CN110443247A - A kind of unmanned aerial vehicle moving small target real-time detecting system and method - Google Patents
A kind of unmanned aerial vehicle moving small target real-time detecting system and method Download PDFInfo
- Publication number
- CN110443247A CN110443247A CN201910778929.XA CN201910778929A CN110443247A CN 110443247 A CN110443247 A CN 110443247A CN 201910778929 A CN201910778929 A CN 201910778929A CN 110443247 A CN110443247 A CN 110443247A
- Authority
- CN
- China
- Prior art keywords
- target
- camera
- aerial vehicle
- unmanned aerial
- image
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/255—Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/67—Focus control based on electronic image sensor signals
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of unmanned aerial vehicle moving small target real-time detecting system and method, the system comprises: varifocal camera and miniature computing platform;The varifocal camera is also used to carry out zoom, photographic subjects video image according to zoom instructions for shooting video image;The miniature computing platform, including CPU processor and GPU processor;The CPU processor, for executing following step: detecting several moving small targets from video image;Calculate position of the center of gravity relative to origin for the target that each is detected;A target is selected, the coordinate position in its real space relative to camera is calculated according to the position of the target in the picture, issues attitude regulating command to unmanned aerial vehicle control system, while sending zoom instructions to varifocal camera;The GPU processor, for executing following step: target video image being inputted trained deep neural network in advance, exports recognized targeted species and probability, and carry out the choosing of profile circle to target.
Description
Technical field
The present invention relates to moving object detection fields, and in particular to a kind of unmanned aerial vehicle moving small target real-time detection system
System and method.
Background technique
Moving object detection and identification are the basic problems in computer vision, especially at a distance to moving small target
Detection is the important problem currently faced.For the Target detection and identification in optical detection, it is based primarily upon two kinds of information: one
Kind is the visual signature of target, including texture, color and shape information;One is the motion informations of target.Common target inspection
Survey method includes inter-frame difference and background extracting, image segmentation and characteristic matching, optical flow method and develops burning hot machine at present
Learning method.The real-time that these methods have is good, but big by background influence or need priori knowledge;Some has preferable inspection
Effect is surveyed, but it is computationally intensive or easy by influence of noise;The method of machine learning has the target of scale, rotational deformation
Well adapting to property, but can not effectively be identified when target very little.For unmanned aerial vehicle moving small target detection and
For identification, have the characteristics that target is small, it is high to calculate requirement of real-time, existing method is not able to satisfy target detection requirement.Cause
This, studies a kind of unmanned aerial vehicle moving small target real-time detection and is of great significance with recognition methods.
Summary of the invention
It is an object of the invention to solve the problems, such as unmanned plane to the real-time detection and identification of ground moving small target, Yi Zhongwu
Man-machine ground motion Small object real-time detection method can in real time find ground Small object real-time detection, pass through camera zoom
Recognition and tracking is carried out to target.
To achieve the goals above, the invention proposes a kind of unmanned aerial vehicle moving small target real-time detecting system, institutes
The system of stating includes: varifocal camera and miniature computing platform;
The varifocal camera is also used to carry out zoom, photographic subjects view according to zoom instructions for shooting video image
Frequency image;
The miniature computing platform, including CPU processor and GPU processor;
The CPU processor, for executing following step:
Video image is obtained, image is carried out down-sampled;
Moving object detection, several moving small targets detected are carried out to down-sampled image;
Using picture centre as coordinate origin, position of the center of gravity relative to origin for the target that each is detected is calculated;
A target is selected, the coordinate in its real space relative to camera is calculated according to the position of the target in the picture
Position issues attitude regulating command to unmanned aerial vehicle control system, camera is made to be directed toward target;Then zoom is sent to varifocal camera
Instruction;
The GPU processor, for executing following step: by target video image input, trained depth is neural in advance
Network exports recognized targeted species and probability, and carries out the choosing of profile circle to target.One kind as above system changes
Into described to down-sampled image progress moving object detection, several targets detected;It specifically includes:
Moving object detection is carried out to down-sampled image, obtains testing result;
Binary conversion treatment is carried out to testing result, the pixel of background position is 0, the pixel of moving target position
Value is 1;
Corrosion parameter and expansion parameters are set according to the number of pixels of detected minimum target and the size of false alarm rate,
Corrosion and expansive working are carried out to detected Small object,
Clustering processing, every a kind of corresponding moving target are carried out for 1 pixel to value.
As a kind of improvement of above system, the moving object detection include: background extracting, inter-frame difference, optical flow method,
Image segmentation, characteristic matching and machine learning.
Based on above system, the invention also provides a kind of moving small target real-time detection side based on unmanned aerial vehicle platform
Method, which comprises
The video image for obtaining varifocal camera carries out image down-sampled;
Moving object detection, several moving small targets detected are carried out to down-sampled image;
Using picture centre as coordinate origin, position of the center of gravity relative to origin for the target that each is detected is calculated;
A target is selected, the coordinate in its real space relative to camera is calculated according to the position of the target in the picture
Position sends instruction adjustment UAV Attitude to unmanned aerial vehicle control system, so that camera is directed toward target, while by varifocal phase
Machine, which sends zoom instructions, makes camera zoom, then obtains the target video image of camera shooting;
Target video image is inputted into trained deep neural network in advance, exports recognized targeted species and general
Rate, and the choosing of profile circle is carried out to target.
It is described that moving object detection is carried out to down-sampled image as a kind of improvement of the above method, it is detected
Several targets;It specifically includes:
Moving object detection is carried out to down-sampled image, obtains testing result;
Binary conversion treatment is carried out to testing result, the pixel of background position is 0, the pixel of moving target position
Value is 1;
Corrosion parameter and expansion parameters are set according to the number of pixels of detected minimum target and the size of false alarm rate,
Corrosion and expansive working are carried out to detected Small object,
Clustering processing, every a kind of corresponding moving target are carried out for 1 pixel to value.
As a kind of improvement of the above method, the moving object detection include: background extracting, inter-frame difference, optical flow method,
Image segmentation, characteristic matching and machine learning.
As a kind of improvement of the above method, the method also includes: the training step of deep neural network, it is specific to wrap
It includes:
Under different location, different angle, different distance, several target types to be identified are acquired using unmanned plane
Destination image data, data collected are labeled, marked content include to target carry out the choosing of profile circle, label target
Classification generates training sample;
The deep neural network for using SSD-MobileNet structure is established, wherein by Standard convolution core in MobileNet
It is separated into depth convolution kernel and puts convolution kernel;The port number of characteristic pattern is M, and convolution kernel size is DK*DK, output channel number is N,
Output characteristic pattern size is DF*DF, then Standard convolution core is M*DK*DK* N, the depth convolution kernel size after separation are DK*DK* M,
Point convolution kernel size is M*N*1*1;In SSD network, VGG structure is substituted for MobileNet structure;
Deep neural network is trained using training sample, the loss function of use includes Classification Loss and positioning damage
Two parts are lost, trained deep neural network is obtained.
Compared with the prior art, the advantages of the present invention are as follows:
Moving object detection is divided into two steps by method of the invention: first is that by using motion information at a distance may be used
Doubtful target is detected, second is that carrying out zoom to camera realizes that the close-in target based on deep neural network identifies, the first step
It is calculated using the CUP of processor, second step is calculated using GPU;In such a way that this two step is walked, can effectively it mention
It rises target detection precision and real-time and reduces false alarm rate.
Detailed description of the invention
Fig. 1 is the structure chart of the unmanned aerial vehicle moving small target real-time detecting system of the embodiment of the present invention 1;
Fig. 2 is the flow chart of the unmanned aerial vehicle moving small target real-time detection method of the embodiment of the present invention 2.
Specific embodiment
Technical solution of the present invention is described in detail in the following with reference to the drawings and specific embodiments.
Embodiment 1:
Basic principle of the invention is to be obtained by the varifocal camera of UAV flight and miniature computing platform by camera
Ground video data detects the moving small target in video using the motion information of target, and testing result is as doubtful
Target carries out zoom to camera, short distance observed object utilizes the texture of target, size then for selected suspicious object
Target is identified with shape information.Miniature computing platform is mainly used for receiving camera data, carries out the detection and knowledge of target
Not, the publication of camera control instruction.
The embodiment of the present invention 1 provides a kind of unmanned aerial vehicle moving small target real-time detecting system, the system packet
It includes: varifocal camera and miniature computing platform;
The varifocal camera is also used to carry out zoom, photographic subjects view according to zoom instructions for shooting video image
Frequency image;
The miniature computing platform, including CPU processor and GPU processor;
The CPU processor, for executing following step:
Video image is obtained, image is carried out down-sampled;
Moving object detection, several moving small targets detected are carried out to down-sampled image;
Using picture centre as coordinate origin, position of the center of gravity relative to origin for the target that each is detected is calculated;
A target is selected, the coordinate in its real space relative to camera is calculated according to the position of the target in the picture
Position issues attitude regulating command to unmanned aerial vehicle control system, camera is made to be directed toward target;Then zoom is sent to varifocal camera
Instruction;
The GPU processor, for executing following step: by target video image input, trained depth is neural in advance
Network exports recognized targeted species and probability, and carries out the choosing of profile circle to target.
Embodiment 2
Target range farther out when using motion information to suspicious object carry out detection and coarse positioning, according to suspicious object
Location information adjustment camera posture simultaneously control camera carry out zoom, realize to can with the close-ups of target, obtain it is apparent
Target information, then target is identified using the method for deep learning.
As shown in Fig. 2, the embodiment of the present invention 2 provides a kind of unmanned aerial vehicle moving small target real-time detection method,
Specific steps are as follows:
Step 1) obtains video by unmanned plane camera mounted, by video input to unmanned plane microcomputer mounted
It calculates and is handled in platform;
Step 2) is down-sampled to image progress in order to promote processing speed, according to the resolution setting for actually entering image
Image resolution ratio is reduced to 720p by down-sampled rate;
Step 3) carries out Moving small targets detection to down-sampled image:
1) object detection method being able to use includes but is not limited to background extracting, inter-frame difference, optical flow method, image point
It cuts, the methods of characteristic matching and machine learning, binary conversion treatment is carried out to testing result, the pixel of background position is 0,
The pixel value of moving target position is 1;
2) to detected Small object carry out corrosion and expansive working, according to the number of pixels of detected minimum target with
And the size setting corrosion parameter and expansion parameters of false alarm rate;
3) after corrosion expansion, clustering processing is carried out for 1 pixel to value, every a kind of corresponding moving target calculates every
The position of centre of gravity of a target;
Step 4) calculates position of the center of gravity relative to origin for the target that each is detected using picture centre as coordinate origin
It sets, the nearest several targets in range image center are filtered out according to the demand of target detection number and is compiled according to distance
Number processing;
Above step carries out operation using the CPU of miniature computing platform;
Step 5) select some number target, according to the position of target in the picture calculate in its real space relative to
The coordinate position of camera adjusts camera posture, so that camera is directed toward target, while carrying out Zoom control to camera, closely sees
Observation of eyes mark;
Step 6) camera is transferred to the target identification program based on deep learning after starting zoom, will preparatory trained depth
Neural network parameter is stored into miniature computing platform, identifies that the step makes to the target in frame image each in video
Operation is carried out with the GPU of miniature computing platform, target identification principle is as follows:
1) training of neural network is used for according to the target type to be identified acquisition target data, using unmanned plane not
Destination image data is acquired under same place, different angle, different distance, acquires amount of images at 10,000 or more, to collected
Data are labeled, and marked content includes that circle choosing, label target classification and attribute are carried out to target;
2) deep neural network structure uses SSD-MobileNet structure, wherein by Standard convolution core in MobileNet
It is separated into depth convolution kernel and puts convolution kernel.If the port number of characteristic pattern is M, convolution kernel size is DK*DK, output channel number is
N, output characteristic pattern size are DF*DF, then Standard convolution core is M*DK*DK* N, the depth convolution kernel size after separation are DK*DK*
M, point convolution kernel size are M*N*1*1;In SSD network, VGG structure is substituted for MobileNet structure, the loss letter of use
Number includes Classification Loss and positioning loss two parts;
3) it after setting network structure, is trained using data set collected, obtains the network for target identification
Parameter simultaneously stores network parameter into mobile processor;
4) identification that target is carried out for the image of each frame input, exports recognized targeted species and probability, circle
Select objective contour;
Step 7) adjusts UAV Attitude in real time and camera is directed toward according to the target position identified in each frame, right
Target is tracked.
It should be noted last that the above examples are only used to illustrate the technical scheme of the present invention and are not limiting.Although ginseng
It is described the invention in detail according to embodiment, those skilled in the art should understand that, to technical side of the invention
Case is modified or replaced equivalently, and without departure from the spirit and scope of technical solution of the present invention, should all be covered in the present invention
Scope of the claims in.
Claims (7)
1. a kind of unmanned aerial vehicle moving small target real-time detecting system, which is characterized in that the system comprises: varifocal camera
With miniature computing platform;
The varifocal camera is also used to carry out zoom, photographic subjects video figure according to zoom instructions for shooting video image
Picture;
The miniature computing platform, including CPU processor and GPU processor;
The CPU processor, for executing following step:
Video image is obtained, image is carried out down-sampled;
Moving object detection, several moving small targets detected are carried out to down-sampled image;
Using picture centre as coordinate origin, position of the center of gravity relative to origin for the target that each is detected is calculated;
A target is selected, the coordinate bit in its real space relative to camera is calculated according to the position of the target in the picture
It sets, issues attitude regulating command to unmanned aerial vehicle control system, camera is made to be directed toward target;Then zoom is sent to varifocal camera to refer to
It enables;
The GPU processor, for executing following step: target video image is inputted trained depth nerve net in advance
Network exports recognized targeted species and probability, and carries out the choosing of profile circle to target.
2. unmanned aerial vehicle moving small target real-time detecting system according to claim 1, which is characterized in that described pair of drop
The image of sampling carries out moving object detection, several targets detected;It specifically includes:
Moving object detection is carried out to down-sampled image, obtains testing result;
Binary conversion treatment is carried out to testing result, the pixel of background position is 0, and the pixel value of moving target position is
1;
According to the number of pixels of detected minimum target and the size of false alarm rate setting corrosion parameter and expansion parameters, to institute
The Small object detected carries out corrosion and expansive working,
Clustering processing, every a kind of corresponding moving target are carried out for 1 pixel to value.
3. unmanned aerial vehicle moving small target real-time detection method according to claim 2, which is characterized in that the movement
Target detection includes: background extracting, inter-frame difference, optical flow method, image segmentation, characteristic matching and machine learning.
4. a kind of moving small target real-time detection method based on unmanned aerial vehicle platform, based on being described in one of claim 1-3
System is realized, which comprises
The video image for obtaining varifocal camera carries out image down-sampled;
Moving object detection, several Small objects detected are carried out to down-sampled image;
Using picture centre as coordinate origin, position of the center of gravity relative to origin for the target that each is detected is calculated;
A target is selected, the coordinate bit in its real space relative to camera is calculated according to the position of the target in the picture
It sets, sends instruction adjustment UAV Attitude to unmanned aerial vehicle control system, so that camera is directed toward target, while by varifocal camera
Sending zoom instructions makes camera zoom, then obtains the target video image of camera shooting;
Target video image is inputted into trained deep neural network in advance, exports recognized targeted species and probability,
And the choosing of profile circle is carried out to target.
5. unmanned aerial vehicle moving small target real-time detection method according to claim 4, which is characterized in that described pair of drop
The image of sampling carries out moving object detection, several targets detected;It specifically includes:
Moving object detection is carried out to down-sampled image, obtains testing result;
Binary conversion treatment is carried out to testing result, the pixel of background position is 0, and the pixel value of moving target position is
1;
According to the number of pixels of detected minimum target and the size of false alarm rate setting corrosion parameter and expansion parameters, to institute
The Small object detected carries out corrosion and expansive working,
Clustering processing, every a kind of corresponding moving target are carried out for 1 pixel to value.
6. unmanned aerial vehicle moving small target real-time detection method according to claim 5, which is characterized in that the movement
Target detection includes: background extracting, inter-frame difference, optical flow method, image segmentation, characteristic matching and machine learning.
7. unmanned aerial vehicle moving small target real-time detection method according to claim 4, which is characterized in that the method
Further include: the training step of deep neural network specifically includes:
Under different location, different angle, different distance, the mesh for the target type that several to be identified is acquired using unmanned plane
Logo image data are labeled data collected, and marked content includes carrying out the choosing of profile circle, label target class to target
Not, training sample is generated;
The deep neural network for using SSD-MobileNet structure is established, wherein separating Standard convolution core in MobileNet
For depth convolution kernel and put convolution kernel;The port number of characteristic pattern is M, and convolution kernel size is DK*DK, output channel number is N, output
Characteristic pattern size is DF*DF, then Standard convolution core is M*DK*DK* N, the depth convolution kernel size after separation are DK*DK* M, point volume
Product core size is M*N*1*1;In SSD network, VGG structure is substituted for MobileNet structure;
Deep neural network is trained using training sample, the loss function of use includes Classification Loss and positioning loss two
Part obtains trained deep neural network.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910778929.XA CN110443247A (en) | 2019-08-22 | 2019-08-22 | A kind of unmanned aerial vehicle moving small target real-time detecting system and method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910778929.XA CN110443247A (en) | 2019-08-22 | 2019-08-22 | A kind of unmanned aerial vehicle moving small target real-time detecting system and method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110443247A true CN110443247A (en) | 2019-11-12 |
Family
ID=68437094
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910778929.XA Pending CN110443247A (en) | 2019-08-22 | 2019-08-22 | A kind of unmanned aerial vehicle moving small target real-time detecting system and method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110443247A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111562220A (en) * | 2020-06-02 | 2020-08-21 | 吉林大学 | Rapid and intelligent detection method for bridge diseases |
CN111626138A (en) * | 2020-04-30 | 2020-09-04 | 南京理工大学 | Multi-scale weak and small target real-time detection system and method under complex ground background |
CN111736190A (en) * | 2020-07-24 | 2020-10-02 | 广东电网有限责任公司 | Unmanned aerial vehicle airborne target detection system and method |
CN113312943A (en) * | 2020-02-27 | 2021-08-27 | 华为技术有限公司 | Video motion recognition method and device |
CN113438399A (en) * | 2021-06-25 | 2021-09-24 | 北京冠林威航科技有限公司 | Target guidance system, method for unmanned aerial vehicle, and storage medium |
CN113853781A (en) * | 2020-05-28 | 2021-12-28 | 深圳市大疆创新科技有限公司 | Image processing method, head-mounted display equipment and storage medium |
CN114200498A (en) * | 2022-02-16 | 2022-03-18 | 湖南天巡北斗产业安全技术研究院有限公司 | Satellite navigation/optical combined target detection method and system |
Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101144716A (en) * | 2007-10-15 | 2008-03-19 | 清华大学 | Multiple angle movement target detection, positioning and aligning method |
CN101930072A (en) * | 2010-07-28 | 2010-12-29 | 重庆大学 | Multi-feature fusion based infrared small dim moving target track starting method |
CN101969548A (en) * | 2010-10-15 | 2011-02-09 | 中国人民解放军国防科学技术大学 | Active video acquiring method and device based on binocular camera shooting |
CN103955912A (en) * | 2014-02-14 | 2014-07-30 | 西安电子科技大学 | Adaptive-window stomach CT image lymph node tracking detection system and method |
CN106707296A (en) * | 2017-01-09 | 2017-05-24 | 华中科技大学 | Dual-aperture photoelectric imaging system-based unmanned aerial vehicle detection and recognition method |
US20180050800A1 (en) * | 2016-05-09 | 2018-02-22 | Coban Technologies, Inc. | Systems, apparatuses and methods for unmanned aerial vehicle |
CN108012083A (en) * | 2017-12-14 | 2018-05-08 | 深圳云天励飞技术有限公司 | Face acquisition method, device and computer-readable recording medium |
CN108229442A (en) * | 2018-02-07 | 2018-06-29 | 西南科技大学 | Face fast and stable detection method in image sequence based on MS-KCF |
CN109788201A (en) * | 2019-02-14 | 2019-05-21 | 四川宏图智慧科技有限公司 | Localization method and device |
CN109815773A (en) * | 2017-11-21 | 2019-05-28 | 北京航空航天大学 | A kind of low slow small aircraft detection method of view-based access control model |
CN109872483A (en) * | 2019-02-22 | 2019-06-11 | 华中光电技术研究所(中国船舶重工集团有限公司第七一七研究所) | A kind of invasion warning photoelectric monitoring system and method |
CN109902578A (en) * | 2019-01-25 | 2019-06-18 | 南京理工大学 | A kind of infrared target detection and tracking |
CN110046619A (en) * | 2019-04-18 | 2019-07-23 | 马杰 | The full-automatic shoal of fish detection method of unmanned fish finding ship and system, unmanned fish finding ship and storage medium |
-
2019
- 2019-08-22 CN CN201910778929.XA patent/CN110443247A/en active Pending
Patent Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101144716A (en) * | 2007-10-15 | 2008-03-19 | 清华大学 | Multiple angle movement target detection, positioning and aligning method |
CN101930072A (en) * | 2010-07-28 | 2010-12-29 | 重庆大学 | Multi-feature fusion based infrared small dim moving target track starting method |
CN101969548A (en) * | 2010-10-15 | 2011-02-09 | 中国人民解放军国防科学技术大学 | Active video acquiring method and device based on binocular camera shooting |
CN103955912A (en) * | 2014-02-14 | 2014-07-30 | 西安电子科技大学 | Adaptive-window stomach CT image lymph node tracking detection system and method |
US20180050800A1 (en) * | 2016-05-09 | 2018-02-22 | Coban Technologies, Inc. | Systems, apparatuses and methods for unmanned aerial vehicle |
CN106707296A (en) * | 2017-01-09 | 2017-05-24 | 华中科技大学 | Dual-aperture photoelectric imaging system-based unmanned aerial vehicle detection and recognition method |
CN109815773A (en) * | 2017-11-21 | 2019-05-28 | 北京航空航天大学 | A kind of low slow small aircraft detection method of view-based access control model |
CN108012083A (en) * | 2017-12-14 | 2018-05-08 | 深圳云天励飞技术有限公司 | Face acquisition method, device and computer-readable recording medium |
CN108229442A (en) * | 2018-02-07 | 2018-06-29 | 西南科技大学 | Face fast and stable detection method in image sequence based on MS-KCF |
CN109902578A (en) * | 2019-01-25 | 2019-06-18 | 南京理工大学 | A kind of infrared target detection and tracking |
CN109788201A (en) * | 2019-02-14 | 2019-05-21 | 四川宏图智慧科技有限公司 | Localization method and device |
CN109872483A (en) * | 2019-02-22 | 2019-06-11 | 华中光电技术研究所(中国船舶重工集团有限公司第七一七研究所) | A kind of invasion warning photoelectric monitoring system and method |
CN110046619A (en) * | 2019-04-18 | 2019-07-23 | 马杰 | The full-automatic shoal of fish detection method of unmanned fish finding ship and system, unmanned fish finding ship and storage medium |
Non-Patent Citations (5)
Title |
---|
MINGFEI GAO 等: "Dynamic Zoom-in Network for Fast Object Detection in Large Images", 《2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION》 * |
严瑾: "CPU-GPU异构云计算环境下视频分析任务调度机制研究与应用", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
查宇飞 等著: "《视频目标跟踪方法》", 31 July 2015 * |
牛文龙: "基于高时相探测的运动点目标检测方法研究", 《中国博士学位论文全文数据库 工程科技Ⅱ辑》 * |
马鹏阁 等著: "《多脉冲激光雷达》", 31 December 2017 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113312943A (en) * | 2020-02-27 | 2021-08-27 | 华为技术有限公司 | Video motion recognition method and device |
CN111626138A (en) * | 2020-04-30 | 2020-09-04 | 南京理工大学 | Multi-scale weak and small target real-time detection system and method under complex ground background |
CN113853781A (en) * | 2020-05-28 | 2021-12-28 | 深圳市大疆创新科技有限公司 | Image processing method, head-mounted display equipment and storage medium |
CN111562220A (en) * | 2020-06-02 | 2020-08-21 | 吉林大学 | Rapid and intelligent detection method for bridge diseases |
CN111736190A (en) * | 2020-07-24 | 2020-10-02 | 广东电网有限责任公司 | Unmanned aerial vehicle airborne target detection system and method |
CN111736190B (en) * | 2020-07-24 | 2022-01-25 | 广东电网有限责任公司 | Unmanned aerial vehicle airborne target detection system and method |
CN113438399A (en) * | 2021-06-25 | 2021-09-24 | 北京冠林威航科技有限公司 | Target guidance system, method for unmanned aerial vehicle, and storage medium |
CN113438399B (en) * | 2021-06-25 | 2022-04-08 | 北京冠林威航科技有限公司 | Target guidance system, method for unmanned aerial vehicle, and storage medium |
CN114200498A (en) * | 2022-02-16 | 2022-03-18 | 湖南天巡北斗产业安全技术研究院有限公司 | Satellite navigation/optical combined target detection method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110443247A (en) | A kind of unmanned aerial vehicle moving small target real-time detecting system and method | |
EP3440428B1 (en) | Remote determination of quantity stored in containers in geographical region | |
Zheng et al. | Air-to-air visual detection of micro-uavs: An experimental evaluation of deep learning | |
US11205274B2 (en) | High-performance visual object tracking for embedded vision systems | |
CN109903331B (en) | Convolutional neural network target detection method based on RGB-D camera | |
Ashraf et al. | Dogfight: Detecting drones from drones videos | |
Zhou et al. | Robust vehicle detection in aerial images using bag-of-words and orientation aware scanning | |
US7957557B2 (en) | Tracking apparatus and tracking method | |
EP2874097A2 (en) | Automatic scene parsing | |
WO2013088175A1 (en) | Image processing method | |
CN107851318A (en) | System and method for Object tracking | |
Štěpán et al. | Vision techniques for on‐board detection, following, and mapping of moving targets | |
CN108171715B (en) | Image segmentation method and device | |
Wang et al. | An overview of 3d object detection | |
CN110992378B (en) | Dynamic updating vision tracking aerial photographing method and system based on rotor flying robot | |
Baykara et al. | Real-time detection, tracking and classification of multiple moving objects in UAV videos | |
CN106250898B (en) | A kind of image local area feature extracting method based on scale prediction | |
CN112683228A (en) | Monocular camera ranging method and device | |
CN116109950A (en) | Low-airspace anti-unmanned aerial vehicle visual detection, identification and tracking method | |
Fang et al. | A real-time anti-distractor infrared UAV tracker with channel feature refinement module | |
CN115018886B (en) | Motion trajectory identification method, device, equipment and medium | |
CN116862832A (en) | Three-dimensional live-action model-based operator positioning method | |
CN115861352A (en) | Monocular vision, IMU and laser radar data fusion and edge extraction method | |
Muddamsetty et al. | Spatio-temporal saliency detection in dynamic scenes using local binary patterns | |
Song et al. | Object tracking with dual field-of-view switching in aerial videos |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20191112 |
|
WD01 | Invention patent application deemed withdrawn after publication |