warning line in the detection area after selecting a related warning line configuration scheme, and finally storing a configuration file;
taking a certain monitoring as an example, a real-time monitoring picture is obtained through a monitoring camera, then a frame of image of the monitoring picture is captured, tripwire configuration software is opened, an interface is shown as an attached drawing 8, the captured monitoring picture is read in at the same time, a detection area and a warning line are drawn manually according to a Z-shaped sequence, as shown in a rectangular frame shown in an attached drawing 9, a bidirectional arrow line is a warning line, such as an entrance and an exit of a company, a passageway opening of a main passageway or a cell, an entrance and an exit of a military camp, an entrance and an exit of a confidential data room, and the like.
As a specific optimization scheme of this embodiment: step S2 specifically includes:
s21, training a pedestrian and vehicle detection model based on a YOLOv3 network framework; further, model pruning and compression are implemented to improve the detection speed of one image of the model under the condition of ensuring that the detection precision is almost unchanged;
s22, detecting vehicles and pedestrians on the real-time monitoring picture of the camera by using the trained vehicle and pedestrian detection model, and outputting a detection image comprising a pedestrian detection frame and a vehicle detection frame;
and S23, tracking the result output in S22 by adopting an FDSST (fast discrete Scale Space tracker) tracking algorithm, and recording the motion trail of the target.
The pedestrian and vehicle training data set used in the invention is MS COCO dataset + Imagenet dataset;
in the field of intelligent security monitoring, the robustness of detection and tracking of pedestrians and vehicles greatly influences the normal operation of other functional modules of a security system and the overall performance of the system, in the embodiment, a network frame used is a fully-convoluted YOLOv3, a detection model training set is MS COCO dataset + Imagenet dataset, during actual training, input images are unified and normalized to be 416x416 in size, multi-scale training is started, so that the network can automatically adapt to the change of images with different scales, the robustness of detection is improved, meanwhile, after the training is finished, a pruning algorithm is used for compressing a network structure and weight, and finally, when the detection is finished, a returned detection frame is input into a tracking algorithm for tracking a target and recording the motion track of the target.
As a specific optimization scheme of this embodiment: s31 specifically includes:
s311, coordinates L1(a1, b1) and L2(a2, b2) at two ends of the alarm ring line; 4 vertex coordinates A (x1, y1), B (x2, y2), C (x3, y3) and D (x4, y4) of the detection frame are determined by utilizing the directivity of vector outer products and judging the relation between 4 sides of the detection frame and the warning line;
s312, further, the concrete judgment rule is described by taking the two endpoints L1 and L2 of the detection frame vertexes A and C and the warning line as an example, namely judging the relation between the line segment AC and the line segment L1L2, and calculating a vectorAnd vectorSum vector
And vector
Outer product of, in vectors
And vector
The outer product of (a) is an example to illustrate the calculation process:
identical vector
And vector
The outer product of (c) is also performed in this manner;
s313, calculating a vector
Sum vector
Outer product and vector of
Sum vector
While computing the vector
Sum vector
Outer product and vector of
Sum vector
If both of the above products are less than 0, it means that the line segment L1L2 intersects the line segment AC, otherwise, it means that the line segment L1L2 does not intersect the line segment AC;
s314, repeating the steps S312 and S313, sequentially calculating the relation between the line segment AB, the line segment CD and the line segment BD and the line segment L1L2, and determining that the pedestrian or the vehicle in the detection frame has the behavior of crossing the warning line as long as a certain line segment is intersected with the line segment L1L2 in the detection frame;
s315, FIG. 3 is a class of extensions of FIG. 2, and other similar extensions are included in the present invention.
S32, establishing tripwire detection categories, wherein the tripwire detection categories are divided into two categories: one is a unidirectional tripwire alarm, the other is a bidirectional tripwire alarm, the unidirectional tripwire class is shown in the attached figures 4 and 5, and the bidirectional tripwire class is shown in the attached figures 6 and 7: (the direction of the line segment L1L2 represents the guard line direction, and the direction of the line segment BA represents the direction of the travel locus of the pedestrian or the vehicle).
FIGS. 4 and 5 show two kinds of one-way tripwire diagrams, and the extension of the two kinds of diagrams belongs to the content of the invention, and FIGS. 6 and 7 show two kinds of two-way tripwire diagrams, and the extension of the two kinds of diagrams belongs to the content of the invention.
Further optimizing the technical scheme, the step S32 is specifically:
s321, judging the rule of the bidirectional tripwire: the same as the implementation step of S31, searching from the result returned in S31, if a pedestrian or a vehicle crosses the warning line and the tripwire type in the configuration file is selected as a bidirectional tripwire, directly alarming and transmitting the result back to the client;
s322, judging a unidirectional tripwire rule: in the present invention, as shown in fig. 4 and 5, fig. 4 is taken as an example to explain, two end points L1(a1, B1) and L2(a2, B2) of the alarm line and the travel direction of the object at a predetermined time are set, B (x1, y1) and a (x2, y2) are set, whether the tripwire rule is met is judged by using the vector outer product with directivity, and the tripwire rule is countedCalculating vector quantity
Andouter product of (i.e.)
If vector
Andif the outer product of the alarm is larger than zero, the pedestrian or the vehicle is judged to have the behavior of crossing the alarm line in one direction, and then the result is transmitted back to the client side to give an alarm.
And carrying out line tripping judgment on the pedestrian frame and the vehicle frame in the detection image, and outputting related information whether the pedestrian or the vehicle trips the line: according to the designed tripwire rule, a unidirectional tripwire alarm and a bidirectional tripwire alarm are respectively set for a doorway of a certain park, the unidirectional alarm is set in a setting mode shown in an attached drawing 4, and the operation result in a certain time period is shown in an attached drawing 10 and an attached drawing 11.
As a specific optimization scheme of this embodiment: counting the number of tripwires based on the result of S3, setting a counter, counting the number of pedestrians or vehicles crossing the alarm wire at the current moment, and simultaneously returning to the client for real-time reminding and early warning of monitoring personnel, wherein the counting is to count the number of rectangular frames crossing the alarm wire at a certain moment according to different tripwire detection rules in S3, and then sending the number to the client for displaying in a network form.
The invention has the beneficial effects that:
1. according to the video tripwire alarm counting method and the video tripwire alarm counting process in intelligent monitoring, the key area and the warning line are defined for the obtained real-time monitoring image, so that on one hand, the searching range is narrowed, the accuracy and the speed of an intelligent detection algorithm are improved, on the other hand, the warning line is set, and once pedestrians or vehicles violate the set warning line rule, the system provides an automatic alarm function and directly returns the result to a client.
2. The invention is based on a YOLOv3 network framework, and trains a detection model of people and vehicles by using MS COCO dataset and Imagenet dataset as a training set, then prunes and compresses the model by adopting a pruning algorithm, and finally finely adjusts the model, thereby not only reducing the space occupied by a depth model, but also improving the prediction speed and accuracy of the model, and further improving the performance of an algorithm system due to the excellent prediction capability of the YOLOv3 network on images with different scales, and finally, the detection algorithm is combined with a tracking algorithm to ensure that the robustness of the system is higher.
3. According to the video tripwire alarm counting method and the video tripwire alarm counting process in intelligent monitoring, different rules are set for key areas or key ports, and alarm modes configured for areas with different key degrees in multiple modes are more flexibly realized.
4. The video tripwire alarm counting method and the video tripwire alarm counting process in the intelligent monitoring are suitable for various practical environments, can effectively judge whether pedestrians or vehicles break into an early warning area, realize early warning in advance and treatment in the event, and provide key evidence for later consultation and evidence collection.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (6)
1. A video tripwire alarm counting method and flow in intelligent monitoring are characterized by comprising the following steps:
s1, acquiring a real-time monitoring image of the camera, and defining a warning line on the real-time monitoring image, wherein a single warning line or a plurality of warning lines can be set according to requirements, and different warning lines can be set for a plurality of cameras simultaneously;
s2, detecting and tracking pedestrians or vehicles on the real-time monitoring picture, and outputting a detection image comprising a pedestrian frame or a vehicle frame;
s3, carrying out wire tripping judgment on the pedestrian frame or the vehicle frame in the detection image, and outputting relevant information whether the pedestrian or the vehicle trips the wire;
and S4, counting the number of tripwires in the current monitoring picture according to the information output in S3, and transmitting the current monitoring picture accompanied with the warning wire and the object frame to the client in real time for reminding and storing.
2. The method and process for counting video tripwire alarms in intelligent monitoring according to claim 1, wherein the step S1 is specifically: s11, opening a video alarm service configuration program, capturing an image of a camera monitoring picture, firstly drawing a detection area on the image according to a Z-shaped sequence, then drawing a warning line in the detection area after selecting a related warning line configuration scheme, and finally storing a configuration file.
3. The method and process for counting video tripwire alarms in intelligent monitoring according to claim 1, wherein the step S2 is specifically:
s21, training a pedestrian and vehicle detection model based on a YOLOv3 network framework; further, model pruning and compression are implemented to improve the detection speed of one image of the model under the condition of ensuring that the detection precision is almost unchanged;
s22, detecting vehicles and pedestrians on the real-time monitoring picture of the camera by using the trained vehicle and pedestrian detection model, and outputting a detection image comprising a pedestrian detection frame or a vehicle detection frame;
and S23, tracking the result output by the S22 by adopting an FDSST tracking algorithm, and recording the motion trail of the target.
4. The method and process for counting video tripwire alarms in intelligent monitoring according to claim 1, wherein the step S3 is specifically:
s31, establishing a rule for crossing the warning line;
and S32, establishing a tripwire detection category.
5. The method and process according to claim 4, wherein the step S31 is specifically as follows:
s311, coordinates L1(a1, b1) and L2(a2, b2) at two ends of the alarm ring line; 4 vertex coordinates A (x1, y1), B (x2, y2), C (x3, y3) and D (x4, y4) of the detection frame are determined by utilizing the directivity of vector outer products and judging the relation between 4 sides of the detection frame and the warning line;
s312, further, the concrete judgment rule is described by taking the two endpoints L1 and L2 of the detection frame vertexes A and C and the warning line as an example, namely judging the relation between the line segment AC and the line segment L1L2, and calculating a vector
And vectorSum vector
And vector
Outer product of, in vectorsAnd vector
The outer product of (a) is an example to illustrate the calculation process:
also, the same applies toVector of (2)
And vector
The outer product of (c) is also performed in this manner;
s313, calculating a vector
Sum vector
Outer product and vector of
Sum vector
While computing the vector
Sum vector
Outer product and vector ofSum vector
If both of the above products are less than 0, it means that the line segment L1L2 intersects the line segment AC, otherwise, it means that the line segment L1L2 does not intersect the line segment AC;
and S314, repeating the steps S312 and S313, sequentially calculating the relation between the line segment AB, the line segment CD and the line segment BD and the line segment L1L2, and regarding that the pedestrian or the vehicle in the detection frame has the behavior of crossing the warning line as long as a certain line segment is intersected with the line segment L1L2 in the detection frame.
6. The method and process according to claim 4, wherein the step S32 is specifically as follows:
s321, judging the rule of the bidirectional tripwire: the same as the implementation step of S31, searching from the result returned in S31, if a pedestrian or a vehicle crosses the warning line and the tripwire type in the configuration file is selected as a bidirectional tripwire, directly alarming and transmitting the result back to the client;
s322, judging a unidirectional tripwire rule: based on S321, further judgment is made, and two endpoints L1(a1, B1) and L2(a2, B2) of the alarm ring line and the traveling direction of the object at a predetermined time are set, B (x1, y1) and a (x2, y2) are set, whether the alarm ring line meets the tripwire rule is judged by using the fact that the vector outer product has directivity, and the vector is calculatedAnd
outer product of (i.e.)
If vectorAndif the outer product of the alarm is larger than zero, the pedestrian or the vehicle is judged to have the behavior of crossing the alarm line in one direction, and then the result is transmitted back to the client side to give an alarm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910836375.4A CN110674703A (en) | 2019-09-05 | 2019-09-05 | Video tripwire alarm counting method and flow in intelligent monitoring |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910836375.4A CN110674703A (en) | 2019-09-05 | 2019-09-05 | Video tripwire alarm counting method and flow in intelligent monitoring |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110674703A true CN110674703A (en) | 2020-01-10 |
Family
ID=69076024
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910836375.4A Pending CN110674703A (en) | 2019-09-05 | 2019-09-05 | Video tripwire alarm counting method and flow in intelligent monitoring |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110674703A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111432172A (en) * | 2020-03-20 | 2020-07-17 | 浙江大华技术股份有限公司 | Fence alarm method and system based on image fusion |
| CN112449093A (en) * | 2020-11-05 | 2021-03-05 | 北京德火科技有限责任公司 | Three-dimensional panoramic video fusion monitoring platform |
| CN113205704A (en) * | 2021-03-19 | 2021-08-03 | 深圳市点创科技有限公司 | Blind area detection method and device for large vehicle and storage medium |
| CN114627155A (en) * | 2022-03-18 | 2022-06-14 | 广州云从人工智能技术有限公司 | Passenger flow statistical method, system, device and medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080106599A1 (en) * | 2005-11-23 | 2008-05-08 | Object Video, Inc. | Object density estimation in video |
| CN103456024A (en) * | 2012-06-02 | 2013-12-18 | 浙江西谷数字技术有限公司 | Moving object line crossing judgment method |
| CN103544806A (en) * | 2013-10-31 | 2014-01-29 | 江苏物联网研究发展中心 | Important cargo transportation vehicle monitoring and prewarning system based on video tripwire rule |
| CN104680557A (en) * | 2015-03-10 | 2015-06-03 | 重庆邮电大学 | Intelligent detection method for abnormal behavior in video sequence image |
| CN105959655A (en) * | 2016-07-18 | 2016-09-21 | 四川君逸数码科技股份有限公司 | Alarm method and device for identifying region invasion by intelligent eye |
| CN108416250A (en) * | 2017-02-10 | 2018-08-17 | 浙江宇视科技有限公司 | Demographic method and device |
| CN108596129A (en) * | 2018-04-28 | 2018-09-28 | 武汉盛信鸿通科技有限公司 | A kind of vehicle based on intelligent video analysis technology gets over line detecting method |
| CN109117702A (en) * | 2018-06-12 | 2019-01-01 | 深圳中兴网信科技有限公司 | The detection and count tracking method and system of target vehicle |
| CN109886117A (en) * | 2019-01-21 | 2019-06-14 | 青岛海信网络科技股份有限公司 | A kind of method and apparatus of goal behavior detection |
-
2019
- 2019-09-05 CN CN201910836375.4A patent/CN110674703A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080106599A1 (en) * | 2005-11-23 | 2008-05-08 | Object Video, Inc. | Object density estimation in video |
| CN103456024A (en) * | 2012-06-02 | 2013-12-18 | 浙江西谷数字技术有限公司 | Moving object line crossing judgment method |
| CN103544806A (en) * | 2013-10-31 | 2014-01-29 | 江苏物联网研究发展中心 | Important cargo transportation vehicle monitoring and prewarning system based on video tripwire rule |
| CN104680557A (en) * | 2015-03-10 | 2015-06-03 | 重庆邮电大学 | Intelligent detection method for abnormal behavior in video sequence image |
| CN105959655A (en) * | 2016-07-18 | 2016-09-21 | 四川君逸数码科技股份有限公司 | Alarm method and device for identifying region invasion by intelligent eye |
| CN108416250A (en) * | 2017-02-10 | 2018-08-17 | 浙江宇视科技有限公司 | Demographic method and device |
| CN108596129A (en) * | 2018-04-28 | 2018-09-28 | 武汉盛信鸿通科技有限公司 | A kind of vehicle based on intelligent video analysis technology gets over line detecting method |
| CN109117702A (en) * | 2018-06-12 | 2019-01-01 | 深圳中兴网信科技有限公司 | The detection and count tracking method and system of target vehicle |
| CN109886117A (en) * | 2019-01-21 | 2019-06-14 | 青岛海信网络科技股份有限公司 | A kind of method and apparatus of goal behavior detection |
Non-Patent Citations (1)
| Title |
|---|
| 黄孝建: "基于视频图像的绊线检测方法研究", 《计算机与现代化》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111432172A (en) * | 2020-03-20 | 2020-07-17 | 浙江大华技术股份有限公司 | Fence alarm method and system based on image fusion |
| CN112449093A (en) * | 2020-11-05 | 2021-03-05 | 北京德火科技有限责任公司 | Three-dimensional panoramic video fusion monitoring platform |
| CN113205704A (en) * | 2021-03-19 | 2021-08-03 | 深圳市点创科技有限公司 | Blind area detection method and device for large vehicle and storage medium |
| CN114627155A (en) * | 2022-03-18 | 2022-06-14 | 广州云从人工智能技术有限公司 | Passenger flow statistical method, system, device and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110674703A (en) | Video tripwire alarm counting method and flow in intelligent monitoring | |
| CN110428522B (en) | Intelligent security system of wisdom new town | |
| CN112052804B (en) | Video intelligent analysis and alarm system for realizing safety management and method thereof | |
| CN108965826A (en) | Monitoring method, device, processing equipment and storage medium | |
| CN107222660B (en) | Distributed network vision monitoring system | |
| CN110866642A (en) | Security monitoring method and device, electronic equipment and computer readable storage medium | |
| JP5992681B2 (en) | Unusual condition detection system for congestion | |
| CN111223260A (en) | Method and system for intelligently monitoring goods theft prevention in warehousing management | |
| CN111311630A (en) | Method and system for intelligently counting quantity of goods through videos in warehousing management | |
| CN110659391A (en) | Video detection method and device | |
| CN106331633A (en) | Method and system for displaying and quickly accessing a variety of monitoring resources | |
| CN112422909B (en) | Video behavior analysis management system based on artificial intelligence | |
| CN110491060A (en) | A kind of robot and its method for safety monitoring, device and storage medium | |
| CN103281518A (en) | Multifunctional networking all-weather intelligent video monitoring system | |
| CN112087604A (en) | Intelligent monitoring video management and control method based on image recognition | |
| CN115103157A (en) | Video analysis method and device based on edge cloud cooperation, electronic equipment and medium | |
| CN105095891A (en) | Human face capturing method, device and system | |
| CN114140719A (en) | AI traffic video analysis technology | |
| CN107301373B (en) | Data processing method, device and storage medium | |
| CN107920224A (en) | A kind of abnormality alarming method, equipment and video monitoring system | |
| Singh et al. | An intelligent video surveillance system using edge computing based deep learning model | |
| CN114187666B (en) | Identification method and system for watching mobile phone while walking | |
| CN116187634A (en) | Intelligent queuing system and prediction method for same | |
| CN114565870A (en) | Production line control method, device and system based on vision, and electronic equipment | |
| CN113286119A (en) | Unity 3D-based warehouse digital twinning system, method and apparatus |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200110 |