CN112804487B - Double-path video target tracking method and system - Google Patents
Double-path video target tracking method and system Download PDFInfo
- Publication number
- CN112804487B CN112804487B CN202011617873.9A CN202011617873A CN112804487B CN 112804487 B CN112804487 B CN 112804487B CN 202011617873 A CN202011617873 A CN 202011617873A CN 112804487 B CN112804487 B CN 112804487B
- Authority
- CN
- China
- Prior art keywords
- video
- path
- target
- tracking
- size
- 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
Images
Classifications
-
- 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
- H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/67—Focus control based on electronic image sensor signals
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/695—Control of camera direction for changing a field of view, e.g. pan, tilt or based on tracking of objects
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Studio Devices (AREA)
Abstract
The invention provides a method and a system for tracking a two-way video target, wherein the two-way video is a first-way video and a second-way video, the field angles of the first-way video and the second-way video are equal when the first-way video and the second-way video start to be monitored normally, and the method and the system can be used for: when an alarm target is detected in a first path of video, starting a target tracking mode of the first path of video and starting a target detection mode of a second path of video; in the process of starting the target tracking mode of the first path of video and the target detection mode of the second path of video, and in the process of starting the target detection mode of the first path of video and the target tracking mode of the second path of video, the mutual switching between the target tracking mode and the target detection mode of the first path of video and the target tracking mode of the second path of video can be realized based on the difference value of the field angles of the first path of video and the second path of video. The method and the device are used for improving the target tracking effect.
Description
Technical Field
The invention relates to the field of photoelectric monitoring, in particular to a double-channel video target tracking method and a double-channel video target tracking system, which are mainly used for target tracking of a photoelectric monitoring system.
Background
The photoelectric monitoring system is widely applied to remote monitoring scenes. The photoelectric monitoring system mainly comprises a visible light camera assembly, a thermal imaging camera assembly, a rotary table servo control unit and an information processing assembly. The visible light camera and the thermal imaging camera are installed in the same direction and can rotate horizontally or vertically in the same direction under the drive of the turntable servo motor. The visible light camera assembly collects visible light videos and the thermal imaging camera assembly collects thermal imaging videos.
When a suspicious target (i.e. an alarm target) is found, a conventional target tracking method, such as a target tracking method based on image cross-correlation (i.e. a KCF target tracking method), often adopts tracking of a single visible light video or a single thermal imaging video, i.e. tracking is usually performed in a single visible light or a single thermal imaging video. In the process of moving the warning target from far to near or from near to far, when the target is enlarged, the actually tracked area is only one part of the warning target, when the target is reduced, the actually tracked area contains one part of background information, and if the warning target is continuously reduced in the picture, the background information accounts for a higher rate than the warning target information, so that the warning target is easy to lose. Although the video picture can be adjusted by manually zooming in or zooming out the lens focal length, the video picture is often changed greatly by manually zooming in or zooming out the lens focal length, so that the size of the target is changed relatively greatly, and therefore, a conventionally adopted target tracking algorithm (such as a KCF (Kernel Correlation Filter) target tracking algorithm) is often unable to adapt to the change of the size of the alarm target, and the tracking loss of the alarm target is also easily caused.
Therefore, the invention provides a method and a system for tracking a two-way video target, which are used for solving the problems.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a two-way video target tracking method and a two-way video target tracking system, which are used for improving the target tracking effect.
In a first aspect, the invention provides a two-way video target tracking method, where the two-way video is a first way video and a second way video, the field angles of the first way video and the second way video when starting normal monitoring are equal, and both the first way video and the second way video are provided with a target tracking mode and a target detection mode;
the double-path video target tracking method comprises the following steps:
starting target detection in the first path of video for detecting an alarm target;
when the alarm target is detected in the first path of video, starting a target tracking mode of the first path of video, and carrying out video tracking on the detected alarm target based on the first path of video;
when an alarm target is detected in the first path of video, a target detection mode of the second path of video is started, the size of the alarm target in the second path of video is detected in real time, and the video focal length of the second path of video is adjusted according to the detected size, so that the size of the alarm target reaches a first preset proportion of the video picture size of the second path of video;
in the process of starting a target tracking mode of a first path of video and starting a target detection mode of a second path of video, calculating the field angles of the first path of video and the second path of video in real time, and when the difference value of the field angles of the first path of video and the second path of video reaches a preset first field angle difference value threshold value, switching the target tracking mode of the first path of video into a target detection mode, detecting the size of a warning target in the first path of video in real time, and adjusting the video focal length of the first path of video according to the detected size to enable the size of the warning target to reach a second preset proportion of the video picture size of the first path of video; switching the target detection mode of the second path of video into a target tracking mode, and continuously tracking the video of the alarm target based on the second path of video;
in the process of starting a target detection mode in the first path of video and starting a target tracking mode in the second path of video, calculating the field angles of the first path of video and the second path of video in real time, and switching the target detection mode of the first path of video into the target tracking mode when the difference value of the field angles of the first path of video and the second path of video reaches a preset second field angle difference value threshold value, and continuing to perform video tracking on the alarm target based on the first path of video; and switching the target tracking mode of the second path of video into a target detection mode, continuously detecting the size of the alarm target in the second path of video in real time, and adjusting the video focal length of the second path of video according to the detected size to enable the size of the alarm target to reach a first preset proportion of the video picture size of the second path of video.
Further, the starting of the target tracking mode of the first path of video and the video tracking of the detected alarm target based on the first path of video includes: and starting a target tracking mode of the first path of video, framing the alarm target detected in the current frame of the first path of video to obtain the position information and the size of the alarm target under the current frame, and then performing video tracking on the alarm target selected by the frame based on the first path of video by utilizing a KCF tracking algorithm.
Further, performing video tracking on the alarm target selected by the frame based on the first path of video by using a KCF tracking algorithm, wherein the video tracking comprises the following steps:
when the KCF tracking algorithm outputs new position information of the alarm target every time, calculating the miss distance of the alarm target based on the new position information and the previous position information of the alarm target corresponding to the new position information, and driving the rotary table to rotate along with the alarm target according to the miss distance obtained through calculation.
Further, the first path of video and the second path of video are visible light video and thermal imaging video.
Further, in the two-way video target tracking method:
target detection in the first path of video is realized by adopting a target detection algorithm based on deep learning and is used for finding a warning target and detecting the size of the warning target in the first path of video;
and the target detection in the second path of video is realized by adopting a target detection algorithm based on deep learning and is used for detecting the size of the alarm target in the second path of video.
In a second aspect, the invention provides a two-way video target tracking system, where the two-way video is a first way video and a second way video, the field angles of the first way video and the second way video when starting normal monitoring are equal, and both the first way video and the second way video are provided with a target tracking mode and a target detection mode; this two-way video target tracking system includes:
the first target detection unit is used for detecting an alarm target in the first path of video;
the second target detection unit is used for detecting the size of the alarm target in the second path of video;
the first path of video tracking unit is used for starting a target tracking mode of the first path of video when the alarm target is detected in the first path of video and carrying out video tracking on the detected alarm target based on the first path of video;
the second path of video tracking unit is used for starting a target detection mode of the second path of video when the alarm target is detected in the first path of video, calling the second target detection unit to detect the size of the alarm target in the second path of video in real time, and adjusting the video focal length of the second path of video according to the detected size so that the size of the alarm target reaches a first preset proportion of the video picture size of the second path of video;
the first video tracking switching unit is used for calculating the field angles of the first path of video and the second path of video in real time in the process of starting a target tracking mode of the first path of video and starting a target detection mode of the second path of video, and switching the target tracking mode of the first path of video into a target detection mode when the difference value of the field angles of the first path of video and the second path of video reaches a preset first field angle difference value threshold value, calling the first target detection unit to detect the size of the alarm target in the first path of video in real time, and adjusting the video focal length of the first path of video according to the detected size to enable the size of the alarm target to reach a second preset proportion of the video picture size of the first path of video; switching the target detection mode of the second path of video into a target tracking mode, and continuously tracking the video of the alarm target based on the second path of video;
the second video tracking switching unit is used for calculating the field angles of the first video and the second video in real time in the process of starting the target detection mode in the first video and starting the target tracking mode in the second video, and switching the target detection mode of the first video into the target tracking mode when the difference value of the field angles of the first video and the second video reaches a preset second field angle difference threshold value, and continuously carrying out video tracking on the alarm target based on the first video; and switching the target tracking mode of the second path of video into a target detection mode, continuously calling a second target detection unit to detect the size of the alarm target in the second path of video in real time, and adjusting the video focal length of the second path of video according to the detected size to enable the size of the alarm target to reach a first preset proportion of the video picture size of the second path of video.
Further, the first path video tracking unit includes:
the first target tracking mode automatic starting module is used for automatically starting the target tracking mode of the first path of video when the alarm target is detected in the first path of video;
the first target detection module is used for framing the alarm target detected in the current frame of the first path of video after the first target tracking mode automatic starting module starts the target tracking mode of the first path of video, so as to obtain the position information and the size of the alarm target under the current frame;
and the first path video tracking module is used for performing video tracking on the alarm target selected from the middle frame of the first target detection module based on the position information and the size of the alarm target acquired by the first target detection module by utilizing a KCF tracking algorithm.
Further, the video tracking module of the first path performs video tracking on the alarm target selected by the frame in the first target detection module based on the video of the first path by using a KCF tracking algorithm, and includes:
when a KCF tracking algorithm outputs new position information of an alarming target each time, calculating the miss distance of the alarming target based on the new position information and the last position information of the alarming target corresponding to the new position information, and driving the turntable to rotate along with the alarming target according to the miss distance obtained by calculation.
Further, the first path of video and the second path of video are visible light video and thermal imaging video.
Further, the first target detection unit detects the alarm target in the first video by adopting a target detection algorithm based on deep learning, and the detection comprises the steps of finding the alarm target in the first video and detecting the size of the alarm target in the first video; and the second target detection unit detects the size of the alarm target in the second path of video by adopting a target detection algorithm based on deep learning.
The invention has the beneficial effects that,
(1) The two-way video target tracking method and the two-way video target tracking system provided by the invention have the advantages that the two-way video target detection and the target tracking can be mutually switched, the defect that a target tracking algorithm (such as a KCF target tracking algorithm) cannot adapt to the target size is favorably eliminated, meanwhile, in the target tracking process, one-way video can be always ensured to be used for detecting the target size, the video focal length of the one-way video is adjusted according to the detected size, the size of an alarm target reaches the preset proportion of the video picture size of the one-way video, the image virtual focus caused in the video zooming process cannot influence the target tracking, and the target tracking effect is favorably optimized.
(2) The two-way video target tracking method and system provided by the invention have the advantages that the first preset proportion and the second preset proportion are used for adjusting the size of the alarm target, particularly the size of the alarm target which becomes relatively larger or relatively smaller in the process of moving from far to near or moving from near to far is adjusted, the size of the alarm target is conveniently adjusted to enable the size of the alarm target to reach the proper proportion of the size of a video picture, the loss of the alarm target can be avoided to a certain extent, and the integral video definition of the tracked alarm target is conveniently adjusted to a certain extent.
In addition, the invention has reliable design principle, simple structure and very wide application prospect.
Drawings
In order to more clearly illustrate the embodiments or technical solutions in the prior art of the present invention, the drawings used in the description of the embodiments or prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained based on these drawings without creative efforts.
FIG. 1 is a schematic flow diagram of a method of one embodiment of the invention.
FIG. 2 is a schematic block diagram of a system of one embodiment of the present invention.
Detailed Description
In order to make those skilled in the art better understand the technical solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in 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.
FIG. 1 is a schematic flow diagram of a method of one embodiment of the invention. The two paths of videos are a first path of video and a second path of video, the field angles of the first path of video and the second path of video are equal when the first path of video and the second path of video start normal monitoring, and the first path of video and the second path of video are both provided with a target tracking mode and a target detection mode.
As shown in fig. 1, the method 100 includes:
step 110, starting target detection in the first path of video for detecting an alarm target;
step 120, when the alarm target is detected in the first path of video, starting a target tracking mode of the first path of video, and performing video tracking on the detected alarm target based on the first path of video;
step 130, when an alarm target is detected in the first path of video, a target detection mode of the second path of video is started, the size of the alarm target in the second path of video is detected in real time, and the video focal length of the second path of video is adjusted according to the detected size, so that the size of the alarm target reaches a first preset proportion of the video picture size of the second path of video;
step 140, in the process of starting a target tracking mode for the first path of video and starting a target detection mode for the second path of video, calculating the field angles of the first path of video and the second path of video in real time, and when the difference value of the field angles of the first path of video and the second path of video reaches a preset first field angle difference value threshold value, switching the target tracking mode of the first path of video to be a target detection mode, detecting the size of the alarm target in the first path of video in real time, and adjusting the video focal length of the first path of video according to the detected size to enable the size of the alarm target to reach a second preset proportion of the video picture size of the first path of video; switching the target detection mode of the second path of video into a target tracking mode, and continuously carrying out video tracking on the alarm target based on the second path of video;
step 150, in the process of starting a target detection mode in the first path of video and starting a target tracking mode in the second path of video, calculating the field angles of the first path of video and the second path of video in real time, and when the difference value of the field angles of the first path of video and the second path of video reaches a preset second field angle difference threshold value, switching the target detection mode of the first path of video into the target tracking mode, and continuing to perform video tracking on the alarm target based on the first path of video; and switching the target tracking mode of the second path of video into a target detection mode, continuously detecting the size of the alarm target in the second path of video in real time, and adjusting the video focal length of the second path of video according to the detected size to enable the size of the alarm target to reach a first preset proportion of the video picture size of the second path of video.
Optionally, as an embodiment of the present invention, in step 120, the starting a target tracking mode of the first path of video, and performing video tracking on the detected alarm target based on the first path of video includes:
starting a target tracking mode of the first path of video, framing the alarm target detected in the current frame of the first path of video to obtain the position information and the size of the alarm target under the current frame, and then performing video tracking on the alarm target selected by the frame based on the first path of video by utilizing a KCF tracking algorithm.
Optionally, as an embodiment of the present invention, the performing video tracking on the alarm target selected by the frame based on the first path of video by using a KCF tracking algorithm includes:
when the KCF tracking algorithm outputs new position information of the alarm target every time, calculating the miss distance of the alarm target based on the new position information and the previous position information of the alarm target corresponding to the new position information, and driving the rotary table to rotate along with the alarm target according to the miss distance obtained through calculation.
Optionally, as an embodiment of the present invention, the first path of video and the second path of video are visible light video and thermal imaging video.
Optionally, as an embodiment of the present invention, target detection in the first video is implemented by using a target detection algorithm based on deep learning, and is used to find an alarm target and detect a size of the alarm target in the first video; and the target detection in the second path of video is realized by adopting a target detection algorithm based on deep learning and is used for detecting the size of the alarm target in the second path of video.
In order to facilitate understanding of the present invention, the following further describes the two-way video target tracking method provided by the present invention with reference to the principle of the two-way video target tracking method of the present invention and the process of tracking the alarm target in the embodiment.
Specifically, for example, a first path of video is a visible light video (acquired by an optical camera component), a second path of video is a thermal imaging video (acquired by a thermal imaging camera component), it is assumed that the visible light video is a channel a, the thermal imaging video is a channel B, the field angles of the channel a and the channel B when the channel a and the channel B start normal monitoring are equal, and both the channel a and the channel B are provided with a target tracking mode and a target detection mode, where the two-path video target tracking method is applied to a photoelectric monitoring system and specifically includes:
step S1: and starting the target in the channel A to detect the alarm target in the channel A.
Wherein the target detection is initiated in channel a, i.e. the target detection mode of channel a is initiated.
And the target detection mode of the channel A is used for detecting the alarm target in the channel A.
And starting target detection in the channel A, and detecting the alarm target in the channel A.
When the alarm target is detected in the channel A, the step S2 and the step S3 are executed.
Step S2: and when the alarm target is detected in the channel A, starting a target tracking mode of the channel A, and carrying out video tracking on the detected alarm target based on the channel A.
And after the target tracking mode of the channel A is started, video tracking can be carried out on the detected alarm target in the channel A.
And step S3: when the alarm target is detected in the channel a, the target detection mode of the channel B is started, the size of the alarm target in the channel B is detected in real time (in this embodiment, at a frame rate of 25FPS, referred to as "real time"), and the video focal length of the channel B is adjusted according to the detected size, so that the size of the alarm target reaches a first preset proportion of the video picture size of the channel B.
For example, the video picture of channel B has a size D B The first preset ratio is recorded as mu, 0<μ<1, obtaining the channel B center information after the video focal length of the channel B is adjusted at the current timeThe size of the target is D, then D = D B ×μ。
In a specific implementation, a specific value of the first preset ratio μmay be selected and set by a person skilled in the art according to an actual scene.
When the alarm target is detected in the channel a in the step S1, the target tracking mode for opening the channel a in the step S2 and the target detection mode for opening the channel B in the step S3 are simultaneously executed.
And step S4: in the process that the channel A is opened in a target tracking mode and the channel B is opened in a target detection mode, the field angles of the channel A and the channel B are calculated in real time, and when the difference value of the field angles of the channel A and the channel B reaches a preset first field angle difference value threshold, the following two aspects are executed:
on one hand, switching the target tracking mode of the channel A into a target detection mode, detecting the size of the alarm target in the channel A in real time, and adjusting the video focal length of the channel A according to the detected size to enable the size of the alarm target to reach a second preset proportion of the video picture size of the channel A;
and on the other hand, switching the target detection mode of the channel B into a target tracking mode, and continuously carrying out video tracking on the alarm target based on the channel B.
In step S4, the target tracking mode of the switching channel a is the target detection mode, and the target detection mode of the switching channel B is the target tracking mode, and the target tracking mode and the target detection mode are executed in parallel/simultaneously.
The target tracking mode of the switching channel a is a target detection mode, and specifically includes: and switching the working mode of the channel A from the target tracking mode to the target detection mode.
The target detection mode of the switching channel B is a target tracking mode, and specifically includes: and switching the working mode of the channel B from a target detection mode to a target tracking mode.
After the working mode of the channel A is switched from the target tracking mode to the target detection mode: and detecting the size of the alarm target in the channel A in real time, and adjusting the video focal length of the channel A according to the detected size to enable the size of the alarm target to reach a second preset proportion of the video picture size of the channel A.
E.g., channel a, has a video picture size D A The second preset ratio is recorded as σ,0<σ<1, if the size of the alarm target obtained after the video focal length of the channel a is adjusted at the current time is k in the channel a, k = D A ×σ。
In a specific implementation, a specific value of the second preset ratio σ may be selected and set by a person skilled in the art according to an actual scene.
The first preset proportion and the second preset proportion are used for adjusting the size of the warning target, and particularly for the warning target which is relatively larger or relatively smaller in the process of moving from far to near or moving from near to far, the size of the warning target is conveniently adjusted to enable the size of the warning target to reach the proper proportion of the size of a video picture, so that the warning target can be prevented from being lost to a certain extent.
After the working mode of the channel B is switched from the target detection mode to the target tracking mode: and continuing to perform video tracking on the alarm target based on the channel B.
For example, after the working mode of the channel B is switched from the target detection mode to the target tracking mode, the KCF tracking algorithm is used to continue video tracking on the alarm target based on the channel B.
After which step S5 is performed.
Step S5: when the target detection mode is started in the channel A and the target tracking mode is started in the channel B, in the process that the target detection mode is started in the channel A and the target tracking mode is started in the channel B, the field angles of the channel A and the channel B are calculated in real time, and when the difference value of the field angles of the channel A and the channel B reaches a preset second field angle difference threshold value, the following two aspects are executed:
on the first hand, switching the target detection mode of the channel A into a target tracking mode, and continuously carrying out video tracking on the alarm target based on the channel A;
and in the second aspect, switching the target tracking mode of the channel B into a target detection mode, continuously detecting the size of the alarm target in the channel B in real time, and adjusting the video focal length of the channel B according to the detected size to enable the size of the alarm target to reach the preset proportion threshold range of the video picture size of the channel B.
In step S5, the target detection mode of the switching channel a is the target tracking mode and the target tracking mode of the switching channel B is the target detection mode, and the two modes are executed simultaneously.
Based on this step S5, the change in the size of the alert object in channel B is detected in channel B, and the focal length is automatically adjusted so that the size of the alert object reaches the set proportion of the video screen size of channel B.
Step S6: when the target detection mode of the switching channel a is the target tracking mode and the target tracking mode of the switching channel B is the target detection mode in step S5, execution of step S4 is continued instead.
Therefore, the method 100 adopts two-way video for target tracking, can effectively reduce the loss probability of alarm target tracking, and is beneficial to improving the target tracking effect.
Fig. 2 is a diagram of an embodiment of a two-way video target tracking system according to the present invention. The two paths of videos are a first path of video and a second path of video, the field angles of the first path of video and the second path of video are equal when the first path of video and the second path of video start to be monitored normally, and the first path of video and the second path of video are both provided with a target tracking mode and a target detection mode.
As shown in fig. 2, the system 200 includes:
a first target detection unit 201, configured to detect an alarm target in a first path of video;
the second target detection unit 202 is used for detecting the size of the alarm target in the second path of video;
the first path video tracking unit 203 is configured to, when an alarm target is detected in the first path video, start a target tracking mode of the first path video, and perform video tracking on the detected alarm target based on the first path video;
the second path video tracking unit 204 is configured to, when the warning target is detected in the first path video, start a target detection mode of the second path video, invoke the second target detection unit 202 to detect the size of the warning target in the second path video in real time, and adjust the video focal length of the second path video according to the detected size, so that the size of the warning target reaches a first preset proportion of the video picture size of the second path video;
the first video tracking switching unit 205 is configured to calculate field angles of the first video and the second video in real time in a process of starting a target tracking mode of the first video and starting a target detection mode of the second video, and when a difference value between the field angles of the first video and the second video reaches a preset first field angle difference threshold value, switch the target tracking mode of the first video to a target detection mode, call the first target detection unit 201 to detect a size of an alarm target in the first video in real time, and adjust a video focal length of the first video according to the detected size, so that the size of the alarm target reaches a second preset proportion of a video picture size of the first video; switching the target detection mode of the second path of video into a target tracking mode, and continuously carrying out video tracking on the alarm target based on the second path of video;
the second video tracking switching unit 206 is configured to calculate, in real time, field angles of the first path of video and the second path of video in a process of starting a target detection mode in the first path of video and starting a target tracking mode in the second path of video, and switch the target detection mode of the first path of video to the target tracking mode and continue to perform video tracking on the alarm target based on the first path of video when a difference value between the field angles of the first path of video and the second path of video reaches a preset second field angle difference threshold value; and switching the target tracking mode of the second path of video into a target detection mode, continuously calling the second target detection unit 202 to detect the size of the alarm target in the second path of video in real time, and adjusting the video focal length of the second path of video according to the detected size to enable the size of the alarm target to reach a first preset proportion of the video picture size of the second path of video.
Optionally, as an embodiment of the present invention, the first way video tracking unit 203 includes:
the first target tracking mode automatic starting module is used for automatically starting the target tracking mode of the first path of video when the alarm target is detected in the first path of video;
the first target detection module is used for framing the alarm target detected in the current frame of the first path of video after the first target tracking mode automatic starting module starts the target tracking mode of the first path of video, so as to obtain the position information and the size of the alarm target under the current frame;
and the first path video tracking module is used for performing video tracking on the alarm target selected from the middle frame of the first target detection module in the first path video by utilizing a KCF tracking algorithm based on the position information and the size of the alarm target acquired by the first target detection module.
Optionally, as an embodiment of the present invention, the performing, by a first path video tracking module, video tracking on an alarm target selected from a frame in a first target detection module based on a first path video by using a KCF tracking algorithm includes:
when a KCF tracking algorithm outputs new position information of an alarming target each time, calculating the miss distance of the alarming target based on the new position information and the last position information of the alarming target corresponding to the new position information, and driving the turntable to rotate along with the alarming target according to the miss distance obtained by calculation.
Optionally, as an embodiment of the present invention, the first path of video and the second path of video are visible light video and thermal imaging video.
Optionally, as an embodiment of the present invention, the first target detecting unit 201 detects the warning object in the first video by using a target detection algorithm based on deep learning, including finding the warning object in the first video and detecting a size of the warning object in the first video; the second target detection unit 202 detects the size of the alarm target in the second path of video by using a target detection algorithm based on deep learning.
The same and similar parts in the various embodiments in this specification may be referred to each other. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and the relevant points can be referred to the description in the method embodiment.
In the several embodiments provided in the present invention, it should be understood that the disclosed system and method may be implemented in other manners. For example, the above-described system embodiments are merely illustrative, and for example, the division of the units is only one logical functional division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed.
It should be noted that, when the two-way video target tracking method and system are used, time limit can be preset, and when the tracking time reaches the set time limit, the tracking is stopped; space limitation can be preset, and tracking is stopped when the direction of the rotary table exceeds the set space limitation; and stopping tracking when the alarm target disappears in the video pictures of the first path of video and the second path of video without any limit.
Although the present invention has been described in detail by referring to the drawings in connection with the preferred embodiments, the present invention is not limited thereto. Various equivalent modifications or substitutions can be made on the embodiments of the present invention by those skilled in the art without departing from the spirit and scope of the present invention, and these modifications or substitutions should be within the scope of the present invention/any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present disclosure and 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 (10)
1. A two-path video target tracking method is characterized in that two paths of videos are a first path of video and a second path of video, the field angles of the first path of video and the second path of video when the first path of video and the second path of video start to be normally monitored are equal, and the first path of video and the second path of video are both provided with a target tracking mode and a target detection mode;
the double-path video target tracking method comprises the following steps:
starting target detection in the first path of video for detecting an alarm target;
when the alarm target is detected in the first path of video, starting a target tracking mode of the first path of video, and carrying out video tracking on the detected alarm target based on the first path of video;
when an alarm target is detected in the first path of video, a target detection mode of the second path of video is started, the size of the alarm target in the second path of video is detected in real time, and the video focal length of the second path of video is adjusted according to the detected size, so that the size of the alarm target reaches a first preset proportion of the video picture size of the second path of video;
in the process of starting a target tracking mode of a first path of video and starting a target detection mode of a second path of video, calculating the field angles of the first path of video and the second path of video in real time, and when the difference value of the field angles of the first path of video and the second path of video reaches a preset first field angle difference value threshold value, switching the target tracking mode of the first path of video into a target detection mode, detecting the size of an alarm target in the first path of video in real time, and adjusting the video focal length of the first path of video according to the detected size to enable the size of the alarm target to reach a second preset proportion of the video picture size of the first path of video; switching the target detection mode of the second path of video into a target tracking mode, and continuously carrying out video tracking on the alarm target based on the second path of video;
in the process of starting a target detection mode in the first path of video and starting a target tracking mode in the second path of video, calculating the field angles of the first path of video and the second path of video in real time, and when the difference value of the field angles of the first path of video and the second path of video reaches a preset second field angle difference value threshold value, switching the target detection mode of the first path of video into the target tracking mode, and continuing to perform video tracking on the alarm target based on the first path of video; and switching the target tracking mode of the second path of video into a target detection mode, continuously detecting the size of the alarm target in the second path of video in real time, and adjusting the video focal length of the second path of video according to the detected size to enable the size of the alarm target to reach a first preset proportion of the video picture size of the second path of video.
2. The method for tracking two-way video targets according to claim 1, wherein the starting of the target tracking mode of the first video and the video tracking of the detected alarm target based on the first video comprises: starting a target tracking mode of the first path of video, framing the alarm target detected in the current frame of the first path of video to obtain the position information and the size of the alarm target under the current frame, and then performing video tracking on the alarm target selected by the frame based on the first path of video by utilizing a KCF tracking algorithm.
3. The two-way video target tracking method according to claim 2, wherein the video tracking of the alarm target selected by the frame based on the first way video by using the KCF tracking algorithm comprises:
when a KCF tracking algorithm outputs new position information of an alarming target each time, calculating the miss distance of the alarming target based on the new position information and the last position information of the alarming target corresponding to the new position information, and driving the rotary table to rotate along with the alarming target according to the miss distance obtained through calculation.
4. The two-way video target tracking method according to any one of claims 1-3, wherein the first way video and the second way video are visible light video and thermal imaging video.
5. The dual-path video target tracking method according to any of claims 1-3,
target detection in the first path of video is realized by adopting a target detection algorithm based on deep learning and is used for finding a warning target and detecting the size of the warning target in the first path of video;
and the target detection in the second path of video is realized by adopting a target detection algorithm based on deep learning and is used for detecting the size of the alarm target in the second path of video.
6. A two-path video target tracking system is characterized in that two paths of videos are a first path of video and a second path of video, the field angles of the first path of video and the second path of video when the first path of video and the second path of video start to be normally monitored are equal, and the first path of video and the second path of video are both provided with a target tracking mode and a target detection mode; this double-circuit video target tracking system includes:
the first target detection unit is used for detecting an alarm target in the first path of video;
the second target detection unit is used for detecting the size of the alarm target in the second path of video;
the first path of video tracking unit is used for starting a target tracking mode of the first path of video when the alarm target is detected in the first path of video and carrying out video tracking on the detected alarm target based on the first path of video;
the second path of video tracking unit is used for starting a target detection mode of the second path of video when the alarm target is detected in the first path of video, calling the second target detection unit to detect the size of the alarm target in the second path of video in real time, and adjusting the video focal length of the second path of video according to the detected size so that the size of the alarm target reaches a first preset proportion of the video picture size of the second path of video;
the first video tracking switching unit is used for calculating the field angles of the first video and the second video in real time in the process of starting a target tracking mode of the first video and starting a target detection mode of the second video, and switching the target tracking mode of the first video into a target detection mode when the difference value of the field angles of the first video and the second video reaches a preset first field angle difference threshold value, calling the first target detection unit to detect the size of the alarm target in the first video in real time, and adjusting the video focal length of the first video according to the detected size to enable the size of the alarm target to reach a second preset proportion of the video picture size of the first video; switching the target detection mode of the second path of video into a target tracking mode, and continuously tracking the video of the alarm target based on the second path of video;
the second video tracking switching unit is used for calculating the field angles of the first path of video and the second path of video in real time in the process of starting a target detection mode in the first path of video and starting a target tracking mode in the second path of video, and switching the target detection mode of the first path of video into a target tracking mode and continuously carrying out video tracking on the alarm target based on the first path of video when the difference value of the field angles of the first path of video and the second path of video reaches a preset second field angle difference value threshold value; and switching the target tracking mode of the second path of video into a target detection mode, continuously calling a second target detection unit to detect the size of the alarm target in the second path of video in real time, and adjusting the video focal length of the second path of video according to the detected size to enable the size of the alarm target to reach a first preset proportion of the video picture size of the second path of video.
7. The dual-channel video target tracking system of claim 6, wherein the first video tracking unit comprises:
the first target tracking mode automatic starting module is used for automatically starting the target tracking mode of the first path of video when the alarm target is detected in the first path of video;
the first target detection module is used for framing the alarm target detected in the current frame of the first path of video after the first target tracking mode automatic starting module starts the target tracking mode of the first path of video, so as to obtain the position information and the size of the alarm target under the current frame;
and the first path video tracking module is used for performing video tracking on the alarm target selected from the middle frame of the first target detection module based on the position information and the size of the alarm target acquired by the first target detection module by utilizing a KCF tracking algorithm.
8. The dual-channel video target tracking system of claim 7, wherein the first channel video tracking module performs video tracking based on the alarm target selected from the frame of the first target detection module in the first channel video by using a KCF tracking algorithm, and the method comprises:
when a KCF tracking algorithm outputs new position information of an alarming target each time, calculating the miss distance of the alarming target based on the new position information and the last position information of the alarming target corresponding to the new position information, and driving the turntable to rotate along with the alarming target according to the miss distance obtained by calculation.
9. The dual-channel video target tracking system of claim 6, wherein the first channel of video and the second channel of video are visible light video and thermal imaging video.
10. The two-way video target tracking system according to any one of claims 6 to 9, wherein the first target detecting unit detects the warning target in the first way of video by using a target detection algorithm based on deep learning, including finding the warning target in the first way of video and detecting the size of the warning target in the first way of video; and the second target detection unit detects the size of the alarm target in the second path of video by adopting a target detection algorithm based on deep learning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011617873.9A CN112804487B (en) | 2020-12-30 | 2020-12-30 | Double-path video target tracking method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011617873.9A CN112804487B (en) | 2020-12-30 | 2020-12-30 | Double-path video target tracking method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112804487A CN112804487A (en) | 2021-05-14 |
| CN112804487B true CN112804487B (en) | 2022-10-28 |
Family
ID=75805952
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011617873.9A Active CN112804487B (en) | 2020-12-30 | 2020-12-30 | Double-path video target tracking method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112804487B (en) |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2866183B2 (en) * | 1990-10-12 | 1999-03-08 | コニカ株式会社 | Tracking camera |
| JP4568009B2 (en) * | 2003-04-22 | 2010-10-27 | パナソニック株式会社 | Monitoring device with camera cooperation |
| CN103581614A (en) * | 2012-08-01 | 2014-02-12 | 通号通信信息集团有限公司 | Method and system for tracking targets in video based on PTZ |
| CN103093654B (en) * | 2013-01-28 | 2014-12-03 | 北京航空航天大学 | Double video camera interactive intelligent tracking teaching system |
| US9684056B2 (en) * | 2014-05-29 | 2017-06-20 | Abdullah I. Khanfor | Automatic object tracking camera |
| CN104853150A (en) * | 2015-02-11 | 2015-08-19 | 居锦武 | Multi-camera objective cooperatively tracking technology |
| CN107024175A (en) * | 2017-06-20 | 2017-08-08 | 长春工程学院 | Motorbus vehicle body critical size detecting system solution based on multi-vision visual |
| CN108055501A (en) * | 2017-11-22 | 2018-05-18 | 天津市亚安科技有限公司 | A kind of target detection and the video monitoring system and method for tracking |
| CN111866437B (en) * | 2020-06-30 | 2022-04-08 | 厦门亿联网络技术股份有限公司 | Automatic switching method and device for double cameras of video conference, terminal equipment and storage medium |
-
2020
- 2020-12-30 CN CN202011617873.9A patent/CN112804487B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN112804487A (en) | 2021-05-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11290639B2 (en) | Enhanced image capture | |
| US9729784B2 (en) | Enhanced image capture | |
| EP3146709B1 (en) | Enhanced image capture | |
| US9774779B2 (en) | Enhanced image capture | |
| CN108198199B (en) | Moving object tracking method, moving object tracking device and electronic equipment | |
| KR20170033126A (en) | Method and apparatus of estimating a motion of an image, method and apparatus of image stabilization and computer-readable recording medium for executing the method | |
| JP2012186743A (en) | Photographing device with tracking function, and photographing method | |
| WO2012086188A1 (en) | Camera device, image processing system, image processing method and image processing program | |
| EP3358823A1 (en) | Method and device for switching between operation modes of video monitoring apparatus | |
| US20110228977A1 (en) | Image capturing device and method for adjusting a position of a lens of the image capturing device | |
| CN106488127A (en) | Camera switching control method based on football detecting and tracking and device | |
| KR20150130901A (en) | Camera apparatus and method of object tracking using the same | |
| CN112804487B (en) | Double-path video target tracking method and system | |
| KR20190026625A (en) | Image displaying method, Computer program and Recording medium storing computer program for the same | |
| US20120075467A1 (en) | Image capture device and method for tracking moving object using the same | |
| KR101619956B1 (en) | Apparatus for image processing of surveilance camera by using auto multi-tracing | |
| TWI476735B (en) | Abnormal classification detection method for a video camera and a monitering host with video image abnormal detection | |
| TW201415381A (en) | Abnormal detection method for a video camera | |
| WO2021138870A1 (en) | Multi-frame noise reduction method, terminal, and system | |
| US20240048844A1 (en) | Method and apparatus for shooting video, and electronic device | |
| Delis et al. | Automatic 3d defects identification in stereoscopic videos | |
| CN114641984A (en) | Video image display method and device | |
| JPH04148282A (en) | Image pickup state detecting method for picture image pickup device | |
| WO2010030170A2 (en) | A surveillance system | |
| WO2023134850A1 (en) | Image capturing apparatus and method for combining data from a digital camera and an event camera |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |