CN104077592A - Automatic extraction method for high-resolution remote-sensing image navigation mark - Google Patents
Automatic extraction method for high-resolution remote-sensing image navigation mark Download PDFInfo
- Publication number
- CN104077592A CN104077592A CN201310101493.3A CN201310101493A CN104077592A CN 104077592 A CN104077592 A CN 104077592A CN 201310101493 A CN201310101493 A CN 201310101493A CN 104077592 A CN104077592 A CN 104077592A
- Authority
- CN
- China
- Prior art keywords
- navigation mark
- doubtful
- sensing image
- navigation
- mark
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses an automatic extraction method for a high-resolution remote-sensing image navigation mark. The automatic extraction method for the high-resolution remote-sensing image navigation mark comprises the following steps: 1) firstly, selecting the pixel of a water area as a training sample, and training a one-class support vector machine classifier; 2) predicting all pixels of a remote-sensing image by the trained one-class support vector machine classifier; 3) searching other targets which are not water area in the pixel of the water area; 4) calculating the areas of all detected cavities; 5) calculating the gray level mean value of a peripheral area in a window where the cavities are positioned; 6) carrying out relevant marshaling to reserved targets; 7) on the basis of an on-line principle of learning, detecting an omission navigation mark; 8) searching on the basis of the image multiband gray level characteristics of the detected navigation mark; and 9) updating a curve fitting result and a navigation mark template. According to the automatic extraction method for the high-resolution remote-sensing image navigation mark, which is disclosed by the invention, the high-resolution remote-sensing image navigation mark can be fully automatically extracted without man-machine interaction.
Description
Technical field
The present invention relates to a kind of image navigation mark extraction method, particularly relate to a kind of high-resolution remote sensing image navigation mark extraction method.
Background technology
Along with the progress of remote sensing technology, the improving constantly of remote sensing image resolution, remote sensing all will have broad application prospects in all trades and professions.In maritime affairs and shipping management, remote sensing and AIS(Automatic Identification System, ship automatic identification system), GIS(Geographic Information System, Geographic Information System) combination is day by day tight, need to generate high precision orthography be combined for data analysis with AIS and GIS in waters, navigation channel.But prior art needs people to detect navigation mark information for driving, and has the situation of undetected navigation mark.
Summary of the invention
Technical matters to be solved by this invention is to provide a kind of high-resolution remote sensing image navigation mark extraction method, its method based on one-class support vector machines and related coefficient marshalling can detect most of navigation mark information automatically, and can complete automatically undetected navigation mark detects, whole process does not need interpersonal mutual, automatically completes.
The present invention solves above-mentioned technical matters by following technical proposals: a kind of high-resolution remote sensing image navigation mark extraction method, it is characterized in that, and it comprises the following steps:
Step 1, for remote sensing image, first chooses the pixel in waters as training sample, training one-class support vector machines sorter;
Step 2, utilizes the one-class support vector machines sorter training to predict all pixels of remote sensing image, retains all predicting the outcome as the pixel in waters;
Step 3, in the pixel in waters, search is not other targets in waters, the candidate target that cavity in Water on Remote Sensing Image is detected as navigation mark;
Step 4, calculates the area in all detected cavities, rejects the target of the less and obvious non-navigation mark that area is larger of area;
Step 5, for empty area and the approaching target of navigation mark, calculate the gray average of outer peripheral areas in the window of empty place, obtain the difference of difference, average and the minimum value of maximal value and average, if the difference of the difference of maximal value and average, average and minimum value is all greater than given empirical value, target is retained, otherwise by its eliminating;
Step 6, then to the target the retaining marshalling of being correlated with;
Step 7, then based on on-line study principle, undetected navigation mark is detected, the space distribution of the navigation mark that first foundation obtains after testing, utilize cubic polynomial to estimate the possible position of undetected navigation mark, more accurately detect in estimated position according to the priori of the navigation mark having detected;
Step 8, to there being undetected region, the image multiband gamma characteristic based on detecting navigation mark is searched for; In order to search for, to detecting navigation mark, according to pixels get average, build and detect template; In the doubtful undetected position of estimating, carry out match search, if there is window to there is the related coefficient that meets threshold value at each wave band, this window is retained as navigation mark;
Step 9, joins the undetected navigation mark detecting in known navigation mark, upgrades curve-fitting results and navigation mark template, carries out iterative detection until do not have new doubtful undetected position to occur.
Preferably, described one-class support vector machines sorter is binary classifier, can make a prediction to pixel, predicts the outcome and comprises waters and other.
Preferably, described step 6 specifically comprises the following steps:
Step 6 A, traversal candidate target, rejects with other targets and does not have correlativity target, obtains doubtful navigation mark sequence;
Step 6 B, travels through doubtful navigation mark, and the navigation mark not traveling through is set up to new relevant group, and this doubtful navigation mark of mark travels through, searches its all similar doubtful navigation marks, and adds in new group;
Step 6 C, travels through doubtful navigation mark in new group, and this doubtful navigation mark of mark travels through, and the doubtful navigation mark similar to doubtful boat not occurring in new group added in new group;
Step 6 D, in new group the similar navigation mark of all doubtful navigation marks all in new group, the relevant group of output; Otherwise repeating step six C;
Step 6 E, all navigation marks are all traversed, export all relevant group; Otherwise repeating step six B, step 6 C, step 6 D;
Step 6 F, using maximal correlation group as navigation mark group.
Positive progressive effect of the present invention is: the method that the present invention is based on one-class support vector machines and related coefficient marshalling can detect most of navigation mark information automatically.In addition, the undetected navigation mark extracting method that the present invention is based on curve can complete automatically undetected navigation mark and detect, and whole process does not need interpersonal mutual, automatically completes.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of high-resolution remote sensing image navigation mark extraction method of the present invention.
Embodiment
Below in conjunction with accompanying drawing, provide preferred embodiment of the present invention, to describe technical scheme of the present invention in detail.
Navigation mark, generally with real time GPS (Global Positioning System, GPS) information, obtains image coordinate after accurately extracting, and in conjunction with GPS information, can be used as reference mark for the orthophotoquad generation in water body region.In addition, remote sensing image wide coverage, contains much information, and the supplementary means that can be used as GPS equipment is used for the navigation mark state in region, navigation channel on a large scale to monitor.Because remote sensing satellite can complete return visit to ground point a few days or tens of days, therefore the same area has the remote sensing image of a large amount of different periods, navigation mark by different period remote sensing images extracts, the alteration in the navigation channel that can obtain the change in location of navigation mark and be extrapolated by navigation mark, for the digitizing in navigation mark navigation channel provides Data support.
As shown in Figure 1, high-resolution remote sensing image navigation mark extraction method of the present invention comprises the following steps:
Step 1, for remote sensing image, first chooses the pixel in waters as training sample, training one-class support vector machines sorter.One-class support vector machines sorter is binary classifier, can make a prediction to pixel, predicts the outcome and comprises waters and other.
Step 2, utilizes the one-class support vector machines sorter training to predict all pixels of remote sensing image, retains all predicting the outcome as the pixel in waters.
Step 3, in the pixel in waters, search is not other targets in waters, the candidate target that cavity in Water on Remote Sensing Image is detected as navigation mark.
Step 4, calculates the area in all detected cavities, rejects the target of the less and obvious non-navigation mark that area is larger of area.
Step 5, for empty area and the approaching target of navigation mark, calculate the gray average of outer peripheral areas in the window of empty place, obtain the difference D1 of maximal value and average and the difference D2 of average and minimum value, if D1 and D2 are greater than given empirical value, target is retained, otherwise by its eliminating.
Step 6, then according to following flow process to the target the retaining marshalling of being correlated with:
Step 6 A, traversal candidate target, rejects with other targets and does not have correlativity target, obtains doubtful navigation mark sequence;
Step 6 B, travels through doubtful navigation mark, and the navigation mark not traveling through is set up to new relevant group, and this doubtful navigation mark of mark travels through, searches its all similar doubtful navigation marks, and adds in new group;
Step 6 C, travels through doubtful navigation mark in new group, and this doubtful navigation mark of mark travels through, and the doubtful navigation mark similar to doubtful boat not occurring in new group added in new group;
Step 6 D, in new group the similar navigation mark of all doubtful navigation marks all in new group, the relevant group of output.Otherwise repeating step six C;
Step 6 E, all navigation marks are all traversed, export all relevant group.Otherwise repeating step six B, step 6 C, step 6 D;
Step 6 F, using maximal correlation group as navigation mark group.
Step 7, then based on on-line study principle, undetected navigation mark is detected, the space distribution of the navigation mark that first foundation obtains after testing, utilize cubic polynomial to estimate the possible position of undetected navigation mark, more accurately detect in estimated position according to the priori of the navigation mark having detected.
Step 8, to there being undetected region, the image multiband gamma characteristic based on detecting navigation mark is searched for.In order to search for, to detecting navigation mark, according to pixels get average, build and detect template.In the doubtful undetected position of estimating, carry out match search, if there is window to there is the related coefficient that meets threshold value at each wave band, this window is retained as navigation mark.
Step 9, joins the undetected navigation mark detecting in known navigation mark, upgrades curve-fitting results and navigation mark template, carries out iterative detection until do not have new doubtful undetected position to occur.
The method that the present invention is based on one-class support vector machines and related coefficient marshalling can detect most of navigation mark information automatically.In addition, the undetected navigation mark extracting method that the present invention is based on curve can complete automatically undetected navigation mark and detect, and whole process does not need interpersonal mutual, automatically completes.
Those skilled in the art can carry out various remodeling and change to the present invention.Therefore, the present invention has covered various remodeling and the change in the scope that falls into appending claims and equivalent thereof.
Claims (3)
1. a high-resolution remote sensing image navigation mark extraction method, is characterized in that, it comprises the following steps:
Step 1, for remote sensing image, first chooses the pixel in waters as training sample, training one-class support vector machines sorter;
Step 2, utilizes the one-class support vector machines sorter training to predict all pixels of remote sensing image, retains all predicting the outcome as the pixel in waters;
Step 3, in the pixel in waters, search is not other targets in waters, the candidate target that cavity in Water on Remote Sensing Image is detected as navigation mark;
Step 4, calculates the area in all detected cavities, rejects the target of the less and obvious non-navigation mark that area is larger of area;
Step 5, for empty area and the approaching target of navigation mark, calculate the gray average of outer peripheral areas in the window of empty place, obtain the difference of difference, average and the minimum value of maximal value and average, as being all greater than given empirical value, the difference of the difference of maximal value and average, average and minimum value target is retained, otherwise by its eliminating;
Step 6, then to the target the retaining marshalling of being correlated with:
Step 7, then based on on-line study principle, undetected navigation mark is detected, the space distribution of the navigation mark that first foundation obtains after testing, utilize cubic polynomial to estimate the possible position of undetected navigation mark, more accurately detect in estimated position according to the priori of the navigation mark having detected;
Step 8, to there being undetected region, the image multiband gamma characteristic based on detecting navigation mark is searched for; In order to search for, to detecting navigation mark, according to pixels get average, build and detect template; In the doubtful undetected position of estimating, carry out match search, if there is window to there is the related coefficient that meets threshold value at each wave band, this window is retained as navigation mark;
Step 9, joins the undetected navigation mark detecting in known navigation mark, upgrades curve-fitting results and navigation mark template, carries out iterative detection until do not have new doubtful undetected position to occur.
2. high-resolution remote sensing image navigation mark extraction method as claimed in claim 1, is characterized in that, described one-class support vector machines sorter is binary classifier, can make a prediction to pixel, predicts the outcome and comprises waters and other.
3. high-resolution remote sensing image navigation mark extraction method as claimed in claim 1, is characterized in that, described step 6 specifically comprises the following steps:
Step 6 A, traversal candidate target, rejects with other targets and does not have correlativity target, obtains doubtful navigation mark sequence;
Step 6 B, travels through doubtful navigation mark, and the navigation mark not traveling through is set up to new relevant group, and this doubtful navigation mark of mark travels through, searches its all similar doubtful navigation marks, and adds in new group;
Step 6 C, travels through doubtful navigation mark in new group, and this doubtful navigation mark of mark travels through, and the doubtful navigation mark similar to doubtful boat not occurring in new group added in new group;
Step 6 D, in new group the similar navigation mark of all doubtful navigation marks all in new group, the relevant group of output; Otherwise repeating step six C;
Step 6 E, all navigation marks are all traversed, export all relevant group; Otherwise repeating step six B, step 6 C, step 6 D;
Step 6 F, using maximal correlation group as navigation mark group.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310101493.3A CN104077592B (en) | 2013-03-27 | 2013-03-27 | Automatic extraction method for high-resolution remote-sensing image navigation mark |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310101493.3A CN104077592B (en) | 2013-03-27 | 2013-03-27 | Automatic extraction method for high-resolution remote-sensing image navigation mark |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104077592A true CN104077592A (en) | 2014-10-01 |
| CN104077592B CN104077592B (en) | 2017-02-08 |
Family
ID=51598838
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310101493.3A Active CN104077592B (en) | 2013-03-27 | 2013-03-27 | Automatic extraction method for high-resolution remote-sensing image navigation mark |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104077592B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106384081A (en) * | 2016-08-30 | 2017-02-08 | 水利部水土保持监测中心 | Slope farmland extracting method and system based on high-resolution remote sensing image |
| CN107610114A (en) * | 2017-09-15 | 2018-01-19 | 武汉大学 | Optical satellite remote sensing image cloud snow mist detection method based on SVMs |
| CN112651277A (en) * | 2020-09-16 | 2021-04-13 | 武昌理工学院 | Remote sensing target analysis method based on multi-source image |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101446642A (en) * | 2008-04-11 | 2009-06-03 | 国家卫星气象中心 | Automatic matching method for remote sensing satellite data ground control point based on knowledge learning |
| CN101893711A (en) * | 2009-05-20 | 2010-11-24 | 中国科学院电子学研究所 | Ocean distributed target detection method based on collective evolution of agents |
| CN101980294A (en) * | 2010-09-25 | 2011-02-23 | 西北工业大学 | Remote sensing image-based method for detecting ice flood of Yellow River |
-
2013
- 2013-03-27 CN CN201310101493.3A patent/CN104077592B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101446642A (en) * | 2008-04-11 | 2009-06-03 | 国家卫星气象中心 | Automatic matching method for remote sensing satellite data ground control point based on knowledge learning |
| CN101893711A (en) * | 2009-05-20 | 2010-11-24 | 中国科学院电子学研究所 | Ocean distributed target detection method based on collective evolution of agents |
| CN101980294A (en) * | 2010-09-25 | 2011-02-23 | 西北工业大学 | Remote sensing image-based method for detecting ice flood of Yellow River |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106384081A (en) * | 2016-08-30 | 2017-02-08 | 水利部水土保持监测中心 | Slope farmland extracting method and system based on high-resolution remote sensing image |
| CN106384081B (en) * | 2016-08-30 | 2020-04-24 | 水利部水土保持监测中心 | Slope farmland extraction method and system based on high-resolution remote sensing image |
| CN107610114A (en) * | 2017-09-15 | 2018-01-19 | 武汉大学 | Optical satellite remote sensing image cloud snow mist detection method based on SVMs |
| CN107610114B (en) * | 2017-09-15 | 2019-12-10 | 武汉大学 | optical satellite remote sensing image cloud and snow fog detection method based on support vector machine |
| CN112651277A (en) * | 2020-09-16 | 2021-04-13 | 武昌理工学院 | Remote sensing target analysis method based on multi-source image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104077592B (en) | 2017-02-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Guan et al. | Automated road information extraction from mobile laser scanning data | |
| CN103389103B (en) | A kind of Characters of Geographical Environment map structuring based on data mining and air navigation aid | |
| Yu et al. | Automated detection of urban road manhole covers using mobile laser scanning data | |
| Yoo et al. | Beach volume change using UAV photogrammetry Songjung beach, Korea | |
| Fouque et al. | Matching raw GPS measurements on a navigable map without computing a global position | |
| CN104504099A (en) | Position-trajectory-based travel state splitting method | |
| Kahraman et al. | Road extraction techniques from remote sensing images: A review | |
| CN116778329A (en) | Urban road underground shallow disease detection method, device, equipment and medium | |
| Zheng et al. | Coconut trees detection on the tenarunga using high-resolution satellite images and deep learning | |
| CN115546551A (en) | Deep learning-based geographic information extraction method and system | |
| CN104077592A (en) | Automatic extraction method for high-resolution remote-sensing image navigation mark | |
| Wang et al. | A knowledge-based method for road damage detection using high-resolution remote sensing image | |
| Li et al. | 3D map system for tree monitoring in hong kong using google street view imagery and deep learning | |
| CN105631849A (en) | Polygon object change detection method and device | |
| US20220171061A1 (en) | Apparatus and method for recognizing high-elevation structure using lidar sensor | |
| Agoub et al. | Generating 3D city models based on the semantic segmentation of lidar data using convolutional neural networks | |
| CN113227713A (en) | Method and system for generating environment model for positioning | |
| Wang | Lane detection of intelligent assisted driving system based on convolutional neural network. | |
| Shokri et al. | POINTNET++ Transfer Learning for Tree Extraction from Mobile LIDAR Point Clouds | |
| Jenerowicz et al. | URBAN GROWTH MONITORING–REMOTE SENSING METHODS FOR SUSTAINABLE DEVELOPMENT | |
| Widyaningrum et al. | Tailored features for semantic segmentation with a DGCNN using free training samples of a colored airborne point cloud | |
| Ali et al. | Automatic extractions of road intersections from satellite imagery in urban areas | |
| Gupta et al. | Urban Growth Trend Analysis of Indore city (2005-2014) through Index Based Models | |
| CN117392347B (en) | Map construction method, device, computer equipment and readable storage medium | |
| CN115908897B (en) | Intelligent identification method for high-speed railway power supply facilities based on high-resolution remote sensing images |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 200011 Tibet South Road, Shanghai, No. 1170, No. Patentee after: Shanghai urban construction design & Research Institute (Group) Co., Ltd. Address before: 200125 Dongfang Road, Shanghai, Pudong New Area, No. 3447 Patentee before: Shanghai Urban Construction Design & Research Institute |