CN110110115A - The method that human-computer interaction screening identifies airborne target image - Google Patents
The method that human-computer interaction screening identifies airborne target image Download PDFInfo
- Publication number
- CN110110115A CN110110115A CN201910245235.XA CN201910245235A CN110110115A CN 110110115 A CN110110115 A CN 110110115A CN 201910245235 A CN201910245235 A CN 201910245235A CN 110110115 A CN110110115 A CN 110110115A
- Authority
- CN
- China
- Prior art keywords
- identification
- image
- human
- computer interaction
- aircraft
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/5866—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using information manually generated, e.g. tags, keywords, comments, manually generated location and time information
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Library & Information Science (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Collating Specific Patterns (AREA)
Abstract
A kind of method that human-computer interaction screening identifies airborne target image disclosed by the invention, it is intended to propose that a kind of calculation amount is small, the method for the high airborne target image of identification of recognition accuracy.The technical scheme is that: the image recognition database of aircraft signature is arranged and formed to the experience of Aircraft Target Identification, knowledge first；Matching identification based on fingerprint characteristic, successively carry out special character on aircraft, special pattern human-computer interaction matching identification fingerprint characteristic, and identification is screened to the human-computer interaction of visual properties on aircraft, geometrical characteristic；Between the matching identification based on fingerprint characteristic and the identification of the screening based on visual properties and geometrical characteristic, judge that matching identification and screening identify whether success, if success, directly export recognition result, otherwise, the similarity matching identification that aircraft print and image recognition database print are carried out by print similarity algorithm assists completing target identification according to similarity degree.
Description
Technical field
The present invention relates to one kind to be applied to images steganalysis field, carries out human-computer interaction knowledge to the Aircraft Targets in image
Method for distinguishing.
Background technique
Target identification is a basic project of field of image processing and field of machine vision.Studies have shown that different mesh
When mark carries out target detection, target identification, target classification, goal description, there is different requirements to image resolution ratio.With airborne
Imaging resolution is higher and higher, to identify target, and the data volume of collection is big considerably beyond the limit manually judged rapidly,
Oneself is difficult to complete so big workload for traditional artificial interpretation, and the subjective errors and misinterpretation of artificial interpretation are thus not
It can avoid, these experiences due to not being organized into knowledge and forming systematic knowledge base by this identification method, to airborne behaviour
The competency profiling for making personnel is high, and recognition performance is also affected with the operating pressure of airborne operator and state.In addition, due to
The presence of SAR special imaging mechanism and speckle noise, SAR image can be intuitively understood unlike optical imagery, image
The quality of performance directly influences subsequent identification work.The slight fluctuations of SAR imaging parameters, such as depression angle, azimuth of target
And its variation of configuration, it can all cause the acute variation of characteristics of image.The complexity of SAR image brings automatic target detection ATR
The complexity of system, calculation amount are very big.The presence of noise largely reduces the quality of image, mould in SAR image
The characteristic information for having pasted area-of-interest affects the application such as subsequent target identification.
Onboard image automatic target detection always exists that recognition accuracy is low, wants to the quality and quantity of image training sample
Ask the problems such as high.Since SAR image has very strong speckle noise, the mutability of characteristics of image, is carried out using SAR image in addition
Automatic target detection is a relatively difficult job.Since the target identification of airborne imaging radar is mostly high altitude operation, identification
Target image information vulnerable to high-altitude external noise, the shake of fuselage, the shake of testee and sample rate it is excessively low because
The interference of element can recognize that deep fades occur for feature, so that the identification image resolvability of target is deteriorated.The target of automation is known
Other algorithm can use template matching, characteristic matching and advanced machine learning, deep learning algorithm etc., these methods are usual
There is recognition accuracies it is low, computationally intensive, high to quality and the quantitative requirement of image training sample the disadvantages of, distance is practical
There is certain gap.Template matching method using sample average as template, but template number with the diminution of azimuthal separation and
Increase, preferable Classification and Identification rate is exchanged for biggish memory space.Machine learning method is also very much, as Kottke D.P is used
Hidden Markov model models, and characteristic sequence is by every a kind of region (background, four class of shade, shallow target area and bright target area)
Each angle Radon transformed value is end to end to be formed, and preferable target identification effect is obtained.
Summary of the invention
The purpose of the present invention is the recognition accuracy for aircraft automatic target detection is low, made Target is identified by airborne behaviour
Make the peopleware of personnel, working condition influences the problems such as big, proposes that a kind of calculation amount is small, recognition accuracy is high, is suitable for machine
Carry the method that the human-computer interaction screening of operator identifies airborne target image.
Above-mentioned purpose of the invention can be obtained by following measures, a kind of human-computer interaction screening identification airborne target figure
The method of picture has following technical characteristic: being arranged first to the experience of Aircraft Target Identification, knowledge and form aircraft signature
Image recognition database；Matching identification based on fingerprint characteristic, successively carry out aircraft on special character, special pattern it is man-machine
The fingerprint characteristic of interaction matching identification, and identification is screened to the human-computer interaction of visual properties, geometrical characteristic on aircraft；Based on finger
The matching identification of line feature and based on visual properties and geometrical characteristic screening identification between, judge matching identification and screening identification
Whether succeed, if it is successful, directly exporting recognition result, otherwise, aircraft print and image is carried out by print similarity algorithm
The similarity matching identification of identification database print assists completing target identification according to similarity degree.
The present invention has following technical effect that
(1) calculation amount is small.The present invention is directed to the advantage and disadvantage of the above-mentioned prior art, proposes a kind of aircraft suitable for airborne operator
The man-machine interactive identification method of target, by the experience of Aircraft Target Identification, knowledge, special character, pattern such as aircraft, aircraft can
Depending on feature and geometries characteristic, the image recognition knowledge base for being able to ascend Aircraft Target Identification speed is established, is realizing it
It is upper to be bound tightly together with image recognition knowledge storehouse matching, reach and operand is greatly reduced, improves recognition speed, subtract
Recognition performance is promoted while the operating pressure of few airborne operator, on this basis, selects recognition performance and calculation amount suitable
In image recognition processing algorithm auxiliary complete target identification.Quality and working condition shadow of the recognition performance by airborne operator
Sound is small.By the experience of Aircraft Target Identification, knowledge, it is systematically organized into identification knowledge base, it is low to the requirement of image training sample,
It assists airborne operator to work by machine, alleviates peopleware of the recognition performance by airborne operator, working condition
Influence.
(2) recognition accuracy is high.The present invention first arranges the experience of Aircraft Target Identification, knowledge and forms image
Then identification database successively carries out the human-computer interaction matching identification of the fingerprint characteristic such as special character, special pattern on aircraft,
Carry out visual properties on aircraft, identification is screened in the human-computer interaction of geometrical characteristic, progress aircraft print and image data base print
Similarity matching identification is screened by the database of multiple levels, eliminates numerous jamming targets, describe mesh using feature
Mark reduces the information content of target, removes redundancy, improves recognition speed and precision.Eventually by feature extraction, effectively obtain
It can be used for each category feature of target detection, identification, classification in image, so that the efficiency and performance of target identification are improved,
Recognition result is exported according to similarity degree, mitigates recognition performance while promoting the accuracy rate of Aircraft Target Identification by airborne operation
The influence of the quality and working condition of personnel.On this basis, then by the measurement of aircraft geometric dimension, print similarity than equity
Automation algorithm assists in identifying target out, greatly improves recognition accuracy.
Detailed description of the invention
Fig. 1 is the flow chart that human-computer interaction screening of the present invention identifies airborne target image.
Fig. 2 is human-computer interaction matching identification flow chart of the Fig. 1 based on fingerprint characteristic.
Fig. 3 is human-computer interaction screening identification process figure of the Fig. 1 based on visual properties and geometrical characteristic.
Fig. 4 is human-computer interaction similarity identification flow chart of the Fig. 1 based on image print.
Below by specific embodiment and in conjunction with attached drawing, the present invention is described in further detail.
Specific embodiment
According to the present invention, refering to fig. 1, human-computer interaction screening identifies that airborne target image is divided into four steps: (1) establishing and fly
Machine target identification data library.(2) the human-computer interaction matching identification based on fingerprint characteristic；(3) visual properties and geometrical characteristic are based on
Human-computer interaction screen identification；(4) based on the human-computer interaction similarity identification of image print.Firstly, to Aircraft Target Identification
Experience, knowledge are arranged and form the image recognition database of aircraft signature；Then, based on the matching identification of fingerprint characteristic,
Successively carry out aircraft on special character, special pattern human-computer interaction matching identification fingerprint characteristic；Then, carrying out can on aircraft
Identification is screened in human-computer interaction depending on feature, geometrical characteristic；Based on fingerprint characteristic matching identification with based on visual properties and several
Between the screening identification of what feature, judge that matching identification and screening identify whether success, if it is successful, directly output identification knot
Otherwise fruit carries out the similarity matching identification of aircraft print and image recognition database print, root by print similarity algorithm
It assists completing target identification according to similarity degree.
(1) image recognition knowledge base is established.According to the experience of Aircraft Target Identification, knowledge, image recognition database is established,
Image recognition database includes the day of the machine side of a ship number, number of registration, aircraft identification code, radio call, machine emblem, disk aerial, protuberance
Irdome, wing, fuselage, tailplane, vertical fin, the number of engine, installation site, fundamental type, flat shape, fuselage, machine
Length, perimeter, area, length-width ratio, the angle of the wing, aircraft print etc..Image recognition database includes fingerprint characteristic, visual
Feature, geometrical characteristic, aircraft print etc., fingerprint characteristic refer to represent the special character or Special Graphs of aircraft identity attribute feature
Case, special character include the machine side of a ship number, number of registration, aircraft identification code, radio call etc., and special pattern includes machine emblem, disk day
Line, antenna house of protuberance etc., visual properties include wing, fuselage, tailplane, vertical fin, the number of engine, installation site,
Fundamental type, flat shape etc., geometrical characteristic include fuselage, the length of wing, perimeter, area, length-width ratio, angle etc..
(2) referring to Fig.2, in the human-computer interaction matching identification based on fingerprint characteristic, using experience, knowledge to reconnaissance image
In special character, special pattern carry out visual identification, identification feature is manually entered, and using visual identification result as database
Querying condition carries out inquiry matching identification in image recognition database, requires if meeting identification, ties this as identification
Fruit exports comparison result, otherwise enters next step identification process.
(3) experience, knowledge are utilized in the human-computer interaction screening identification based on visual properties and geometrical characteristic refering to Fig. 3
Visual identification is carried out to the visual properties in reconnaissance image, carries out figure using geometries characteristic of the image measurement tool to target
As measurement, image measurement tool selects visual properties according to recognition result, determines that geometrical characteristic parameter is gone forward side by side according to measurement result
Row image recognition data base querying calls the common dimensional measurement algorithm based on image and carries out operation, obtains geometrical characteristic ginseng
Several and its precision, and using the combination of visual identification result, geometries characteristic parameter or said two devices as image recognition data
Library inquiry condition carries out screening identification in image recognition database, requires if meeting identification, using this as recognition result,
Query result is exported, next step identification process is otherwise entered.
(4) refering to Fig. 4, in the human-computer interaction similarity identification based on image print, tool is cut to mesh using image
Mark carries out image print and cuts, and obtains target image print, the image recognition database after reading screening, and utilizes the part SIFT
Characteristic point detective operators, detection target print and the local feature in multiple prints in image recognition database query result
Point, if the SIFT feature number extracted from target print X is m, various kinds picture Y (i), (i in image recognition database
=1,2,3 ... K) the SIFT feature number extracted is respectively n (i) (i=1,2,3 ... K), m SIFT in target print X spy
The number that sign point is matched with a SIFT feature of n (i) in the various kinds picture Y (i) in database is respectively k (i) (i=
1,2,3 ... K), k (i) is normalized, final similarity value is obtained, finally, the height according to similarity degree exports
Recognition result.SIFT local feature region detective operators are by asking the characteristic point in a width figure and its description in relation to scale and direction
Son obtains feature and carries out Image Feature Point Matching, obtains good result, and SIFT feature not only has scale invariability, even if
Change rotation angle, brightness of image or shooting visual angle, the feature detection effect still being able to.
Above in conjunction with attached drawing to the present invention have been described in detail, it is to be noted that being described in examples detailed above
Preferred embodiment only of the invention, is not intended to restrict the invention, and for those skilled in the art, the present invention can
To there is various modifications and variations, all within the spirits and principles of the present invention, any modification, equivalent replacement, improvement and so on,
It should be included within scope of the presently claimed invention.
Claims (10)
1. a kind of method that human-computer interaction screening identifies airborne target image, has following technical characteristic: first to Aircraft Targets
Experience, the knowledge of identification are arranged and form the image recognition database of aircraft signature；Matching identification based on fingerprint characteristic,
Successively carry out special character on aircraft, special pattern human-computer interaction matching identification fingerprint characteristic, and to visual special on aircraft
Identification is screened in the human-computer interaction of sign, geometrical characteristic；In the matching identification based on fingerprint characteristic and based on visual properties and geometry spy
Between the screening identification of sign, judge that matching identification and screening identify whether success, if it is successful, recognition result is directly exported, it is no
Then, the similarity matching identification that aircraft print and image recognition database print are carried out by print similarity algorithm, according to phase
It assists completing target identification like degree.
2. the method that human-computer interaction screening as described in claim 1 identifies airborne target image, it is characterised in that: image recognition
Database includes the antenna house of the machine side of a ship number, number of registration, aircraft identification code, radio call, machine emblem, disk aerial, protuberance, machine
The wing, fuselage, tailplane, vertical fin, the number of engine, installation site, fundamental type, flat shape, fuselage, wing length,
Width, height, perimeter, area, length-width ratio, angle, aircraft print, fingerprint characteristic, visual properties, geometrical characteristic and aircraft print.
3. the method that human-computer interaction screening as claimed in claim 2 identifies airborne target image, it is characterised in that: fingerprint characteristic
Refer to represent the special character or special pattern of aircraft identity attribute feature.
4. the method that human-computer interaction screening as claimed in claim 3 identifies airborne target image, it is characterised in that: special character
Including the machine side of a ship number, number of registration, aircraft identification code and radio call.
5. the method that human-computer interaction screening as claimed in claim 3 identifies airborne target image, it is characterised in that: special pattern
Including machine emblem, disk aerial, protuberance antenna house.
6. the method that human-computer interaction screening as claimed in claim 2 identifies airborne target image, it is characterised in that: visual properties
Including wing, fuselage, tailplane, vertical fin, the number of engine, installation site, fundamental type, flat shape.
7. the method that human-computer interaction screening as claimed in claim 2 identifies airborne target image, it is characterised in that: geometrical characteristic
Including fuselage, the length of wing, perimeter, area, length-width ratio and angle.
8. the method that human-computer interaction screening as described in claim 1 identifies airborne target image, it is characterised in that: based on finger
In the human-computer interaction matching identification of line feature, mesh is carried out to special character, the special pattern in reconnaissance image using experience, knowledge
Depending on identification, identification feature is manually entered, and using visual identification result as database query, in image recognition database
Inquiry matching identification is carried out, is required if meeting identification, using this as recognition result, exports comparison result, is otherwise entered next
Walk identification process.
9. the method that human-computer interaction as described in claim 1 screening identifies airborne target image, it is characterised in that: based on can
Depending on being carried out using experience, knowledge to the visual properties in reconnaissance image in the human-computer interaction screening identification of feature and geometrical characteristic
Visual identification carries out image measurement using geometries characteristic of the image measurement tool to target, and image measurement tool is according to knowledge
Other result selects visual properties, determines geometrical characteristic parameter according to measurement result and carries out image recognition data base querying, calls
Common dimensional measurement algorithm based on image simultaneously carries out operation, obtains geometrical characteristic parameter and its precision, and by visual identification knot
The combination of fruit, geometries characteristic parameter or said two devices is as image recognition database query, in image recognition data
Screening identification is carried out in library, is required if meeting identification, using this as recognition result, exports query result, is otherwise entered next
Walk identification process.
10. the method that human-computer interaction screening as described in claim 1 identifies airborne target image, it is characterised in that: be based on
In the human-computer interaction similarity identification of image print, tool is cut using image, target progress image print is cut, obtain mesh
Logo image print, the image recognition database after reading screening, and SIFT local feature region detective operators are utilized, detect target sample
The local feature region in multiple prints in piece and image recognition database query result, and extracted from target print X
SIFT feature number is m, the various kinds picture Y (i) in image recognition database, and the SIFT that (i=1,2,3 ... K) are extracted is special
Sign point number is respectively n (i) (i=1,2,3 ... K), m SIFT feature in target print X and each print in database
The number that a SIFT feature of n (i) in image Y (i) matches is respectively k (i) (i=1,2,3 ... K), is returned to k (i)
One change processing, obtains final similarity value, finally, the height according to similarity degree exports recognition result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910245235.XA CN110110115B (en) | 2019-03-28 | 2019-03-28 | Method for screening and identifying airborne target image through man-machine interaction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910245235.XA CN110110115B (en) | 2019-03-28 | 2019-03-28 | Method for screening and identifying airborne target image through man-machine interaction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110110115A true CN110110115A (en) | 2019-08-09 |
| CN110110115B CN110110115B (en) | 2023-05-23 |
Family
ID=67484776
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910245235.XA Active CN110110115B (en) | 2019-03-28 | 2019-03-28 | Method for screening and identifying airborne target image through man-machine interaction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110110115B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101464951A (en) * | 2007-12-21 | 2009-06-24 | 北大方正集团有限公司 | Image recognition method and system |
| US20130330008A1 (en) * | 2011-09-24 | 2013-12-12 | Lotfi A. Zadeh | Methods and Systems for Applications for Z-numbers |
| WO2016015547A1 (en) * | 2014-08-01 | 2016-02-04 | 深圳中集天达空港设备有限公司 | Machine vision-based method and system for aircraft docking guidance and aircraft type identification |
| CN105913241A (en) * | 2016-04-01 | 2016-08-31 | 袁艳荣 | Application method of customer authentication system based on image identification |
| CN107690659A (en) * | 2016-12-27 | 2018-02-13 | 深圳前海达闼云端智能科技有限公司 | A kind of image identification system and image-recognizing method |
| WO2018212494A1 (en) * | 2017-05-17 | 2018-11-22 | 삼성전자 주식회사 | Method and device for identifying object |
| CN109002744A (en) * | 2017-06-06 | 2018-12-14 | 中兴通讯股份有限公司 | Image-recognizing method, device and video monitoring equipment |
| CN109117808A (en) * | 2018-08-24 | 2019-01-01 | 深圳前海达闼云端智能科技有限公司 | Face recognition method and device, electronic equipment and computer readable medium |
-
2019
- 2019-03-28 CN CN201910245235.XA patent/CN110110115B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101464951A (en) * | 2007-12-21 | 2009-06-24 | 北大方正集团有限公司 | Image recognition method and system |
| US20130330008A1 (en) * | 2011-09-24 | 2013-12-12 | Lotfi A. Zadeh | Methods and Systems for Applications for Z-numbers |
| WO2016015547A1 (en) * | 2014-08-01 | 2016-02-04 | 深圳中集天达空港设备有限公司 | Machine vision-based method and system for aircraft docking guidance and aircraft type identification |
| CN105913241A (en) * | 2016-04-01 | 2016-08-31 | 袁艳荣 | Application method of customer authentication system based on image identification |
| CN107690659A (en) * | 2016-12-27 | 2018-02-13 | 深圳前海达闼云端智能科技有限公司 | A kind of image identification system and image-recognizing method |
| WO2018212494A1 (en) * | 2017-05-17 | 2018-11-22 | 삼성전자 주식회사 | Method and device for identifying object |
| CN109002744A (en) * | 2017-06-06 | 2018-12-14 | 中兴通讯股份有限公司 | Image-recognizing method, device and video monitoring equipment |
| CN109117808A (en) * | 2018-08-24 | 2019-01-01 | 深圳前海达闼云端智能科技有限公司 | Face recognition method and device, electronic equipment and computer readable medium |
Non-Patent Citations (3)
| Title |
|---|
| S.SCLAROFF等: "Modal matching for correspondence and recognition", 《IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE》 * |
| 居恒哲: "面向对象的高分辨率遥感影像中飞机的识别方法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 方晓峰: "基于图像处理的LCD航空仪表信息的识别方法与实现", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110110115B (en) | 2023-05-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108510467B (en) | SAR image target identification method based on depth deformable convolution neural network | |
| CN109061600B (en) | Target identification method based on millimeter wave radar data | |
| CN109344825A (en) | A kind of licence plate recognition method based on convolutional neural networks | |
| CN105975913B (en) | Road network extraction method based on adaptive cluster learning | |
| WO2021129105A1 (en) | Mask rcnn network model-based target identification method and apparatus | |
| CN107085708B (en) | High-resolution remote sensing image change detection method based on multi-scale segmentation and fusion | |
| CN108241870A (en) | For distributing specific class method for distinguishing interested in measurement data | |
| CN109614936B (en) | Layered identification method for remote sensing image airplane target | |
| CN104573744B (en) | Fine granulation classification identifies and the part of object positions and feature extracting method | |
| CN109271991A (en) | A kind of detection method of license plate based on deep learning | |
| CN103577838A (en) | Face recognition method and device | |
| CN110598690A (en) | End-to-end optical character detection and identification method and system | |
| US20240005148A1 (en) | System and method for finding and classifying patterns in an image with a vision system | |
| CN104751166A (en) | Spectral angle and Euclidean distance based remote-sensing image classification method | |
| CN108109139A (en) | Airborne LIDAR three-dimensional building object detecting method based on gray scale volume element model | |
| CN100592334C (en) | Method for processing image data | |
| CN112907520A (en) | Single tree crown detection method based on end-to-end deep learning method | |
| CN109117776A (en) | Aircraft and meteorological clutter classifying identification method based on track information | |
| CN110599463A (en) | Tongue image detection and positioning algorithm based on lightweight cascade neural network | |
| CN111368682A (en) | Method and system for detecting and identifying station caption based on faster RCNN | |
| CN104050460B (en) | The pedestrian detection method of multiple features fusion | |
| CN108537169A (en) | A kind of high-resolution remote sensing image method for extracting roads based on center line and detection algorithm of having a lot of social connections | |
| Neuhausen et al. | Automatic window detection in facade images | |
| CN110659374A (en) | Method for searching images by images based on neural network extraction of vehicle characteristic values and attributes | |
| CN110852207A (en) | Blue roof building extraction method based on object-oriented image classification technology |
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 |