CN107862713B - Camera deflection real-time detection early warning method and module for polling meeting place - Google Patents
Camera deflection real-time detection early warning method and module for polling meeting place Download PDFInfo
- Publication number
- CN107862713B CN107862713B CN201710866772.7A CN201710866772A CN107862713B CN 107862713 B CN107862713 B CN 107862713B CN 201710866772 A CN201710866772 A CN 201710866772A CN 107862713 B CN107862713 B CN 107862713B
- Authority
- CN
- China
- Prior art keywords
- image
- camera
- early warning
- submodule
- deflection
- 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
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
- G06T2207/30201—Face
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a camera deflection real-time detection early warning method for a polling meeting place, aiming at the defects of the existing mainstream camera deflection detection method, the real-time matching algorithm is carried out on a video image by adopting a rapid template matching technology, and the camera angle deflection real-time detection algorithm combining geometric relation and camera parameters is utilized, so that the angle deflection of a camera in the video meeting place can be timely detected and early warned, the video quality reduction of a meeting caused by the deflection of the camera is avoided, the camera deflection real-time detection early warning method can be integrated with a common video meeting platform, the camera deflection real-time monitoring early warning function of the polling meeting place is achieved, the abnormal condition in the video meeting process is timely recognized, and the quality and the safety of the video meeting are improved. The invention also provides a camera deflection real-time detection early warning module for the polling meeting place.
Description
Technical Field
The invention relates to the technical field of video processing, in particular to a camera deflection real-time detection early warning method for a polling meeting place and a camera deflection real-time detection early warning module for the polling meeting place.
Background
Video conferencing has rapidly developed and widely used in recent years. However, in some cases, such as manual touch, the angle of the conference monitoring camera may be deflected, so that the monitoring range of the camera is changed. The video picture of the expected monitoring range or the monitoring target can not be displayed on the display, and the video picture is not suitable to be used as the conference video, so that the performance of the conference video is greatly reduced. Therefore, real-time intelligent video monitoring of camera angles is necessary.
At present, two methods are mainly used for detecting the deflection of a mainstream camera, namely whether the deflection occurs or not and the angle of the deflection is detected by depending on hardware facilities such as a cradle head and the like which are matched with the camera; secondly, calibrating the target camera and modeling the calibration result to finish the detection of the deflection of the camera. However, the detection of the deflection of the camera is completed by a hardware method, which is high in cost, and meanwhile, the existing meeting place camera which does not contain matched hardware is more complex to modify. If the camera is calibrated and the deflection of the meeting place camera is detected in a mode of modeling a calibration result, the calculation steps are complex, the calculation amount is large, and the real-time detection of the meeting place camera can hardly be finished.
Disclosure of Invention
In view of the above, one of the objectives of the present invention is to provide a method for detecting and warning the deflection of a camera in a polling meeting place in real time, which can effectively detect whether the camera in the meeting place deflects and can warn in time. The invention also aims to provide a camera deflection real-time detection early warning module for the polling meeting place.
One of the purposes of the invention is realized by the following technical scheme:
the real-time detection and early warning method for the deflection of the camera aiming at the polling meeting place comprises the following steps:
(1) acquiring a video frame image from a conference video which is not deflected, and simply referring the image to be an original image;
(2) intercepting an image block (the size of the image block is selected according to the size of a video image) from the center of the original image, and simply using the image block as a template image;
(3) comparing the difference between the current video frame image and the previous video frame image, setting a difference threshold, intercepting the template again under the condition that the difference is greater than the threshold, and polling to the next meeting place under the condition of polling the meeting places;
(4) performing face detection on the video frame image under the condition that the difference degree is smaller than the threshold value, and adding 1 to an empty field counter if no face is detected;
(5) the counting of the empty field counter reaches an empty field threshold value, the threshold value is set according to the actual situation of the meeting place, and the system gives an early warning;
(6) for the condition that the human face is detected, carrying out multi-step long template matching on the current video frame image and the template image to obtain a matched image;
(7) calculating offset pixels of the template image and the matching image;
(8) calculating the real-time deflection angle of the camera according to the offset pixels calculated in the step (7) by combining the parameters of the camera with the geometric relation;
(9) and if the deflection angle of the camera is larger than the angle threshold value, setting the threshold value according to the actual situation of the meeting place, and performing early warning by the system.
Further, in step (2), the coordinates of the center point pixel coordinate system of the intercepted template image are extracted.
Further, in the step (4), a face detection algorithm based on SVM and HOG features is adopted, firstly, an SVM is used for training collected samples to obtain a face feature library, then, a sliding window is adopted to traverse all regions of the image to search regions where faces may exist, the size of a sub-window is required to be consistent with that of the training samples, HOG features of each sub-window are calculated, an SVM classifier is used for judging feature values, if the face is the face, coordinate values of the face are recorded, otherwise, the window is abandoned, all the coordinate values are finally collected, and face calibration is completed on an original image.
Further, in step (5), the matching image should extract coordinates of the center point pixel coordinate system.
Further, in the step (6), the matching of the multi-step long template refers to: the template image and the search subgraph are respectively in the x direction and the y direction according to the current step SxAnd SyCollecting data, Sx，SySetting according to the size of the actual video image, calculating the correlation coefficient R between the template image and the searched subgraph according to the data, and if the current correlation coefficient R is smaller than the maximum value R of the correlation coefficientmaxThen move the search subgraph to the next oneAnd otherwise, taking half of the current step length in the x direction and the y direction as a new step length, and calculating until a matched image is found.
Further, in the step (7), the offset pixel of the image block is calculated according to the coordinates of the central pixel coordinate system of the two image blocks.
The second purpose of the invention is realized by the following technical scheme:
the camera deflection real-time detection early warning module for the polling meeting place comprises:
the difference detection submodule is used for detecting whether the meeting place scene changes;
the face recognition submodule is used for detecting meeting place information;
the template matching submodule is used for calculating the number of deflection pixels;
the offset angle submodule is used for converting the actual deflection angle into an actual deflection angle;
the early warning information generation submodule is used for generating final early warning information; and
a human-computer interaction interface, parameter setting and a display screen;
the human face recognition system comprises a human face recognition submodule, a template matching submodule, a deviation angle submodule, an early warning generation submodule, a parameter setting submodule and a display, wherein the difference detection submodule is connected with the input end of the human face recognition submodule, the output end of the human face recognition submodule is connected with the input end of the template matching submodule, the output end of the template matching submodule is connected with the input end of the deviation angle submodule, the output end of the deviation angle submodule is connected with the input end of the early warning generation submodule, the output end of the early warning generation submodule is connected with the input end of a human-computer interaction.
The invention has the beneficial effects that:
aiming at the defects of the existing mainstream camera deflection detection method, the invention can realize the timely detection and early warning of the angular deflection of the camera in the video conference meeting place by adopting the real-time matching algorithm of the rapid template matching technology to the video image and the camera angular deflection real-time detection algorithm which combines the geometric relation with the camera parameter, avoid the reduction of the video quality of the conference caused by the deflection of the camera, can be integrated with a common video conference platform, achieve the camera deflection real-time monitoring and early warning function of a polling meeting place, timely identify the abnormal condition in the video conference process and improve the quality and the safety of the video conference.
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof.
Drawings
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described in detail with reference to the accompanying drawings, in which:
FIG. 1 is a flow chart of the algorithm of the present invention.
Fig. 2 is a geometric schematic diagram of a pixel-angle relationship.
FIG. 3 is a block diagram of the modules of the present invention;
Detailed Description
Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. It should be understood that the preferred embodiments are illustrative of the invention only and are not limiting upon the scope of the invention.
As shown in fig. 1, the method for detecting and warning the deflection of the camera in real time for polling meeting places of the present invention includes the following steps:
(1) acquiring a video frame image from a conference video which is not deflected, and simply referring the image to be an original image;
(2) intercepting an image block from the center of the original image, selecting the size of the image block according to the size of the video image, and simply using the image block as a template image; in the embodiment, the coordinates of the central point pixel coordinate system of the intercepted template image are extracted;
(3) carrying out difference contrast on the current video frame image and the previous video frame image, setting a difference threshold value, and intercepting the template again under the condition that the difference is greater than the threshold value; in the implementation of the embodiment, according to the characteristics of the polling meeting places, the polling to the next meeting place is considered under the condition that the difference degree is greater than the threshold value;
(4) performing face detection on the video frame image under the condition that the difference degree is smaller than the threshold value, and adding 1 to an empty field counter if no face is detected;
in the embodiment, according to the characteristics of the polling meeting place, the camera which does not detect the face may deflect to a large angle and does not shoot the participant, or the meeting place is an empty place, and the patent adopts a face detection algorithm based on SVM and HOG characteristics. Firstly, training a collected sample by using an SVM (support vector machine) to obtain a face feature library; traversing each region of the image by adopting a sliding window to search regions in which human faces may exist, wherein the size of the sub-window is required to be consistent with that of the training sample; calculating HOG characteristics of each sub-window, judging the characteristic value through an SVM classifier, recording coordinate values of the sub-window if the characteristic value is a human face, and discarding the window if the characteristic value is not the human face; and finally, summarizing all coordinate values, and finishing the calibration of the face on the original image.
(5) When the count of the empty field counter reaches an empty field threshold (the threshold is set according to the actual situation of the meeting place), the system gives an early warning;
(6) for the condition that the human face is detected, carrying out multi-step long template matching on the current video frame image and the template image to obtain a matched image; in this embodiment, the matching image should extract coordinates of the pixel coordinate system of the center point.
In this step, the similarity detection adopts a multi-step template matching detection method. Template matching is an efficient pattern recognition technique that can more directly reflect the similarity between images using image information and a priori knowledge about the recognition pattern. The method matches the actual image block with the input image by sliding the image block on the input image, and searches for the most similar area with the template image in one image. The template matching algorithm comprises the following steps: comparing a known template with an area with the same size in an original image from the upper left corner (0,0) of the input image; recording a comparison result c; thirdly, translating to the next pixel, comparing the template with an area with the same size in the original image again, and recording the comparison result; and fourthly, repeating the steps till the lower right corner of the input image, wherein the image with the minimum difference with the original image is the target image. The traditional template matching algorithm mathematical model is easy to establish, the algorithm accuracy is high, but the speed is slow, and the requirement of real-time performance cannot be met sometimes along with the increase of data images.
In order to improve the problems of low efficiency and low calculation speed of the traditional template matching algorithm, the invention adopts a multi-step long template matching method. The multi-step template matching steps are as follows: the template image and the search subgraph are respectively in the x direction and the y direction according to the current step SxAnd SyCollecting data (S)x，SySet according to the size of the actual video image), and calculating a correlation coefficient R between the template image and the searched subgraph according to the data; if the current correlation coefficient R is smaller than the maximum value R of the correlation coefficientmaxAnd moving the search subgraph to the next position, otherwise taking half of the current step length in the x and y directions as a new step length, and then calculating until a matched image is found. The multi-step template matching does not need to perform traversal calculation on each pixel point of the image, and compared with the traditional template matching algorithm, the multi-step template matching greatly improves the calculation efficiency, reduces the calculation time and can meet the real-time requirement.
(7) Calculating offset pixels of the template image and the matching image;
according to the step (2) and the step (6), the offset pixel is calculated according to the coordinates of the central pixel coordinate system of the two image blocks in the embodiment.
(8) Calculating the real-time deflection angle of the camera according to the offset pixels obtained by calculation in the step (7) by utilizing the parameter and the geometric relation of the camera;
the calculation principle and formula are as follows:
and calculating whether the camera deflects or not and a real-time deflection angle in real time by using the visible angle and the geometric relation of the camera according to the coordinates of the central points of the known template image and the matched image. The focal length information and the visual angle of the camera can be known by referring to the relevant specification.
As shown in fig. 2, in the horizontal direction, α is a vertex angle of the triangle, which is also called a viewing angle (as referred to in the relevant specification), and β is a deflection angle in the horizontal direction; l is the distance from the camera lens to the imaging plane; x is a straight line corresponding to the offset pixel n; x is a straight line corresponding to the total pixel N; phi is the base angle of the triangle, phi is pi/2-alpha, and the relationship between the pixel and the angle mapping obtained by the sine theorem is
In a mirror image coordinate system, a is equal to l, b is equal to X/2, A is equal to phi, omega is equal to pi/2-beta, and n is an offset pixel in the horizontal direction; n is the total pixels in the horizontal direction. The substitution formula is:
namely, it is
The simplified formula is as follows:
the following steps are provided:
after operation, have
In the same X direction, the image pixel and angle mapping in Y direction can be obtained
M is an offset pixel in the vertical direction; m is the total pixel in the vertical direction;
is an offset angle in the vertical direction; δ is the viewing angle of the cross section.
(9) The deflection angle of the camera is larger than an angle threshold value, the threshold value is set according to the actual situation of the meeting place, and the system gives an early warning.
As shown in fig. 3, the present invention further provides a camera deflection real-time detection and early warning module for a polling meeting place, which includes a difference detection sub-module for detecting whether a scene of the meeting place changes, a face recognition sub-module for detecting information of the meeting place, a template matching sub-module for calculating the number of deflection pixels, an offset angle sub-module for converting into an actual deflection angle, an early warning information generation sub-module for generating final early warning information, a human-computer interaction interface, a parameter setting module, and a display screen. The input end of the difference detection submodule is connected with an external video network core switch and used for acquiring a video frame image; the input end of the face recognition submodule is connected with the output end of the difference detection submodule; the input end of the template matching submodule is connected with the output end of the face recognition submodule; the input end of the offset angle submodule is connected with the output end of the template matching submodule; the input end of the early warning information generation submodule is connected with the output end of the offset angle submodule; the input end of the human-computer interaction interface is respectively connected with the input end of the early warning information generation submodule and the parameter setting module, and the output end of the human-computer interaction interface is connected with the display.
Through the detection early warning module, the video conference meeting place camera can be timely detected and early warned of angle deflection, the reduction of conference video quality caused by the deflection of the camera is avoided, and the requirement of a polling conference on the conference video quality is met.
Finally, the above embodiments are only intended to illustrate the technical solutions of the present invention and not to limit the present invention, and although the present invention has been described in detail with reference to the preferred embodiments, it will be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions, and all of them should be covered by the claims of the present invention.
Claims (7)
1. The real-time detection and early warning method for the deflection of the camera aiming at the polling meeting place is characterized in that: the method comprises the following steps:
(1) acquiring a video frame image from a conference video which is not deflected, and simply referring the image to be an original image;
(2) intercepting an image block from the center of the original image, selecting the size of the image block according to the size of the video image, and simply using the image block as a template image;
(3) comparing the difference between the current video frame image and the previous video frame image, setting a difference threshold, and polling the next meeting place when the difference is greater than the threshold;
(4) performing face detection on the video frame image under the condition that the difference degree is smaller than the threshold value, and adding 1 to an empty field counter if no face is detected;
(5) the counting of the empty field counter reaches an empty field threshold value, the threshold value is set according to the actual situation of the meeting place, and the system gives an early warning;
(6) for the condition that the human face is detected, carrying out multi-step long template matching on the current video frame image and the template image to obtain a matched image;
(7) calculating offset pixels of the template image and the matching image;
(8) calculating the real-time deflection angle of the camera according to the offset pixels calculated in the step (7) by combining the parameters of the camera with the geometric relation;
(9) and if the deflection angle of the camera is larger than the angle threshold value, setting the threshold value according to the actual situation of the meeting place, and performing early warning by the system.
2. The real-time camera deflection detection and early warning method for polling meeting places according to claim 1, characterized in that: in the step (2), the coordinates of the central point pixel coordinate system of the intercepted template image are extracted.
3. The real-time camera deflection detection and early warning method for polling meeting places according to claim 1, characterized in that: in the step (4), a face detection algorithm based on SVM and HOG features is adopted, firstly, an SVM is used for training collected samples to obtain a face feature library, then, a sliding window is adopted to traverse all regions of the image to search regions where faces may exist, the size of a sub-window is required to be consistent with that of a training sample, HOG features of each sub-window are calculated, feature values of the HOG features are judged through an SVM classifier, if the face is the face, coordinate values of the face are recorded, otherwise, the window is abandoned, all the coordinate values are finally collected, and face calibration is completed on an original image.
4. The real-time camera deflection detection and early warning method for polling meeting places according to claim 1, characterized in that: in step (5), the matching image should extract coordinates of the pixel coordinate system of the center point.
5. The method for real-time detection and early warning of camera deflection for polling meeting places according to claim 1, 2 or 3, characterized in that: in the step (6), the matching of the multi-step long templates refers to: the template image and the search subgraph are respectively in the x direction and the y direction according to the current step SxAnd SyCollecting data, Sx，SyAnd setting according to the size of the actual video image, calculating a correlation coefficient R of the template image and the searched subgraph according to the data, moving the searched subgraph to the next position if the current correlation coefficient R is smaller than the maximum value Rmax of the correlation coefficient, and otherwise, taking half of the current step length in the x and y directions as a new step length, and calculating until a matched image is found.
6. The method for real-time detection and early warning of camera deflection for polling meeting places according to claim 1, 2 or 3, characterized in that: in the step (7), the offset pixels of the two image blocks are calculated according to the coordinates of the central pixel coordinate systems of the two image blocks.
7. The camera deflection real-time detection early warning device for the polling meeting place is characterized in that: the method comprises the following steps:
the difference detection submodule is used for detecting whether the meeting place scene changes;
the face recognition submodule is used for detecting meeting place information;
the template matching submodule is used for calculating the number of deflection pixels;
the offset angle submodule is used for converting the actual deflection angle into an actual deflection angle;
the early warning information generation submodule is used for generating final early warning information; and
a human-computer interaction interface, parameter setting and a display screen;
the human face recognition system comprises a human face recognition submodule, a template matching submodule, a deviation angle submodule, an early warning generation submodule, a parameter setting submodule and a display, wherein the difference detection submodule is connected with the input end of the human face recognition submodule, the output end of the human face recognition submodule is connected with the input end of the template matching submodule, the output end of the template matching submodule is connected with the input end of the deviation angle submodule, the output end of the deviation angle submodule is connected with the input end of the early warning generation submodule, the output end of the early warning generation submodule is connected with the input end of a human-computer interaction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710866772.7A CN107862713B (en) | 2017-09-22 | 2017-09-22 | Camera deflection real-time detection early warning method and module for polling meeting place |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710866772.7A CN107862713B (en) | 2017-09-22 | 2017-09-22 | Camera deflection real-time detection early warning method and module for polling meeting place |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107862713A CN107862713A (en) | 2018-03-30 |
| CN107862713B true CN107862713B (en) | 2021-04-06 |
Family
ID=61699599
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710866772.7A Active CN107862713B (en) | 2017-09-22 | 2017-09-22 | Camera deflection real-time detection early warning method and module for polling meeting place |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107862713B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108712627A (en) * | 2018-08-21 | 2018-10-26 | 云南电网有限责任公司 | Video conference content and quality detection apparatus and detection method |
| CN111178113B (en) * | 2018-11-09 | 2023-10-13 | 深圳技威时代科技有限公司 | Information processing method, device and storage medium |
| CN109948655A (en) * | 2019-02-21 | 2019-06-28 | 华中科技大学 | It is a kind of based on multi-level endoscope operation instruments detection method |
| CN110210461B (en) * | 2019-06-27 | 2021-03-05 | 北京澎思科技有限公司 | Multi-view collaborative abnormal behavior detection method based on camera grid |
| CN117459688B (en) * | 2023-12-26 | 2024-05-03 | 海纳云物联科技有限公司 | Camera angle marking method, device and medium based on map system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103139554A (en) * | 2011-11-22 | 2013-06-05 | 浙江大学 | Method and device for optimizing three-dimensional video frequency distortion |
| CN106503615A (en) * | 2016-09-20 | 2017-03-15 | 北京工业大学 | Indoor human body detecting and tracking and identification system based on multisensor |
-
2017
- 2017-09-22 CN CN201710866772.7A patent/CN107862713B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103139554A (en) * | 2011-11-22 | 2013-06-05 | 浙江大学 | Method and device for optimizing three-dimensional video frequency distortion |
| CN106503615A (en) * | 2016-09-20 | 2017-03-15 | 北京工业大学 | Indoor human body detecting and tracking and identification system based on multisensor |
Non-Patent Citations (2)
| Title |
|---|
| Projection Moire measurement of the deflection of composite plates subject to bird strike impact;A. Shulev 等;《SPIE Optical Metrology》;20070618;第1-2页 * |
| 视频监控中视频质量诊断系统的设计与实现;吴艳平 等;《计算机光盘软件与应用》;20140815(第15期);第1-2页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107862713A (en) | 2018-03-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107862713B (en) | Camera deflection real-time detection early warning method and module for polling meeting place | |
| CN104685513B (en) | According to the high-resolution estimation of the feature based of the low-resolution image caught using array source | |
| CN108470356B (en) | Target object rapid ranging method based on binocular vision | |
| CN111160291B (en) | Human eye detection method based on depth information and CNN | |
| CN110287907B (en) | Object detection method and device | |
| CN110189375B (en) | Image target identification method based on monocular vision measurement | |
| CN106845338B (en) | Pedestrian detection method and system in video stream | |
| Cvišić et al. | Recalibrating the KITTI dataset camera setup for improved odometry accuracy | |
| TW201926244A (en) | Real-time video stitching method | |
| CN111257329A (en) | Smartphone camera defect detection method and detection system | |
| CN109711267A (en) | A kind of pedestrian identifies again, pedestrian movement's orbit generation method and device | |
| CN112017212B (en) | Training and tracking method and system of face key point tracking model | |
| CN111325828B (en) | Three-dimensional face acquisition method and device based on three-dimensional camera | |
| CN113393524A (en) | Target pose estimation method combining deep learning and contour point cloud reconstruction | |
| CN103607558A (en) | Video monitoring system, target matching method and apparatus thereof | |
| CN105741326A (en) | Target tracking method for video sequence based on clustering fusion | |
| CN113255608B (en) | Multi-camera face recognition positioning method based on CNN classification | |
| CN111399634A (en) | Gesture-guided object recognition method and device | |
| WO2019041447A1 (en) | 3d video frame feature point extraction method and system | |
| CN112215036B (en) | Cross-mirror tracking method, device, equipment and storage medium | |
| JP7230722B2 (en) | Image processing device and image processing method | |
| TW202020732A (en) | Objective identification method and device thereof | |
| Fan et al. | Human-M3: A Multi-view Multi-modal Dataset for 3D Human Pose Estimation in Outdoor Scenes | |
| WO2020175085A1 (en) | Image processing apparatus and image processing method | |
| CN111854651A (en) | Indoor building area real-time measuring method based on SLAM |
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 |