CN106056624A - Unmanned aerial vehicle high-definition image small target detecting and tracking system and detecting and tracking method thereof - Google Patents
Unmanned aerial vehicle high-definition image small target detecting and tracking system and detecting and tracking method thereof Download PDFInfo
- Publication number
- CN106056624A CN106056624A CN201610349248.8A CN201610349248A CN106056624A CN 106056624 A CN106056624 A CN 106056624A CN 201610349248 A CN201610349248 A CN 201610349248A CN 106056624 A CN106056624 A CN 106056624A
- Authority
- CN
- China
- Prior art keywords
- image
- video
- target
- frame
- pixel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an unmanned aerial vehicle (UAV) high-definition image small target detecting and tracking system. The UAV high-definition image small target detection and tracking system comprises a video image collecting module, and the video image collecting module is in data connection with an image processing module and a video processing result display module in sequence. The video image collecting module is used for collecting a high-frequency video frame image shot by a UAV and sending the high-frequency video frame image to the image processing module. The image processing module is used for processing the received high-frequency video frame image, detecting a moving target, and outputting the position coordinate of the moving target in the image to the video processing result display module in real time. The invention further discloses a target detecting and tracking method. The problem that the existing high-definition image small target detecting and tracking algorithm applied to actual tasks is of poor real-time performance and a small target can be hardly detected using the algorithm is solved.
Description
Technical field
The invention belongs to unmanned plane technical field of image processing, be specifically related to unmanned plane high-definition image small target deteection with
Track system and detecting and tracking method thereof.
Background technology
Unmanned plane has that cost is low, good concealment, flexible operation, reusable, reduce the plurality of advantages such as casualties,
It is a kind of with high content of technology, airborne aircraft that effective utilization is good, obtains fast development in recent years.U.S. Department of Defense is " unmanned
Machine system development figure (2005-2030) " in the emphasis that clearly unmanned plane developed as nearly 30 years airborne vehicles.Meanwhile, current army
The applications such as thing scouting, earthquake relief work, target search, area anti-terrorism, information search, the detection of core Biochemical Information, rescue and relief work
Urgent needs, promoting the development of unmanned plane function technology.
Unmanned plane frontier defense investigation is main based on vedio data, due to unmanned on-line monitoring, it is necessary to pass through video
Image processing means forms the ability of unmanned plane frontier defense investigation.The method used in the world at present is generally by unmanned aerial vehicle vision
Frequently image transmission system, transmits view data to earth station, earth station staff monitors the situation of monitor area, but this side
Formula expends monitoring personnel's energy very much, for general frontier defense inspection task, needs to find Artificial intellectual technology, carrys out generation
This simple target detection identification mission is performed for people.Automatically the detection of research ground target and the technology of identification, can help
Monitoring personnel effectively monitor and find suspicious object, significantly alleviate the control and monitoring of monitoring personnel.
Correlation technique involved in technology is patrolled and examined in unmanned plane frontier defense both at home and abroad, at China's unmanned plane aircraft by analyzing
In the case of system platform is more ripe, the theoretical research of target detection tracking technique also has certain basis, find these sides
The applied research of method is the weakest.This invention will launch research for this technology application difficult problem, strive as early as possible by target
Detecting and tracking technology is applied to frontier defense inspection.The mission requirements geared to actual circumstances, it is achieved quickly, in real time, accurately find target also
And identification target, this is the application prospect getting a good eye value for frontier defense.It is not only for frontier defense, for the people
With field, all it is with a wide range of applications such as the tracking of environment investigation, the condition of a disaster investigation and law-breaker is arrested etc..
Summary of the invention
It is an object of the invention to provide a kind of unmanned plane high-definition image detecting and tracking dim small target system, to solve target
Detection and tracking technique are applied in the actual task of unmanned plane frontier defense investigation, patrol.
It is a further object to provide the detection of a kind of unmanned plane high-definition image detecting and tracking dim small target system
With tracking, to solve the real-time of the existing high-definition image detecting and tracking dim small target algorithm being applied in actual task
Difference, Small object is difficult to the problems such as detection.
The first technical scheme of the present invention is, unmanned plane high-definition image detecting and tracking dim small target system, bag
Including video image acquisition module, video image acquisition module data cube computation successively has image processing module and video processing results to show
Show module；
Video image acquisition module, for gathering the high frequency video two field picture of unmanned plane shooting, and by high frequency video frame figure
As sending to image processing module；
Image processing module, for the high frequency video two field picture received is carried out image procossing, and detects moving target,
Output moving target position coordinates in the picture is to video processing results display module in real time.
Further, video image acquisition module connects Image transmission equipment, and Image transmission equipment is for receiving unmanned plane shooting
High frequency video two field picture, and send it to video image acquisition module；
Video image acquisition module turns usb signal interface by HDMI and is connected to image processing module, and HDMI turns usb signal
Interface is for being converted to USB video signal by high frequency video two field picture.
The second technical scheme of the present invention is, a kind of above-mentioned unmanned plane high-definition image Small object adhere to survey with
The detecting and tracking method of track system, specifically implements according to following steps:
Step 1, carries out pretreatment to the high frequency video two field picture of unmanned plane shooting, then reads at regular intervals
The sequence frame of high frequency video two field picture；
Step 2, the target area in frame difference method detection target video image:
Consecutive frame image is used time difference based on pixel, by closing the moving region that value is extracted in image；
Step 3, uses the moving target of Lucas-Kanade optical flow method detection moving region, and calculates moving target
Position in the picture；
Step 4, according to the positional information of the moving target that step 3 obtains, marks moving target in original image, and
Carry out real-time tracking.
Further, step 1 method particularly includes:
In image processing module, create a video interface catcher, obtain computer USB interface passage IP, set video
Image sequence frame reads time interval, reads video image with the time interval that sets, and reads the parameter information of this image,
This image is stored in internal memory simultaneously.
Further, step 2 method particularly includes:
In image processing module, first consecutive frame image respective pixel value is subtracted each other and obtain difference image, then to difference
Partial image binaryzation, in the case of ambient brightness change is little, if the change of respective pixel value is less than the threshold value preset, i.e.
Described pixel is background pixel；If the pixel value change of image-region is more than or equal to the threshold value preset, the most described pixel is
Foreground pixel, determines moving target position in the picture by background pixel and foreground pixel.
Further, the concrete grammar of the difference of image sequence frame by frame is,
D=| IL(x,y,i)-IL(x, y, i-1) |,
Wherein, IDLIt is neighbor frame difference figure, IL(x, y, i) and IL(x, y, i-1) is that the brightness of the i-th and i-th-1 frame divides respectively
Amount, i represent frame number (i-1 ..., N), N is sequence totalframes, and T is threshold value, and d represents the luminance component of the i-th and i-th-1 two field picture
Difference, x and y represents the transverse and longitudinal coordinate of pixel respectively.
Further, step 3 method particularly includes:
Utilize Harris angular-point detection method to detect the characteristic point of moving region, and calculate the coordinate P of characteristic pointi
(x y), utilizes the Lucas-Kanade optical flow method improved to estimate the position P in next frame image of the feature in present framei′
(x, y),
Work as Pi' (x, number n ＜ a y) andThe value of wherein a, b
For empirical value, then it is assumed that these characteristic points are the characteristic point of moving target.
Further, step 4 method particularly includes:
In video processing results display module, according to the coordinate of the target pixel points that step 3 obtains, at original video figure
As this point of labelling in frame, reading at a certain time interval and process this image simultaneously, the result detected the most at last is shown to use
Interface, family.
Accompanying drawing explanation
Fig. 1 is the structural representation of frontier defense unmanned plane high-definition image detecting and tracking dim small target system of the present invention；
Fig. 2 is the method flow diagram using Lucas-Kanade optical flow method detection moving region target in the present invention.
In figure, 1. video image acquisition module, 2. image processing module, 3. Video processing display module, 4. Image transmission equipment,
5.HDMI turns usb signal interface, 6.PC processor.
Detailed description of the invention
The present invention is described in detail with embodiment below in conjunction with the accompanying drawings.
The invention provides a kind of unmanned plane high-definition image detecting and tracking dim small target system, see structure chart 1, video figure
As acquisition module 1, image processing module 2, Video processing display module 3, Image transmission equipment 4, HDMI turns usb signal interface 5, at PC
Reason device 6.
Wherein, the input of video image acquisition module 1 connects Image transmission equipment 4, receives video signal, output termination HDMI
Turn usb signal interface 5, HDMI video signal is converted to the usb signal that PC processor 6 can receive；In PC processor mainly
Comprise image processing module 2 and Video processing display module 3；Image processing module 2 reads the USB end video figure of PC processor 6
As carrying out Computer Vision, detect target and calculate the positional information of target；Video processing display module 3 is by defeated for module 2
The target position information gone out is loaded in raw video image, and labelling also outlines target, it is achieved detects in real time and follows the tracks of target.
Present invention also offers a kind of detecting and tracking method of unmanned plane high-definition image detecting and tracking dim small target system, tool
Body is implemented according to following steps:
Step 1, high clear video image pretreatment, create a video interface catcher, obtain computer USB interface passage
IP, sets sequence of video images frame and reads time interval, reads video image with the time interval set, and reads this image
Parameter information, this image is stored in internal memory simultaneously, is labeled as present frame, treat that the image processing algorithm of lower step processes.
Wherein, the work of pretreatment is the parameter information reading image, is stored in internal memory by this image simultaneously, is labeled as present frame.
Step 2, the target area in frame difference method detection image, it is poor first consecutive frame image respective pixel value to be subtracted each other
Partial image, then to difference image binaryzation, in the case of ambient brightness change is little, if the change of respective pixel value is less than
During pre-determined threshold value, it is believed that be background pixel herein；If the pixel value change of image-region is more than or equal in advance
During the threshold value determined, it is believed that this due in image moving object cause, be foreground pixel by these zone markers, profit
Moving target position in the picture is may determine that with the pixel region of labelling.The difference of image sequence frame by frame, is equivalent to figure
As sequence carries out the high-pass filtering in time domain.Its formula is as follows:
D=| IL(x,y,i)-IL(x, y, i-1) |,
Wherein, IDLIt is neighbor frame difference figure, IL(x, y, i) and IL(x, y, i-1) is that the brightness of the i-th and i-th-1 frame divides respectively
Amount, i represent frame number (i-1 ..., N),NFor sequence totalframes, T is threshold value, and d represents the luminance component of the i-th and i-th-1 two field picture
Difference, x and y represents the transverse and longitudinal coordinate of pixel respectively.
Step 3, Lucas-Kanade optical flow method detection moving region target, method particularly includes:
See the flow chart of Fig. 2, utilize Harris angular-point detection method, detect the spy of described moving region at present frame
Levy a little, and calculate the coordinate P of characteristic pointi(x,y)；
The Lucas-Kanade optical flow method improved is utilized to estimate the position P in next frame image of the feature in present framei′
(x,y)；
Work as Pi' (x, number n ＜ a y), andWherein a, b's
Value is empirical value, then it is assumed that these characteristic points are the characteristic point of moving target；
These characteristic points are classified, adjacent characteristic point coordinate is taken average Pj' (x, y), as coordinates of targets.
Step 4, according to the coordinate of the target pixel points that step 3 obtains, this point of labelling in raw video image frame, and use
Rectangle circle lives this target.Reading at a certain time interval and process this image simultaneously, the result detected the most at last is shown to
User interface.
The present invention uses digital high-definition Image transmission equipment, and the video image that unmanned plane shoots is passed to earth station, and real-time
Gather sequence of frames of video.The video image gathered is the high-definition image of 1080P, distinguishes existing simulation low-resolution image,
Can reduce and video image is carried out the online or difficulty of later image process, such as target detection, identify, follow the tracks of and position
Deng.
The present invention uses the algorithm of target detection of improvement, i.e. frame difference method to combine with the Lucas-Kanade optical flow method of improvement
Hybrid detection algorithm.Frame difference method is for the detection of moving target under static background, has the fireballing advantage of calculating；And light
Stream Master is for the detection of moving target under dynamic background (motion platform), but optical flow method amount of calculation is bigger, it is difficult to meet real
Shi Xing.The method using both approaches to combine, had both used frame difference method to detect motion target area, had carried out in this region
Optical flow method detects, and is possible not only to reduce amount of calculation, meets the mission requirements under unmanned plane motion platform simultaneously.Additionally, in tradition
Lucas-Kanade optical flow method on the basis of, the present invention uses the thought of mean filter, reduces the false drop rate of target, for rear
The target recognition of phase is followed the tracks of and is reduced difficulty.
Claims (8)
1. unmanned plane high-definition image detecting and tracking dim small target system, it is characterised in that include video image acquisition module (1),
Described video image acquisition module (1) data cube computation successively has image processing module (2) and video processing results display module
(3)；
Described video image acquisition module (1), for gathering the high frequency video two field picture of unmanned plane shooting, and regards described high frequency
Frequently two field picture sends to described image processing module (2)；
Described image processing module (2), for the described high frequency video two field picture received is carried out image procossing, and detects fortune
Moving-target, output moving target position coordinates in the picture is to described video processing results display module (3) in real time.
2. the system as claimed in claim 1, it is characterised in that described video image acquisition module (1) connects Image transmission equipment
(4), described Image transmission equipment (4) is for receiving the high frequency video two field picture of unmanned plane shooting, and sends it to described video figure
As acquisition module (1)；
Described video image acquisition module (1) turns usb signal interface (5) by HDMI and is connected to described image processing module (2),
Described HDMI turns usb signal interface (5) for high frequency video two field picture is converted to USB video signal.
3. a unmanned plane high-definition image Small object as claimed in claim 1 or 2 adheres to surveying and the detecting and tracking of the system of tracking
Method, it is characterised in that specifically implement according to following steps:
Step 1, carries out pretreatment to the high frequency video two field picture of unmanned plane shooting, then reads high frequency at regular intervals
The sequence frame of video frame images；
Step 2, the target area in frame difference method detection target video image:
Consecutive frame image is used time difference based on pixel, by closing the moving region that value is extracted in image；
Step 3, uses Lucas-Kanade optical flow method to detect the moving target of described moving region, and calculates moving target
Position in the picture；
Step 4, according to the positional information of the moving target that step 3 obtains, marks moving target in original image, and carries out
Real-time tracking.
4. target detection tracking method as claimed in claim 3, it is characterised in that described step 1 method particularly includes:
In image processing module (2), create a video interface catcher, obtain computer USB interface passage IP, set video
Image sequence frame reads time interval, reads video image with the time interval that sets, and reads the parameter information of this image,
This image is stored in internal memory simultaneously.
5. target detection tracking method as claimed in claim 3, it is characterised in that described step 2 method particularly includes:
In image processing module (2), first consecutive frame image respective pixel value is subtracted each other and obtain difference image, then to difference
Image binaryzation, in the case of ambient brightness change is little, if the change of respective pixel value is less than the threshold value preset, i.e. institute
Stating pixel is background pixel；If the pixel value change of image-region is more than or equal to the threshold value preset, the most described pixel is front
Scene element, determines moving target position in the picture according to background pixel and foreground pixel.
6. target detection tracking method as claimed in claim 5, it is characterised in that described image sequence frame by frame difference concrete
Method is,
D=| IL(x,y,i)-IL(x, y, i-1) |,
Wherein, IDLIt is neighbor frame difference figure, IL(x, y, i) and IL(x, y, i-1) is the luminance component of the i-th and i-th-1 frame respectively, i
Expression frame number (i-1 ..., N), N is sequence totalframes, and T is threshold value, and d represents the luminance component difference of the i-th and i-th-1 two field picture,
X and y represents the transverse and longitudinal coordinate of pixel respectively.
7. target detection tracking method as claimed in claim 3, it is characterised in that described step 3 method particularly includes:
Utilize Harris angular-point detection method to detect the characteristic point of described moving region, and calculate the coordinate P of characteristic pointi(x,
Y), the Lucas-Kanade optical flow method improved is utilized to estimate the position P in next frame image of the feature in present framei′(x,
Y),
Work as Pi' (x, number n ＜ a y) andThe value of wherein a, b is warp
Test value, then it is assumed that these characteristic points are the characteristic point of moving target.
8. target detection tracking method as claimed in claim 3, it is characterised in that described step 4 method particularly includes:
In video processing results display module (3), according to the coordinate of the target pixel points that step 3 obtains, at original video figure
As this point of labelling in frame, reading at a certain time interval and process this image simultaneously, the result detected the most at last is shown to use
Interface, family.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610349248.8A CN106056624A (en) | 2016-05-24 | 2016-05-24 | Unmanned aerial vehicle high-definition image small target detecting and tracking system and detecting and tracking method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610349248.8A CN106056624A (en) | 2016-05-24 | 2016-05-24 | Unmanned aerial vehicle high-definition image small target detecting and tracking system and detecting and tracking method thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106056624A true CN106056624A (en) | 2016-10-26 |
Family
ID=57175839
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610349248.8A Pending CN106056624A (en) | 2016-05-24 | 2016-05-24 | Unmanned aerial vehicle high-definition image small target detecting and tracking system and detecting and tracking method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106056624A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106845364A (en) * | 2016-12-28 | 2017-06-13 | 中国航天电子技术研究院 | A kind of fast automatic object detection method |
| CN107993252A (en) * | 2017-11-29 | 2018-05-04 | 天津聚飞创新科技有限公司 | Subscriber tracing system, method and device |
| CN108508912A (en) * | 2018-03-20 | 2018-09-07 | 陈超 | Detecting system and method based on computer control |
| CN108537726A (en) * | 2017-03-03 | 2018-09-14 | 杭州海康威视数字技术股份有限公司 | A kind of method of track up, equipment and unmanned plane |
| CN109635649A (en) * | 2018-11-05 | 2019-04-16 | 航天时代飞鸿技术有限公司 | A kind of high speed detection method and system of unmanned plane spot |
| CN110866472A (en) * | 2019-11-04 | 2020-03-06 | 西北工业大学 | Unmanned aerial vehicle ground moving target identification and image enhancement system and method |
| CN111311645A (en) * | 2020-02-25 | 2020-06-19 | 四川新视创伟超高清科技有限公司 | Ultrahigh-definition video cut target tracking and identifying method |
| CN111832379A (en) * | 2019-10-15 | 2020-10-27 | 中国石油化工股份有限公司 | Unmanned aerial vehicle real-time video detection system based on convolutional neural network |
| CN113204986A (en) * | 2020-12-11 | 2021-08-03 | 深圳市科卫泰实业发展有限公司 | Moving target detection method suitable for unmanned aerial vehicle |
| CN114119522A (en) * | 2021-11-17 | 2022-03-01 | 北京华能新锐控制技术有限公司 | Visual detection method for coal blockage of coal conveying belt |
| CN116862953A (en) * | 2023-09-05 | 2023-10-10 | 天津大学 | Motion background-oriented real-time detection tracking device and method for small moving target |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102855467A (en) * | 2012-07-30 | 2013-01-02 | 上海未来软件有限公司 | Method for extracting and identifying plate numbers from video streaming |
| CN103413444A (en) * | 2013-08-26 | 2013-11-27 | 深圳市川大智胜科技发展有限公司 | Traffic flow surveying and handling method based on unmanned aerial vehicle high-definition video |
| CN104102218A (en) * | 2014-06-30 | 2014-10-15 | 西北工业大学 | Visual servo-based sense-and-avoid method and system |
-
2016
- 2016-05-24 CN CN201610349248.8A patent/CN106056624A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102855467A (en) * | 2012-07-30 | 2013-01-02 | 上海未来软件有限公司 | Method for extracting and identifying plate numbers from video streaming |
| CN103413444A (en) * | 2013-08-26 | 2013-11-27 | 深圳市川大智胜科技发展有限公司 | Traffic flow surveying and handling method based on unmanned aerial vehicle high-definition video |
| CN104102218A (en) * | 2014-06-30 | 2014-10-15 | 西北工业大学 | Visual servo-based sense-and-avoid method and system |
Non-Patent Citations (2)
| Title |
|---|
| HOUSSAM SALMANE, ET AL.: "Object tracking using Harris corner points based optical flow propagation and Kalman filter", 《2011 14TH INTERNATIONAL IEEE CONFERENCE ON INTELLIGENT TRANSPORTATION SYSTEMS》 * |
| 袁国武 等: "一种结合光流法与三帧差分法的运动目标检测算法", 《小型微型计算机系统》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106845364A (en) * | 2016-12-28 | 2017-06-13 | 中国航天电子技术研究院 | A kind of fast automatic object detection method |
| CN106845364B (en) * | 2016-12-28 | 2020-06-09 | 中国航天电子技术研究院 | Rapid automatic target detection method |
| CN108537726A (en) * | 2017-03-03 | 2018-09-14 | 杭州海康威视数字技术股份有限公司 | A kind of method of track up, equipment and unmanned plane |
| CN108537726B (en) * | 2017-03-03 | 2022-01-04 | 杭州海康威视数字技术股份有限公司 | Tracking shooting method and device and unmanned aerial vehicle |
| CN107993252A (en) * | 2017-11-29 | 2018-05-04 | 天津聚飞创新科技有限公司 | Subscriber tracing system, method and device |
| CN108508912A (en) * | 2018-03-20 | 2018-09-07 | 陈超 | Detecting system and method based on computer control |
| CN109635649A (en) * | 2018-11-05 | 2019-04-16 | 航天时代飞鸿技术有限公司 | A kind of high speed detection method and system of unmanned plane spot |
| CN111832379A (en) * | 2019-10-15 | 2020-10-27 | 中国石油化工股份有限公司 | Unmanned aerial vehicle real-time video detection system based on convolutional neural network |
| CN110866472A (en) * | 2019-11-04 | 2020-03-06 | 西北工业大学 | Unmanned aerial vehicle ground moving target identification and image enhancement system and method |
| CN111311645A (en) * | 2020-02-25 | 2020-06-19 | 四川新视创伟超高清科技有限公司 | Ultrahigh-definition video cut target tracking and identifying method |
| CN113204986A (en) * | 2020-12-11 | 2021-08-03 | 深圳市科卫泰实业发展有限公司 | Moving target detection method suitable for unmanned aerial vehicle |
| CN114119522A (en) * | 2021-11-17 | 2022-03-01 | 北京华能新锐控制技术有限公司 | Visual detection method for coal blockage of coal conveying belt |
| CN116862953A (en) * | 2023-09-05 | 2023-10-10 | 天津大学 | Motion background-oriented real-time detection tracking device and method for small moving target |
| CN116862953B (en) * | 2023-09-05 | 2024-01-19 | 天津大学 | Motion background-oriented real-time detection tracking device and method for small moving target |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106056624A (en) | Unmanned aerial vehicle high-definition image small target detecting and tracking system and detecting and tracking method thereof | |
| CN109819208B (en) | Intensive population security monitoring management method based on artificial intelligence dynamic monitoring | |
| CN109272530B (en) | Target tracking method and device for space-based monitoring scene | |
| CN106356757B (en) | A kind of power circuit unmanned plane method for inspecting based on human-eye visual characteristic | |
| CN106650620B (en) | A kind of target person identification method for tracing using unmanned plane monitoring | |
| Craye et al. | Spatio-temporal semantic segmentation for drone detection | |
| CN109584213B (en) | Multi-target number selection tracking method | |
| Chen et al. | A deep learning approach to drone monitoring | |
| CN103824070B (en) | A kind of rapid pedestrian detection method based on computer vision | |
| CN109711318B (en) | Multi-face detection and tracking method based on video stream | |
| Sommer et al. | Flying object detection for automatic UAV recognition | |
| CN104794468A (en) | Human face detection and tracking method based on unmanned aerial vehicle mobile platform | |
| Kruthiventi et al. | Low-light pedestrian detection from RGB images using multi-modal knowledge distillation | |
| CN110619276B (en) | Anomaly and violence detection system and method based on unmanned aerial vehicle mobile monitoring | |
| CN110490043A (en) | A kind of forest rocket detection method based on region division and feature extraction | |
| CN103729620B (en) | A kind of multi-view pedestrian detection method based on multi-view Bayesian network | |
| CN106875419A (en) | Small and weak tracking of maneuvering target based on NCC matching frame differences loses weight detecting method | |
| CN107146200A (en) | A kind of unmanned aerial vehicle remote sensing image split-joint method based on image mosaic quality evaluation | |
| CN104794737A (en) | Depth-information-aided particle filter tracking method | |
| CN110276286B (en) | Embedded panoramic video stitching system based on TX2 | |
| CN112037252A (en) | Eagle eye vision-based target tracking method and system | |
| CN106295695A (en) | A kind of takeoff and landing process automatic tracing image pickup method and device | |
| Saif et al. | Crowd density estimation from autonomous drones using deep learning: challenges and applications | |
| CN103870847B (en) | Detecting method for moving object of over-the-ground monitoring under low-luminance environment | |
| Hwang et al. | Aircraft detection using deep convolutional neural network in small unmanned aircraft systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB03 | Change of inventor or designer information |
Inventor after: Zhao Chunhui Inventor after: Gu Jiade Inventor after: Hu Jinwen Inventor after: Lv Yang Inventor before: Gu Jiade Inventor before: Zhao Chunhui Inventor before: Hu Jinwen Inventor before: Lv Yang Inventor before: Zhang Zhiyuan Inventor before: Fan Bin Inventor before: Jiang Shan Inventor before: Li Sijia |
|
| CB03 | Change of inventor or designer information | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20161026 |
|
| RJ01 | Rejection of invention patent application after publication |