CN106355603B - Human body tracing method and human body tracking device - Google Patents

Human body tracing method and human body tracking device Download PDF

Info

Publication number
CN106355603B
CN106355603B CN201610752578.1A CN201610752578A CN106355603B CN 106355603 B CN106355603 B CN 106355603B CN 201610752578 A CN201610752578 A CN 201610752578A CN 106355603 B CN106355603 B CN 106355603B
Authority
CN
China
Prior art keywords
human body
tracked
tracking
human
detected
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
Application number
CN201610752578.1A
Other languages
Chinese (zh)
Other versions
CN106355603A (en
Inventor
卢宇
梁伯均
张伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Sensetime Technology Co Ltd
Original Assignee
Shenzhen Sensetime Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shenzhen Sensetime Technology Co Ltd filed Critical Shenzhen Sensetime Technology Co Ltd
Priority to CN201610752578.1A priority Critical patent/CN106355603B/en
Publication of CN106355603A publication Critical patent/CN106355603A/en
Application granted granted Critical
Publication of CN106355603B publication Critical patent/CN106355603B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/20Analysis of motion
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/10Image acquisition modality
    • G06T2207/10016Video; Image sequence
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30196Human being; Person
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30232Surveillance

Abstract

The invention discloses a kind of human body tracing methods and human body tracking device, wherein human body tracking comprises determining that tracked human body;The accumulative loss number of tracked human body is initialized as zero;The tracking information for being tracked human body is recorded in tracking list;Human testing is persistently carried out based on camera acquired image;When detecting human body and when the secondary human body detected is to be tracked human body, will update when the tracking information of the human body time detected into tracking list, and the accumulative loss number of tracked human body is reset;It is not to be tracked human body when can't detect human body or work as the secondary human body detected, and the accumulative loss number for being tracked human body is not up to the loss frequency threshold value, the accumulative loss number for being tracked human body is added one, and returns and executes above-mentioned the step of human testing is persistently carried out based on camera acquired image and subsequent step.Technical solution provided by the invention can be realized to be continued to track to by the human body with losing.

Description

Human body tracing method and human body tracking device
Technical field
The present invention relates to intelligent monitoring technology fields, and in particular to a kind of human body tracing method and human body tracking device.
Background technique
Human testing and tracking technique, which refer to, carries out human testing to current scene, automatic right when having detected human body Detection human body is tracked.Human testing and tracking technique can be widely applied to video monitoring, security protection, indoor and outdoor robot, different It is the technology with very big researching value and wide application prospect in the application such as normal behavioural analysis and stream of people's statistics.
Currently, common human body detecting method mainly has through camera collection image, then pass through feature extraction and identification Etc. detection techniques position human body.Wherein, common detection technique has based on histogram of gradients (HOG, Histogram of Ori Ented Gradients) with support vector machines (SVM, Support Vector Machine) combine method and deformable part The method etc. of part model (DPM, Deformable Part Model).
It is above-mentioned based on the human body detecting method of camera collection image after detecting human body, human body tracking technology can be passed through Realize automatically tracking for human body.Currently, common human body tracking technology has method based on contour feature, Camshift tracking to calculate The filter related to coring of method, mean shift algorithm (i.e. mean-shift), particle filter algorithm, optical flow method, Kalman filtering algorithm Wave device (KCF, Kernelized Correlation Filters) algorithm etc..These above-mentioned traditional human body tracking technologies are one Good tracking effect can be obtained in a little simple scenes, but when human body and background similarity with higher, human motion are fast The case where speed, human body are blocked or when human body attitudes changes greatly, and above-mentioned human body tracking technology is just easy to appear with losing.Further Ground, when human body is blocked or is appeared in the monitoring sight of camera again afterwards with losing because of other reasons, above-mentioned human body tracking Technology will be unable to continue to track the human body by with losing.
Summary of the invention
The present invention provides a kind of human body tracing method and human body tracking device, for realizing to by the continuation of the human body with losing Tracking.
First aspect present invention provides a kind of human body tracing method, comprising:
Determine tracked human body;
The accumulative loss number of above-mentioned tracked human body is initialized as zero;
The tracking information of above-mentioned tracked human body is recorded in tracking list;
Human testing is persistently carried out based on camera acquired image;
It, will be when the human body that time detect when detecting human body and when the secondary human body detected is above-mentioned tracked human body Tracking information is updated into above-mentioned tracking list, and when the accumulative loss number of above-mentioned tracked human body is not zero by above-mentioned quilt The accumulative loss number for tracking human body is reset;
When can't detect human body or when the secondary human body detected is not above-mentioned tracked human body, above-mentioned tracked people is detected Whether the accumulative loss number of body reaches preset loss frequency threshold value;
If the accumulative loss number of above-mentioned tracked human body reaches above-mentioned loss frequency threshold value, returns and execute above-mentioned determination The step of being tracked human body and subsequent step;
If the accumulative loss number of above-mentioned tracked human body is not up to above-mentioned loss frequency threshold value, by above-mentioned tracked people The accumulative loss number of body adds one, and returns and execute the above-mentioned step for persistently carrying out human testing based on camera acquired image Rapid and subsequent step.
Second aspect of the present invention provides a kind of human body tracking device, comprising:
Determination unit, for determining tracked human body;
Initialization unit, for the accumulative loss number of above-mentioned tracked human body to be initialized as zero;
Updating unit is tracked, for recording the tracking information of above-mentioned tracked human body in tracking list;
Human testing unit, for persistently carrying out human testing based on camera acquired image;
Number detection unit is lost, for can't detect human body or above-mentioned human testing unit when above-mentioned human testing unit When the secondary human body detected is not above-mentioned tracked human body, whether the accumulative loss number for detecting above-mentioned tracked human body reaches Preset loss frequency threshold value;
Accumulated unit is lost, for detecting the accumulative loss of above-mentioned tracked human body when above-mentioned loss number detection unit When number is not up to above-mentioned loss frequency threshold value, the accumulative loss number of above-mentioned tracked human body is added one;
Above-mentioned determination unit detects the accumulative loss number of above-mentioned tracked human body in above-mentioned loss number detection unit It is triggered again when reaching above-mentioned loss frequency threshold value;
Above-mentioned tracking updating unit is also used to: when above-mentioned human testing unit detects human body and above-mentioned human testing unit When the secondary human body detected is above-mentioned tracked human body, by above-mentioned human testing unit when the tracking of the secondary human body detected is believed Breath is updated into above-mentioned tracking list;
Above-mentioned initialization unit is also used to: when above-mentioned human testing unit detects human body and above-mentioned human testing unit is worked as The secondary human body detected is above-mentioned tracked human body, and when the accumulative loss number of above-mentioned tracked human body is not zero, will be above-mentioned The accumulative loss number for being tracked human body is reset.
Therefore the present invention initializes the accumulative loss number for being tracked human body after determining tracked human body And human testing is persistently carried out based on camera acquired image, it is not when can't detect human body or work as the secondary human body detected When above-mentioned tracked human body (when being tracked human body with losing), this is detected by the accumulative loss number with body of loseing face and whether reaches pre- If loss frequency threshold value, if not up to, which is added continue together keep to this by with The tracking of track human body re-starts the determination of tracked human body if reached.On the one hand, human body is tracked by introducing It is accumulative to lose number, it can realize with losing tracked human body and continue to track to by the human body with losing;Another party Frequency threshold value is lost by setting in face, can also be to avoid the figure for not appearing in camera acquisition for a long time because being tracked human body Lead to the problem of can not carrying out human body tracking for a long time as interior, improves the flexibility of human body tracking.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below There is attached drawing needed in technical description to be briefly described.It should be evident that the accompanying drawings in the following description is only this Some embodiments of invention without any creative labor, may be used also for those of ordinary skill in the art To obtain other drawings based on these drawings.
Fig. 1-a is a kind of human body tracing method one embodiment flow diagram provided by the invention;
Fig. 1-b is that the position coverage rate under a kind of application scenarios provided by the invention illustrates schematic diagram;
Fig. 2 is a kind of another embodiment flow diagram of human body tracing method provided by the invention;
Fig. 3 is a kind of human body tracking device one embodiment structural schematic diagram provided by the invention.
Specific embodiment
In order to make the invention's purpose, features and advantages of the invention more obvious and easy to understand, below in conjunction with the present invention Attached drawing in embodiment, technical scheme in the embodiment of the invention is clearly and completely described.Obviously, described reality Applying example is only a part of the embodiment of the present invention, and not all embodiments.Based on the embodiments of the present invention, the common skill in this field Art personnel every other embodiment obtained without making creative work belongs to the model that the present invention protects It encloses.
Embodiment one
Present example provides a kind of human body tracing method.Human body tracking side as shown in Fig. 1-a, in the embodiment of the present invention Method includes:
Step 101 determines tracked human body;
In a step 101, can continue to carry out human testing until detecting human body, the human body that will test later determines To be tracked human body.Alternatively, also can receive the human body image (including the image of human body) or other characteristics of human body letter of input Breath is determined according to the human body image or other characteristics of human body's information of input by terminal and is tracked human body.For example, user can be to end The human body image that any human body image or captured in real-time that end input is chosen from image library obtain, is based on triggering the terminal The human body image of input carries out human testing and is determined as the human body that human body image includes to be tracked human body.
Optionally, only 1 human body is determined as every time being tracked human body.Step 101 are as follows: persistently carry out human testing until Detect human body;If the number of the human body detected is 1, will be determined as being tracked human body when the secondary human body detected;If inspection The number of the human body measured is not 1, then will be determined as being tracked human body when a human body in the secondary multiple human bodies detected.Tool Body, when the number of the human body detected is not 1, the area when the secondary each human body detected can be calculated separately, by face The maximum human body of product is determined as being tracked human body.Below with based on camera acquired image persistently carry out human testing until It detects and is illustrated for human body.It, can be using more rapidly region convolutional neural networks for camera acquired image (i.e. Faster R-CNN, wherein the full name in English of R-CNN be Region-Convolutional Neural Network) or Other human testing algorithms based on image detect in the image with the presence or absence of human body, when detecting the human body number in the image When being 1, the human body that will test is determined as being tracked human body;It, can basis when detecting that the human body number in the image is greater than 1 Formula si=wi*hiEach human body area in the picture is calculated, is determined as the maximum human body of area to be tracked human body later.? In above-mentioned formula, siIndicate the area of human body i in the picture, wiIndicate the human body rectangle where the human body i detected in the picture The width of frame, hiIndicate the height of the human body rectangle frame where the human body i detected in the picture.
It should be noted that 1 human body is determined as being tracked human body every time by above scheme, certainly, in other embodiments In, N number of human body can also be determined as every time to be tracked human body, wherein N is the preset value greater than 2.In a step 101, work as inspection When measuring the number of human body and human body no more than N, the human body that will test all is determined as being tracked human body;When detecting human body And the number of human body be greater than N when, the N number of human body in human body that will test be determined as be tracked human body.The embodiment of the present invention is not Specific implementation used by tracked human body determining in step 101 is defined.
The accumulative loss number of above-mentioned tracked human body is initialized as zero by step 102;
In a step 102, the accumulative loss number for the tracked human body that step 101 determines is initialized as zero.
Specifically, the accumulative loss number for the tracked human body that can be determined by tabular form recording step 101.Example Such as, by losing the accumulative loss number for the tracked human body that list recording step 101 determines, then step 102 is embodied in Initialize above-mentioned loss list, wherein list is empty for the loss after initialization.
Further, which is tracked the accumulative loss number of human body in addition to recording, can also be in tracked human body After losing, the identification information for being tracked human body is stored, it, can be by the loss list in order to after tracked human body is with losing The identification information of the tracked human body of storage, judges the human body newly detected and the tracked human body recorded in the loss list is No is same human body.
Step 103, the tracking information that above-mentioned tracked human body is recorded in tracking list;
In step 103, the tracking information for the tracked human body that recording step 101 determines in tracking list.The tracking Information includes but is not limited to: being tracked the identification information of human body, is tracked the location information of human body.Wherein, this is tracked human body Identification information namely the information that is based on when detecting that this is tracked human body, which should Human body is tracked the location of when being detected.For example, when persistently carrying out human testing based on the image that camera acquires When, if detecting tracked human body from image A1, this is tracked the identification information of human body namely what image A1 was included is somebody's turn to do It is tracked the image section (such as image section in the human body rectangle frame in image A1 where tracked human body) of human body, the quilt The location information of tracking human body indicates that this is tracked human body coordinate position locating in image A1 (such as can will be in image A1 The left upper apex of human body rectangle frame and the coordinate of bottom right vertex where the tracked human body detected are considered as this and are tracked human body The locating coordinate position in image A1).
Step 104 persistently carries out human testing based on camera acquired image;
At step 104, human testing is persistently carried out based on camera acquired image, which can be terminal The camera configured in (such as monitor terminal, robot etc.), alternatively, being also possible to can be with other camera shooting of the terminal called Head instrument, is not construed as limiting herein.
Optionally, at step 104, it is based on camera acquired image, people is persistently carried out using Faster R-CNN Physical examination is surveyed, using the ability of the powerful extraction image further feature of convolutional neural networks, realize more accurate human testing and Tracking.Certainly, in the embodiment of the present invention, the camera can also be collected using other human testing algorithms based on image Image carry out human testing, herein without limitation.
Optionally, when detecting human body, judgement is when whether the human body that time detect is above-mentioned tracking list or above-mentioned is lost Lose the tracked human body of list records;If when the quilt that the secondary human body detected is above-mentioned tracking list or above-mentioned loss list records Human body is tracked, then determines that when the secondary human body detected be above-mentioned tracked human body;If when the human body that time detect be not it is above-mentioned with The tracked human body of track list records and be not above-mentioned loss list records tracked human body, then determine as the people that time detect Body is not above-mentioned tracked human body.In a kind of application scenarios, whether above-mentioned judgement is above-mentioned tracking when the secondary human body detected The tracked human body of list or above-mentioned loss list records may include: judgement when whether the secondary human body detected is above-mentioned tracking The tracked human body of list records;If sentencing when the secondary human body detected is not the tracked human body of above-mentioned tracking list records It is disconnected when the human body that time detect whether be above-mentioned loss list records tracked human body.That is, first judgement is as the secondary people detected Body whether be above-mentioned tracking list records tracked human body, be not above-mentioned tracking list records when the human body that time detect Be tracked human body, then judge when the human body that time detect whether be above-mentioned loss list records tracked human body.
Specifically, the embodiment of the present invention can pass through when above-mentioned tracking information includes the image of above-mentioned tracked human body As under type judgement when the human body time detected whether be above-mentioned tracking list records tracked human body: believed based on above-mentioned tracking The image for the above-mentioned tracked human body that breath includes, use preset human bioequivalence model judge when the human body that time detect whether as The tracked human body of above-mentioned tracking list records.Wherein, above-mentioned human bioequivalence model is obtained based on convolutional neural networks training, on Human bioequivalence model is stated for judging whether two human bodies belong to the same human body.Pass through a large amount of training data and is based on convolution Neural metwork training human body identification model enables to (judge two using human body identification model progress human bioequivalence It is the same human body that whether human body, which belongs to) identification accuracy be better than traditional recognition methods (such as side based on HOG feature Method and the method based on contour feature etc.).Similarly, can also use the above method to determine when the human body that time detect whether be on State the tracked human body for losing list records.
Further, in the case where above-mentioned tracking information further includes the location information for being tracked human body, above-mentioned judgement is worked as It may include: to remember according in above-mentioned tracking list that whether the secondary human body detected, which is the tracked human bodies of above-mentioned tracking list records, This recorded is tracked the location information of human body, calculates this and is tracked human body and when the position of the secondary all human bodies detected covers Rate;Position coverage rate is not less than the human body of preset coverage rate threshold value if it exists, then covers the position that human body is tracked with this The maximum tracked human body for being determined as the tracking list records when the secondary human body detected of rate;Position coverage rate is not if it does not exist Less than the human body of above-mentioned coverage rate threshold value, then the image for the above-mentioned tracked human body for including based on above-mentioned tracking information, in use State the judgement of human bioequivalence model when the human body that time detect whether be above-mentioned tracking list records tracked human body.Wherein, on Rheme sets coverage rate namely human body shared region (such as human body rectangle of the human body where in different images in different images Frame) overlapping area and non-overlap area ratio.Since step 104 is persistently to carry out people based on camera acquired image Physical examination survey, therefore, be tracked human body continuous acquisition to frame image in position should vary less, correspondingly, be tracked The overlapping area of human body region in two continuous frames image should be larger, correspondingly, being tracked human body in two continuous frames figure Position coverage rate as in also can be larger.Therefore in above scheme, before carrying out identification judgement using above-mentioned human bioequivalence model, It first passes through and calculates this and be tracked human body and the position coverage rate when all human bodies time detected, there are position coverage rate is not small It is when the human body of above-mentioned coverage rate threshold value, the position coverage rate that human body is tracked with this is maximum when the secondary human body detected is true It is set to the tracked human body of the tracking list records, carries out identifying that judgement is needed using above-mentioned human bioequivalence model so as to save The time to be expended improves human body tracking efficiency.Specifically, being illustrated in conjunction with Fig. 1-b to above-mentioned position coverage rate, if Fig. 1- C1 in b is the people where the human body (for ease of description, subsequent be described as human body S1 for the human body) detected in image A1 Body rectangle frame, C2 are human body (for ease of description, subsequent be described as human body S2 for the human body) place detected in image A2 Human body rectangle frame, coordinate position of the coordinate of the left upper apex and bottom right vertex that take human body rectangle frame as corresponding human body.Then As shown in Fig. 1-b, the coordinate position of human body S1 are as follows: left upper apex coordinate (xs11,ys11) and bottom right vertex coordinate (xs12,ys12), The coordinate position of human body S2 are as follows: left upper apex coordinate (xs21,ys21) and bottom right vertex coordinate (xs22,ys22), w1And h1It respectively indicates The width and height of C1, w2And h2Respectively indicate the width and height of C2, it is clear that w1=| xs12-xs11|,h1=| ys12-ys11|, w2 =| xs22-xs21|,h2=| ys22-ys21|.Then the overlapping area (subsequent to indicate the overlapping area with interSec) of C1 and C2 is i.e. Non-overlap area, that is, the sum of C1 and the area of C2 of the area in the region of oblique dotted line filling in Fig. 1-b, C1 and C2 subtract the overlapping Area (subsequent to indicate the non-overlap area with union), then the ratio of interSec and union is human body S1 and human body S2 Position coverage rate.The position coverage rate that human body S1 and human body S2 are indicated with overlap, then exist: overlap=interSec/ Union, wherein interSec=(x1+w1-x2)*(y1+h1-y2), union=w1*h1+w2*h21-interSec。
Step 105, when detecting human body and when the secondary human body detected is above-mentioned tracked human body, be detected when secondary The tracking information of human body update into above-mentioned tracking list, and when the accumulative loss number of above-mentioned tracked human body is not zero The accumulative loss number of above-mentioned tracked human body is reset;
In step 105, when detect human body and when the secondary human body detected be above-mentioned tracked human body when, when time examine The tracking information of the human body measured is updated into above-mentioned tracking list, later return step 104.For example, setting above-mentioned tracking Information includes: the identification information and location information of tracked human body.When step 104 detects tracked human body from image A1, The image section for the tracked human body that then this is tracked the identification information of human body namely image A1 is included, this is tracked human body Location information indicate that this is tracked human body coordinate position locating in image A1, update quilt in above-mentioned tracking list at this time Track the tracking information of human body, the tracked human body that the identification information of updated tracked human body includes by image A1 Image section, the location information of updated tracked human body are to be tracked human body coordinate position locating in image A1.Tool Body, can by detected in image A1 this be tracked the left upper apex and bottom right vertex of the human body rectangle frame where human body Coordinate is considered as this and is tracked human body coordinate position locating in image A1, alternatively, can also will detect in image A1 The bottom left vertex of human body rectangle frame and the coordinate of right vertices where the tracked human body are considered as this and are tracked human body in image Locating coordinate position, is not construed as limiting herein in A1.
Further, if the accumulative loss number of above-mentioned tracked human body is not zero, show the tracked people before this with losing Body is traced into again, at this time resets the accumulative loss number of above-mentioned tracked human body.Specifically, if by losing list note The accumulative loss number for being tracked human body is recorded, then the above-mentioned accumulative loss number by above-mentioned tracked human body resets specific manifestation Are as follows: initialize above-mentioned loss list.
Step 106, when can't detect human body or when the secondary human body detected is not above-mentioned tracked human body, detect above-mentioned Whether the accumulative loss number for being tracked human body reaches preset loss frequency threshold value;
In the embodiment of the present invention, when tracked human body does not appear in above-mentioned camera acquired image, will occur Not the case where human body that can't detect human body or detect is not above-mentioned tracked human body.At this point, further detecting this is tracked people Whether the accumulative loss number of body reaches preset loss frequency threshold value, if detecting that this is tracked the accumulative loss of human body time Number reaches preset loss frequency threshold value, then returns to step 101, to redefine tracked human body, and based on new true The step of fixed tracked human body executes step 101 later.If detecting that this is tracked the accumulative loss number of human body and does not reach To above-mentioned loss frequency threshold value, then 107 are entered step.
It further, is not above-mentioned tracked human body when can't detect human body or work as the secondary human body detected, and this is tracked The accumulative loss number of human body is 0, then shows that this is tracked the case where human body is first appeared with losing, can arrange at this time in above-mentioned loss Recorded in table above-mentioned tracked human body image (such as will record in current tracking list this be tracked image deposit of human body In above-mentioned loss list), and empty in above-mentioned tracking list and be tracked the relevant record of human body to this.That is, emptying above-mentioned tracking The tracking information of the tracking human body recorded in list.
The accumulative loss number of above-mentioned tracked human body is added one, and return step 104 by step 107;
In the embodiment of the present invention, when step 106 detects that this is tracked the accumulative loss number of human body and is not up to above-mentioned lose When losing frequency threshold value, holding continues to track to the tracking human body, which is added one.Into one Step, if this is tracked the accumulative loss number of human body, step 107 specific manifestation are as follows: by the loss by losing list records The accumulative loss number for being tracked human body recorded in list adds one.Further, if also record has the quilt in the loss list The image for tracking human body, then can keep the image in the loss list constant.
It should be noted that the human body tracing method in the embodiment of the present invention can be executed by human body tracking device.It is above-mentioned Human body tracking device can integrate in robot, monitor terminal or other terminals, be not construed as limiting herein.
Therefore the embodiment of the present invention initializes tracked the accumulative of human body and loses after determining tracked human body It loses number and human testing is persistently simultaneously carried out based on camera acquired image, as the people that can't detect human body or secondary ought detect When body is not above-mentioned tracked human body (be tracked human body with losing when), detect this by the accumulative loss number with body of loseing face whether Reach preset loss frequency threshold value, if not up to, which being added and continues holding pair together The tracking for being tracked human body re-starts the determination of tracked human body if reached.On the one hand, it is tracked by introducing The accumulative loss number of human body can be realized with losing tracked human body and continue to track to by the human body with losing; On the other hand, frequency threshold value is lost by setting, can also be adopted to avoid because tracked human body does not appear in the camera for a long time Lead to the problem of can not carrying out human body tracking for a long time in the image of collection, improves the flexibility of human body tracking.
Embodiment two
Present example records the tracking information for being tracked human body and tracked people to track list and lose list respectively The accumulative loss number of body, as shown in Fig. 2, the human body tracing method in the embodiment of the present invention includes:
Step 201 determines tracked human body;
Specifically, step 201 is referred to the description of step 101 in Fig. 1-a, details are not described herein again.
List is lost in step 202, initialization;
Wherein, list is empty for the loss after initialization.
Step 203, the tracking information that above-mentioned tracked human body is recorded in tracking list;
In step 203, the tracking information for the tracked human body that recording step 201 determines in tracking list.The tracking Information includes but is not limited to: being tracked the identification information of human body, is tracked the location information of human body.
Specifically, step 203 is referred to the description of step 103 in Fig. 1-a, details are not described herein again.
Step 204 persistently carries out human testing based on camera acquired image;
In step 204, human testing is persistently carried out based on camera acquired image, which can be terminal The camera configured in (such as monitor terminal, robot etc.), alternatively, being also possible to can be with other camera shooting of the terminal called Head instrument, is not construed as limiting herein.
Optionally, in step 204, it is based on camera acquired image, people is persistently carried out using Faster R-CNN Physical examination is surveyed, using the ability of the powerful extraction image further feature of convolutional neural networks, realize more accurate human testing and Tracking.Certainly, in the embodiment of the present invention, the camera can also be collected using other human testing algorithms based on image Image carry out human testing, herein without limitation.
When detecting human body, 205 being entered step, when can't detect human body, entering step 208.
Step 205, judgement when the human body that time detect whether be above-mentioned tracking list records tracked human body;
In the embodiment of the present invention, judge the human body whether be above-mentioned tracking list records tracked human body, if so, into Enter step 207, if it is not, then entering step 206.
Specifically, step 205 can be by as follows when above-mentioned tracking information includes: the image of above-mentioned tracked human body Mode judge when the human body that time detect whether be above-mentioned tracking list records tracked human body: based on above-mentioned tracking information packet The image of the above-mentioned tracked human body contained uses preset human bioequivalence model to judge when whether the human body that time detect is to be above-mentioned Track the tracked human body of list records.Wherein, above-mentioned human bioequivalence model is obtained based on convolutional neural networks training, above-mentioned people Body identification model is for judging whether two human bodies belong to the same human body.Pass through a large amount of training data and is based on convolutional Neural Network training human body identification model enables to (judge two human bodies using human body identification model progress human bioequivalence Whether belong to is the same human body) identification accuracy be better than traditional recognition methods (such as the method based on HOG feature and Method etc. based on contour feature).
Further, in the case where above-mentioned tracking information further includes the location information for being tracked human body, step 205 can be with Include: when detecting human body, according to recorded in above-mentioned tracking list this be tracked human body location information, calculate this by with Track human body and the position coverage rate when the secondary all human bodies detected;Position coverage rate is not less than preset coverage rate threshold if it exists The human body of value, then it is the position coverage rate that human body is tracked with this is maximum when the secondary human body detected is determined as the tracking list The tracked human body of record;Position coverage rate is not less than the human body of above-mentioned coverage rate threshold value if it does not exist, then is based on above-mentioned tracking The image for the above-mentioned tracked human body that information includes, use above-mentioned human bioequivalence model judge when the human body that time detect whether as The tracked human body of above-mentioned tracking list records.Wherein, above-mentioned position coverage rate namely the human body shared region in different images The overlapping area of (such as the human body rectangle frame of human body where in different images) and the ratio of non-overlap area.Due to step 204 be persistently to carry out human testing based on camera acquired image, therefore, is tracked the frame figure that human body is arrived in continuous acquisition Position as in should vary less, and answer correspondingly, being tracked human body overlapping area of region in two continuous frames image This is larger, correspondingly, being tracked position coverage rate of the human body in two continuous frames image also can be larger.Therefore in above scheme, Before carrying out identification judgement using above-mentioned human bioequivalence model, first pass through calculate this be tracked human body with it is all when what is time detected The position coverage rate of human body will be tracked people with this when there are the human body that position coverage rate is not less than above-mentioned coverage rate threshold value The maximum tracked human body for being determined as the tracking list records when the secondary human body detected of the position coverage rate of body, so as to It saves and the time that identification judgement needs to expend is carried out using above-mentioned human bioequivalence model, improve human body tracking efficiency.Specifically, closing It is referred to the description in Fig. 1-a step 104 in the explanation of above-mentioned position coverage rate, details are not described herein again.
Step 206, judgement when the human body that time detect whether be above-mentioned loss list records tracked human body;
In the embodiment of the present invention, when step 205 judges the human body not and be the tracked human body of above-mentioned tracking list records When, further judgement works as whether the secondary human body detected is the tracked human body of above-mentioned loss list records, if so, entering step Rapid 207, if it is not, then entering step 208.
Further, in the embodiment of the present invention in tracked human body for the first time with losing when, i.e., can't detect this for the first time in step 204 When being tracked human body, the image that above-mentioned tracked human body can be recorded in above-mentioned loss list (such as will currently track list The image for being tracked human body of middle record is stored in above-mentioned loss list), and empty in above-mentioned tracking list and be tracked with this The relevant record of human body.That is, emptying the tracking information of the tracking human body recorded in above-mentioned tracking list.Specifically, when above-mentioned When tracking information includes the image of above-mentioned tracked human body, step 206 can also be based on the above-mentioned quilt that above-mentioned tracking information includes The image for tracking human body uses above-mentioned human bioequivalence model to judge when whether the secondary human body detected is as above-mentioned loss list records Tracked human body.
Step 207 will be updated when the tracking information of the secondary human body detected into above-mentioned tracking list, and in above-mentioned loss The accumulative loss number of the tracked human body recorded in list initializes the loss list when being not zero;
In the embodiment of the present invention, when detect human body and when the secondary human body detected be above-mentioned tracked human body when, will work as The tracking information of the secondary human body detected is updated into above-mentioned tracking list, later return step 204.For example, setting above-mentioned Tracking information includes: the identification information and location information of tracked human body.When step 204 detects tracked people from image A1 Body, then this is tracked the identification information of human body namely image A1 is included that this is tracked image section (such as the image of human body In A1 be tracked human body where human body rectangle frame in image section), this be tracked human body location information indicate this by with Track human body coordinate position locating in image A1, updates the tracking information for being tracked human body in above-mentioned tracking list at this time, Image section of the identification information of updated tracked human body by the image A1 tracked human body for including, updated quilt The location information of tracking human body is to be tracked human body coordinate position locating in image A1.Specifically, can will be in image A1 The left upper apex of human body rectangle frame and the coordinate of bottom right vertex where the tracked human body detected are considered as this and are tracked people Body coordinate position locating in image A1, alternatively, this that can also will be detected in image A1 is tracked where human body The bottom left vertex of human body rectangle frame and the coordinate of right vertices are considered as this and are tracked human body coordinate position locating in image A1, It is not construed as limiting herein.
Further, if the accumulative loss number of above-mentioned tracked human body is not zero, show the tracked people before this with losing Body is traced into again, initializes above-mentioned loss list at this time.
Whether step 208, the accumulative loss number for detecting above-mentioned tracked human body reach preset loss frequency threshold value;
In the embodiment of the present invention, when tracked human body does not appear in above-mentioned camera acquired image, will occur Not the case where human body that can't detect human body or detect is not above-mentioned tracked human body, at this point, further detecting above-mentioned loss column Whether the accumulative loss number for being tracked human body recorded in table reaches preset loss frequency threshold value, if detecting the quilt Tracking human body accumulative loss number reach preset loss frequency threshold value, then return to step 201, so as to redefine by Track human body, and the step of step 201 is executed later based on the tracked human body newly determined.If detecting that this is tracked human body Accumulative loss number be not up to above-mentioned loss frequency threshold value, then enter step 209.
Further, when the accumulative loss number 0 of above-mentioned tracked human body, then show that this is tracked human body and first appears with losing The case where, the image that can record above-mentioned tracked human body in above-mentioned loss list at this time (such as will be tracked currently in list and be remembered The image for being tracked human body of record is stored in above-mentioned loss list), and empty in above-mentioned tracking list and be tracked human body with this Relevant record.That is, emptying the tracking information of the tracking human body recorded in above-mentioned tracking list.
The accumulative loss number of above-mentioned tracked human body is added one, and return step 204 by step 209;
In the embodiment of the present invention, when step 208 detects that this is tracked the accumulative loss number of human body and is not up to above-mentioned lose When losing frequency threshold value, holding continues to track to the tracking human body, and this recorded in above-mentioned loss list is tracked human body Accumulative loss number add one.Further, it if also record has this to be tracked the image of human body in above-mentioned loss list, can keep Image in the loss list is constant.
It should be noted that the human body tracing method in the embodiment of the present invention can be executed by human body tracking device, it is above-mentioned Human body tracking device can integrate in robot, monitor terminal or other terminals, be not construed as limiting herein.
Therefore the embodiment of the present invention initializes tracked the accumulative of human body and loses after determining tracked human body It loses number and human testing is persistently simultaneously carried out based on camera acquired image, as the people that can't detect human body or secondary ought detect When body is not above-mentioned tracked human body (be tracked human body with losing when), detect this by the accumulative loss number with body of loseing face whether Reach preset loss frequency threshold value, if not up to, which being added and continues holding pair together The tracking for being tracked human body re-starts the determination of tracked human body if reached.On the one hand, it is tracked by introducing The accumulative loss number of human body can be realized with losing tracked human body and continue to track to by the human body with losing; On the other hand, frequency threshold value is lost by setting, can also be adopted to avoid because tracked human body does not appear in the camera for a long time Lead to the problem of can not carrying out human body tracking for a long time in the image of collection, improves the flexibility of human body tracking.
Embodiment three
The embodiment of the present invention provides a kind of human body tracking device, as shown in figure 3, the human body tracking in the embodiment of the present invention fills Setting 300 includes:
Determination unit 301, for determining tracked human body;
Initialization unit 302, for the accumulative loss number of above-mentioned tracked human body to be initialized as zero;
Updating unit 303 is tracked, for recording the tracking information of above-mentioned tracked human body in tracking list;
Human testing unit 304, for persistently carrying out human testing based on camera acquired image;
Number detection unit 305 is lost, for can't detect human body or human testing unit when human testing unit 304 304 when the secondary human body detected is not above-mentioned tracked human body, and whether the accumulative loss number of the above-mentioned tracked human body of detection Reach preset loss frequency threshold value;
Accumulated unit 306 is lost, for detecting that above-mentioned the accumulative of tracked human body is lost when loss number detection unit 305 When mistake number is not up to above-mentioned loss frequency threshold value, the accumulative loss number of above-mentioned tracked human body is added one;
Determination unit 301 detects that the accumulative loss number of above-mentioned tracked human body reaches in loss number detection unit 305 It is triggered again when to above-mentioned loss frequency threshold value;
Tracking updating unit 303 is also used to: when human testing unit 304 detects human body and human testing unit 304 is worked as When the secondary human body detected is above-mentioned tracked human body, human testing unit 304 is worked as to the tracking information of the secondary human body detected It updates in above-mentioned tracking list;
Initialization unit 302 is also used to: when human testing unit 304 detects human body and human testing unit 304 is when secondary The human body detected is above-mentioned tracked human body, and when the accumulative loss number of above-mentioned tracked human body is not zero, by above-mentioned quilt The accumulative loss number for tracking human body is reset.
Optionally, determination unit 301 includes:
Object detection unit is tracked, for persistently carrying out human testing until detecting human body;
Sub- determination unit, when the number of the human body for detecting when above-mentioned tracking object detection unit is 1, by it is above-mentioned with Track object detection unit is determined as being tracked human body when the secondary human body detected;It is detected when above-mentioned tracking object detection unit When the number of human body is not 1, by above-mentioned tracking object detection unit when a human body in the secondary multiple human bodies detected determines To be tracked human body.
Optionally, above-mentioned sub- determination unit is specifically used for: as the human body that above-mentioned tracking object detection unit detects When number is not 1, the area that above-mentioned tracking object detection unit works as the secondary each human body detected is calculated separately, area is maximum Human body is determined as being tracked human body.
Optionally, pass through the accumulative loss number of the loss above-mentioned tracked human body of list records;Initialization unit 302 is specific For the accumulative loss number of above-mentioned tracked human body to be reset in a manner of initializing above-mentioned loss list;Wherein, it initializes List is empty for loss afterwards.
Optionally, the human body tracking device in the embodiment of the present invention further include: recording unit is lost, for working as human testing The human body that unit 304 can't detect human body or human testing unit 304 and secondary ought detect is not above-mentioned tracked human body, and above-mentioned When the accumulative loss number for being tracked human body is 0, the image of above-mentioned tracked human body is recorded in above-mentioned loss list;Initialization Unit 302 is also used to: when human testing unit 304 can't detect human body or human testing unit 304 when the secondary human body detected Not be above-mentioned tracked human body, and the accumulative loss number of above-mentioned tracked human body be 0 when, empty in above-mentioned tracking list with it is upper State the relevant record of tracked human body.
Optionally, human testing unit 304 is specifically used for: when detecting human body, judging human testing unit 304 when secondary The human body detected whether be above-mentioned tracking list or above-mentioned loss list records tracked human body;If human testing unit 304 when the tracked human body that the secondary human body detected is above-mentioned tracking list or above-mentioned loss list records, then determine that human body is examined Unit 304 is surveyed when the secondary human body detected is above-mentioned tracked human body;If human testing unit 304 works as the secondary human body detected not For above-mentioned tracking list records tracked human body and not be above-mentioned loss list records tracked human body, then determine human body examine Unit 304 is surveyed when the secondary human body detected is not above-mentioned tracked human body.
Optionally, human testing unit 304 is specifically used for: when detecting human body, judging human testing unit 304 when secondary The human body detected whether be above-mentioned tracking list records tracked human body;What if human testing unit 304 secondary ought detect Human body is not the tracked human body of above-mentioned tracking list records, then judges that human testing unit 304 is when the secondary human body detected The no tracked human body for above-mentioned loss list records.
Optionally, above-mentioned tracking information includes: the location information of above-mentioned tracked human body and the figure of above-mentioned tracked human body Picture;Human testing unit 304 is specifically used for: above-mentioned tracked according to what is recorded in above-mentioned tracking list when detecting human body The location information of human body calculates above-mentioned tracked human body and human testing unit 304 when the position of the secondary all human bodies detected Coverage rate;Position coverage rate is not less than the human body of preset coverage rate threshold value if it exists, then by the position with above-mentioned tracked human body Set the maximum tracked human body for being determined as above-mentioned tracking list records when the secondary human body detected of coverage rate;Position if it does not exist Coverage rate is not less than the human body of preset coverage rate threshold value, the then figure for the tracked human body for including based on above-mentioned tracking information Picture uses preset human bioequivalence model to judge human testing unit 304 when whether the secondary human body detected is for above-mentioned tracking column The tracked human body of table record, wherein above-mentioned human bioequivalence model is based on convolutional neural networks training and obtains.
Optionally, human testing unit 304 is specifically used for: camera acquired image is based on, using Faster R- CNN persistently carries out human testing.
It should be noted that the human body tracking device in the embodiment of the present invention can integrate robot, monitor terminal or In other terminals.The function of each functional module of human body tracking device is referred to the description in above method embodiment, Its specific implementation process can refer to the associated description in above method embodiment, and details are not described herein again.
Therefore the human body tracking device in the embodiment of the present invention initializes the quilt after determining tracked human body The accumulative loss number for tracking human body simultaneously persistently carries out human testing based on camera acquired image, when can't detect human body Or this is detected by with body of loseing face (when being tracked human body with losing) when the secondary human body detected is not the tracked human body It is accumulative to lose whether number reaches preset loss frequency threshold value, if not up to, this to be tracked to the loss number of human body Add and continues to keep the tracking for being tracked human body to this, if reached, re-start the determination of tracked human body together.One side Face is tracked the accumulative loss number of human body by introducing, and can realize with losing tracked human body to by with losing Human body continue to track;On the other hand, frequency threshold value is lost by setting, it can also be to avoid because being tracked human body for a long time not It appears in the image of camera acquisition and leads to the problem of can not carrying out human body tracking for a long time, improve human body tracking Flexibility.
In several embodiments provided herein, it should be understood that disclosed device and method can pass through it Its mode is realized.
It should be noted that for the various method embodiments described above, describing for simplicity, therefore, it is stated as a series of Combination of actions, but those skilled in the art should understand that, the present invention is not limited by the sequence of acts described because According to the present invention, certain steps can use other sequences or carry out simultaneously.Secondly, those skilled in the art should also know It knows, the embodiments described in the specification are all preferred embodiments, and related actions and modules might not all be this hair Necessary to bright.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, there is no the portion being described in detail in some embodiment Point, it may refer to the associated description of other embodiments.
The above are to a kind of description of human body tracing method and human body tracking device provided by the present invention, for this field Those skilled in the art, thought according to an embodiment of the present invention has change in specific embodiments and applications Place, to sum up, the contents of this specification are not to be construed as limiting the invention.

Claims (18)

1. a kind of human body tracing method characterized by comprising
Determine tracked human body;
The accumulative loss number of the tracked human body is initialized as zero;
The tracking information of the tracked human body is recorded in tracking list;Wherein, the tracking information includes described is tracked The location information of the identification information of human body and the tracked human body;
Human testing is persistently carried out based on camera acquired image;
Judgement works as whether the secondary human body detected is the tracked human body for tracking list records, wherein judgement is detected when secondary To human body whether be it is described tracking list records tracked human body mode include: according to be tracked human body with when time detection The position coverage rate of all human bodies arrived, alternatively, whether working as the secondary human body detected using the judgement of preset human bioequivalence model For the tracked human body of the tracking list records;
It, will be when the tracking for the human body that time detect when detecting human body and when the secondary human body detected is the tracked human body Information update is tracked into the tracking list, and when the accumulative loss number of the tracked human body is not zero by described The accumulative loss number of human body is reset;
When can't detect human body or when the secondary human body detected is not the tracked human body, the tracked human body is detected It is accumulative to lose whether number reaches preset loss frequency threshold value;
If the accumulative loss number of the tracked human body reaches the loss frequency threshold value, return execute the determination by with The step of track human body and subsequent step;
If the accumulative loss number of the tracked human body is not up to the loss frequency threshold value, by the tracked human body Accumulative number of losing adds one, and return execution described the step of human testing is persistently carried out based on camera acquired image with And subsequent step.
2. the method according to claim 1, wherein the tracked human body of the determination includes:
Persistently human testing is carried out until detecting human body;
If the number of the human body detected is 1, will be determined as being tracked human body when the secondary human body detected;
If the number of the human body detected is not 1, by when a human body in the multiple human bodies that time detect be determined as by with Track human body.
3. according to the method described in claim 2, it is characterized in that, described will be as a people in the secondary multiple human bodies detected Body is determined as being tracked human body
Calculate separately the area when the secondary each human body detected;
The maximum human body of area is determined as to be tracked human body.
4. the method according to claim 1, wherein being tracked the accumulative of human body described in list records by losing Lose number;
It is described that the accumulative loss number of the tracked human body is initialized as zero, specifically: initialize the loss list;
It is described when detecting human body and when the secondary human body detected is the tracked human body, will be when the human body that time detect Tracking information is updated into the tracking list, and when the accumulative loss number of the tracked human body is not zero by the quilt The accumulative loss number for tracking human body is reset;Specifically: it is tracked when detecting human body and working as the secondary human body detected to be described When human body, it will update when the tracking information of the secondary human body detected into the tracking list, and in the tracked human body Accumulative lose when number is not zero initializes the loss list;
Wherein, list is empty for the loss after initialization.
5. according to the method described in claim 4, it is characterized in that, not being when can't detect human body or work as the secondary human body detected The tracked human body, it is described the accumulative loss number of the tracked human body is added one before further include:
If the accumulative loss number of the tracked human body is 0, the tracked human body is recorded in the loss list Image, and empty record relevant to the tracked human body in the tracking list.
6. according to the method described in claim 5, it is characterized in that, described persistently carry out people based on camera acquired image Physical examination is surveyed, comprising:
When detecting human body, whether judgement is the tracking list or the loss list records when the secondary human body detected It is tracked human body;
If determining to work as secondary when the secondary human body detected is the tracking list or the tracked human body for losing list records The human body detected is the tracked human body;
If when the secondary human body detected is not the tracked human body of the tracking list records and is not the loss list records Tracked human body, then determine that when the human body that time detect be not the tracked human body.
7. according to the method described in claim 6, it is characterized in that, the judgement when the human body that time detect whether be it is described with Track list or the tracked human body for losing list records, comprising:
Whether judgement is the tracked human body for tracking list records when the secondary human body detected;
If working as the tracked human body that the secondary human body detected is not the tracking list records, judgement is when the secondary human body detected It whether is the tracked human body for losing list records.
8. the method according to the description of claim 7 is characterized in that the tracking information includes: the position of the tracked human body The image of confidence breath and the tracked human body;
Whether the judgement is the tracked human body for tracking list records when the secondary human body detected, comprising:
According to the location information of the tracked human body recorded in the tracking list, the tracked human body is calculated and when secondary The position coverage rate of all human bodies detected;
Position coverage rate is not less than the human body of preset coverage rate threshold value if it exists, then will cover with the position of the tracked human body The maximum tracked human body for being determined as the tracking list records when the secondary human body detected of lid rate;
Position coverage rate is not less than the human body of preset coverage rate threshold value, the then institute for including based on the tracking information if it does not exist The image for stating tracked human body uses preset human bioequivalence model to judge when whether the secondary human body detected is as tracking column The tracked human body of table record, wherein the human bioequivalence model is based on convolutional neural networks training and obtains.
9. method according to any one of claims 1 to 8, which is characterized in that described to be based on camera acquired image Human testing is persistently carried out, specifically:
Based on camera acquired image, human testing is persistently carried out using more rapidly region convolutional neural networks.
10. a kind of human body tracking device characterized by comprising
Determination unit, for determining tracked human body;
Initialization unit, for the accumulative loss number of the tracked human body to be initialized as zero;
Updating unit is tracked, for recording the tracking information of the tracked human body in tracking list;Wherein, the tracking letter Breath includes the identification information of the tracked human body and the location information of the tracked human body;
Human testing unit, for persistently carrying out human testing based on camera acquired image;
Judging unit, for judging whether when the secondary human body detected be the tracked human body for tracking list records, wherein Judgement includes: according to being tracked people when whether the human body time detected is the modes of tracked human body of the tracking list records Body and the position coverage rate when the secondary all human bodies detected, alternatively, using the judgement of preset human bioequivalence model when time detection To human body whether be it is described tracking list records tracked human body;
Number detection unit is lost, for can't detect human body or the human testing unit when secondary when the human testing unit When the human body detected is not the tracked human body, whether the accumulative loss number for detecting the tracked human body reaches default Loss frequency threshold value;
Accumulated unit is lost, for detecting the accumulative loss number of the tracked human body when the loss number detection unit Not up to the loss frequency threshold value when, the accumulative loss number of the tracked human body is added one;
The determination unit detects that the accumulative loss number of the tracked human body reaches in the loss number detection unit It is triggered again when the loss frequency threshold value;
The tracking updating unit is also used to: when the human testing unit detects human body and the human testing unit is when secondary When the human body detected is the tracked human body, the human testing unit is worked as into the tracking information of the secondary human body detected more Newly into the tracking list;
The initialization unit is also used to: when the human testing unit detects human body and the human testing unit is when time inspection The human body measured is the tracked human body, and when the accumulative loss number of the tracked human body is not zero, by it is described by with The accumulative loss number of track human body is reset.
11. human body tracking device according to claim 10, which is characterized in that the determination unit includes:
Object detection unit is tracked, for persistently carrying out human testing until detecting human body;
When the number of sub- determination unit, the human body for detecting when the tracking object detection unit is 1, by the tracking mesh Mark detection unit is determined as being tracked human body when the secondary human body detected;When the human body that the tracking object detection unit detects Number when not being 1, by the tracking object detection unit when a human body in the multiple human bodies that time detect be determined as by Track human body.
12. human body tracking device according to claim 11, which is characterized in that
The sub- determination unit is specifically used for: when the number for the human body that the tracking object detection unit detects is not 1, point The tracking object detection unit is not calculated when the area for each human body that time detect, by the maximum human body of area be determined as by Track human body.
13. human body tracking device according to claim 10, which is characterized in that be tracked described in list records by losing The accumulative loss number of human body;
The initialization unit is specifically used for losing the accumulative of tracked human body in a manner of initializing the loss list Number is lost to reset;
Wherein, list is empty for the loss after initialization.
14. human body tracking device according to claim 13, which is characterized in that the human body tracking device further include:
Recording unit is lost, for can't detect human body or the human testing unit when time detection when the human testing unit To human body be not the tracked human body, and when the accumulative loss number of the tracked human body is 0, in the loss list The middle image for recording the tracked human body;
The initialization unit is also used to: when the human testing unit can't detect human body or the human testing unit when secondary The human body detected is not the tracked human body, and when the accumulative loss number of the tracked human body is 0, empty it is described with Record relevant to the tracked human body in track list.
15. human body tracking device according to claim 14, which is characterized in that the human testing unit is specifically used for: When detecting human body, judge the human testing unit when whether the human body that time detect is the tracking list or described is lost Lose the tracked human body of list records;If the human testing unit is the tracking list or described when the secondary human body detected The tracked human body for losing list records, then determining that the human testing unit works as the secondary human body detected is the tracked people Body;If the human testing unit is when the secondary human body detected is not the tracked human body of the tracking list records and is not institute State the tracked human body for losing list records, then determine the human testing unit when the human body that time detect be not it is described by with Track human body.
16. human body tracking device according to claim 15, which is characterized in that the human testing unit is specifically used for: When detecting human body, judge that the human testing unit works as whether the secondary human body detected is the quilt for tracking list records Track human body;If the human testing unit is not the tracked human body for tracking list records when the secondary human body detected, Then judge that the human testing unit works as whether the secondary human body detected is the tracked human body for losing list records.
17. human body tracking device according to claim 16, which is characterized in that the tracking information include: it is described by with The image of the location information of track human body and the tracked human body;
The human testing unit is specifically used for: when detecting human body, according to it is described tracking list in record it is described by with The location information of track human body calculates the tracked human body and the human testing unit when the position of the secondary all human bodies detected Set coverage rate;If it exists position coverage rate be not less than preset coverage rate threshold value human body, then by with the tracked human body The maximum tracked human body for being determined as the tracking list records when the secondary human body detected of position coverage rate;Position if it does not exist Set the human body that coverage rate is not less than preset coverage rate threshold value, the then tracked human body for including based on the tracking information Image uses preset human bioequivalence model to judge the human testing unit when whether the secondary human body detected is for the tracking The tracked human body of list records, wherein the human bioequivalence model is based on convolutional neural networks training and obtains.
18. human body tracking device according to any one of claims 10 to 17, which is characterized in that the human testing unit It is specifically used for: based on camera acquired image, persistently carries out human testing using more rapidly region convolutional neural networks.
CN201610752578.1A 2016-08-29 2016-08-29 Human body tracing method and human body tracking device Active CN106355603B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610752578.1A CN106355603B (en) 2016-08-29 2016-08-29 Human body tracing method and human body tracking device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610752578.1A CN106355603B (en) 2016-08-29 2016-08-29 Human body tracing method and human body tracking device

Publications (2)

Publication Number Publication Date
CN106355603A CN106355603A (en) 2017-01-25
CN106355603B true CN106355603B (en) 2019-10-22

Family

ID=57854633

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610752578.1A Active CN106355603B (en) 2016-08-29 2016-08-29 Human body tracing method and human body tracking device

Country Status (1)

Country Link
CN (1) CN106355603B (en)

Families Citing this family (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107273836A (en) * 2017-06-07 2017-10-20 深圳市深网视界科技有限公司 A kind of pedestrian detection recognition methods, device, model and medium
CN110363787A (en) * 2018-03-26 2019-10-22 北京市商汤科技开发有限公司 Information acquisition method and system, electronic equipment, program and medium
CN108762309B (en) * 2018-05-03 2021-05-18 浙江工业大学 Human body target following method based on hypothesis Kalman filtering
CN108921008A (en) * 2018-05-14 2018-11-30 深圳市商汤科技有限公司 Portrait identification method, device and electronic equipment
CN110675426B (en) 2018-07-02 2022-11-22 百度在线网络技术(北京)有限公司 Human body tracking method, device, equipment and storage medium
CN109145564A (en) * 2018-08-22 2019-01-04 中国工商银行股份有限公司 Control the method and device of mobile terminal
CN109271942A (en) * 2018-09-26 2019-01-25 上海七牛信息技术有限公司 A kind of stream of people's statistical method and system
JP7352235B2 (en) * 2019-01-12 2023-09-28 株式会社Zmp Autonomous vehicles, delivery systems, programs, and storage media that store programs
CN110176025B (en) * 2019-04-25 2021-06-18 沈阳图为科技有限公司 Invigilator tracking method based on posture
CN110264498A (en) * 2019-06-26 2019-09-20 北京深醒科技有限公司 A kind of human body tracing method under video monitoring scene
CN110378931A (en) * 2019-07-10 2019-10-25 成都数之联科技有限公司 A kind of pedestrian target motion track acquisition methods and system based on multi-cam
CN111241995B (en) * 2020-01-09 2022-12-30 苏宁云计算有限公司 Loss compensation method and system for tracked object and computer equipment
CN111539283B (en) * 2020-04-15 2023-08-11 上海摩象网络科技有限公司 Face tracking method and face tracking equipment
CN112200841B (en) * 2020-09-30 2021-08-27 杭州海宴科技有限公司 Cross-domain multi-camera tracking method and device based on pedestrian posture
CN112287846A (en) * 2020-10-30 2021-01-29 深圳市优必选科技股份有限公司 Target person following method, device, mobile robot and readable storage medium
CN112894810A (en) * 2021-01-19 2021-06-04 四川阿泰因机器人智能装备有限公司 KCF algorithm-based mobile robot target loss prevention following method

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101339608A (en) * 2008-08-15 2009-01-07 北京中星微电子有限公司 Object tracking method and system based on detection
CN101783015A (en) * 2009-01-19 2010-07-21 北京中星微电子有限公司 Equipment and method for tracking video
CN102831378A (en) * 2011-06-14 2012-12-19 株式会社理光 Method and system for detecting and tracking human object
CN103064086A (en) * 2012-11-04 2013-04-24 北京工业大学 Vehicle tracking method based on depth information
CN103838258A (en) * 2014-02-26 2014-06-04 上海微小卫星工程中心 Automatic tracking method and system applied to space-based space target
CN103927508A (en) * 2013-01-11 2014-07-16 浙江大华技术股份有限公司 Target vehicle tracking method and device
JP2014149716A (en) * 2013-02-01 2014-08-21 Panasonic Corp Object tracking apparatus and method thereof
CN105261035A (en) * 2015-09-15 2016-01-20 杭州中威电子股份有限公司 Method and device for tracking moving objects on highway
CN105825524A (en) * 2016-03-10 2016-08-03 浙江生辉照明有限公司 Target tracking method and apparatus

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101339608A (en) * 2008-08-15 2009-01-07 北京中星微电子有限公司 Object tracking method and system based on detection
CN101783015A (en) * 2009-01-19 2010-07-21 北京中星微电子有限公司 Equipment and method for tracking video
CN102831378A (en) * 2011-06-14 2012-12-19 株式会社理光 Method and system for detecting and tracking human object
CN103064086A (en) * 2012-11-04 2013-04-24 北京工业大学 Vehicle tracking method based on depth information
CN103927508A (en) * 2013-01-11 2014-07-16 浙江大华技术股份有限公司 Target vehicle tracking method and device
JP2014149716A (en) * 2013-02-01 2014-08-21 Panasonic Corp Object tracking apparatus and method thereof
CN103838258A (en) * 2014-02-26 2014-06-04 上海微小卫星工程中心 Automatic tracking method and system applied to space-based space target
CN105261035A (en) * 2015-09-15 2016-01-20 杭州中威电子股份有限公司 Method and device for tracking moving objects on highway
CN105825524A (en) * 2016-03-10 2016-08-03 浙江生辉照明有限公司 Target tracking method and apparatus

Also Published As

Publication number Publication date
CN106355603A (en) 2017-01-25

Similar Documents

Publication Publication Date Title
CN106355603B (en) Human body tracing method and human body tracking device
CN109886130B (en) Target object determination method and device, storage medium and processor
CN109922310B (en) Target object monitoring method, device and system
CN105872477B (en) video monitoring method and video monitoring system
CN109522854A (en) A kind of pedestrian traffic statistical method based on deep learning and multiple target tracking
JP5422018B2 (en) Image processing method and image processing apparatus
CN106447680B (en) The object detecting and tracking method that radar is merged with vision under dynamic background environment
CN108200334B (en) Image shooting method and device, storage medium and electronic equipment
CN110264493A (en) A kind of multiple target object tracking method and device under motion state
CN107452015B (en) Target tracking system with re-detection mechanism
CN105165004B (en) Camera chain
JP6624877B2 (en) Information processing apparatus, information processing method and program
CN110910460B (en) Method and device for acquiring position information and calibration equipment
JP2009510541A (en) Object tracking method and object tracking apparatus
CN106037651A (en) Heart rate detection method and system
EP3593322B1 (en) Method of detecting moving objects from a temporal sequence of images
CN109255360A (en) A kind of objective classification method, apparatus and system
CN110245609A (en) Pedestrian track generation method, device and readable storage medium storing program for executing
CN109886129A (en) Prompt information generation method and device, storage medium and electronic device
JP2004338889A (en) Image recognition device
CN111047622A (en) Method and device for matching objects in video, storage medium and electronic device
CN110651274A (en) Movable platform control method and device and movable platform
CN110580708B (en) Rapid movement detection method and device and electronic equipment
CN104063681B (en) A kind of moving object image-recognizing method and device
CN110321782A (en) A kind of system detecting characteristics of human body's signal

Legal Events

Date Code Title Description
C06 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