CN103413123B - A kind of high-resolution remote sensing image airfield detection method based on conditional random field models - Google Patents
A kind of high-resolution remote sensing image airfield detection method based on conditional random field models Download PDFInfo
- Publication number
- CN103413123B CN103413123B CN201310332701.0A CN201310332701A CN103413123B CN 103413123 B CN103413123 B CN 103413123B CN 201310332701 A CN201310332701 A CN 201310332701A CN 103413123 B CN103413123 B CN 103413123B
- Authority
- CN
- China
- Prior art keywords
- random field
- remote sensing
- field models
- airport
- sparse coding
- 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.)
- Expired - Fee Related
Links
Abstract
The present invention relates to a kind of high-resolution remote sensing image airfield detection method based on conditional random field models, the process of realization is: first extract the dense SIFT feature of high-resolution remote sensing image, ask for dense SIFT feature at the sparse coding crossed on complete dictionary, then in 4 contiguous range of each sparse coding, set up 4 connected graphs, set up the condition random field models on 4 connected graphs, conditional random field models parameter is obtained according to by Max margin Algorithm Learning, inferred each sparse coding by LBP algorithm and belong to the probit of airport target, thus obtain airport target probability graph, finally airport target probability graph is carried out Threshold segmentation, just can get airfield detection result.The present invention carries out airfield detection in High spatial resolution remote sensing, has accuracy rate high, and the advantage that false alarm rate is low has the highest using value.
Description
Technical field
The invention belongs to technical field of remote sensing image processing, be specifically related to a kind of high-resolution based on conditional random field models
Rate remote sensing images airfield detection method.
Background technology
Along with developing rapidly of satellite and sensor technology, high score rate remote sensing images are that the target detection such as airport provide non-
The abundantest information, effectively utilizes these information, can promote the performance of airfield detection well.Current existing greatly
But most of algorithms is by utilizing the geological informations such as the linear character of airfield runway to realize the detection on airport, but simple geometry
Information cannot very effectively distinguish airport and highway, river, artificial structure etc..Chinese Patent Application No.
201110166001.X, describes one " airport target by using remote sensing image based on Selective Attention Mechanism detection and identification
Method ", use SIFT feature that airport characterizes the detection realizing airport, but high-resolution remote sensing image background is multiple
Miscellaneous, only extract SIFT local feature and carry out identification, easily cause flase drop.And Chinese Patent Application No.
201210282568.8, describe one " infrared remote sensing image based on sparse coding and vision significance detection airport
Method ", use vision significance guide sparse coding airport is carried out characteristic present, at low resolution remote sensing images
The testing result that middle acquirement is pretty good, but still be not applied for carrying out airfield detection at high score rate remote sensing images.
Summary of the invention
Solve the technical problem that
In place of the deficiencies in the prior art, the present invention proposes a kind of high-resolution based on conditional random field models
Remote sensing images airfield detection method.
Technical scheme
A kind of high-resolution remote sensing image airfield detection method based on conditional random field models, it is characterised in that include with
Lower step:
Step 1: extract dense SIFT feature x of high-resolution remote sensing image, to the dense SIFT extracted
Feature employing efficient sparse coding algorithm carries out study and obtained complete dictionary D, asks for dense SIFT
Feature x is at the sparse coding s (x, D) crossed on complete dictionary D；
Step 2: with each sparse coding s (x, D) as summit, set up 4 connected graphs in 4 contiguous range on summit
G, set up the condition random field models on 4 connected graph G；
Step 3: specified criteria random field models parameter, is inferred each sparse coding s (x, D) by LBP algorithm and belongs to
The probit on airport, obtains airport target probability graph；
Step 4: use adaptive threshold fuzziness method to split airport target probability graph, obtains airfield detection knot
Really.
The conditional random field models parameter of described step 3, by using Max-margin Algorithm Learning to obtain.
Beneficial effect
A kind of based on conditional random field models the high-resolution remote sensing image airfield detection method that the present invention proposes, first
Extract the dense SIFT feature of high-resolution remote sensing image, ask for dense SIFT feature and crossing on complete dictionary
Sparse coding, then sets up 4 connected graphs in 4 contiguous range of each sparse coding, sets up on 4 connected graphs
Conditional random field models, obtains conditional random field models parameter according to by Max-margin Algorithm Learning, by LBP algorithm
Infer each sparse coding and belong to the probit of airport target, thus obtain airport target probability graph, finally to machine
Field destination probability figure carries out Threshold segmentation, just can get airfield detection result.
Compared with prior art, the present invention uses sparse coding to characterize airport, can effectively catch the knot on airport
Structure feature, and it is placed in conditional random field models the detection realizing airport, the present invention combines sparse coding and bar
The advantage of part random field models, can characterize airport with robust, and be effectively utilized the spatial information of surrounding neighbors, significantly
Reduce the false alarm rate of airfield detection, have the strongest using value.
Accompanying drawing explanation
Fig. 1 is the flowchart of the present invention.
Fig. 2 is High spatial resolution remote sensing.
Fig. 3 is airport target probability graph.
Fig. 4 is airfield detection result figure.
Detailed description of the invention
In conjunction with embodiment, accompanying drawing, the invention will be further described:
Hardware environment for implementing is: Intel Xeon (R) CPU, E5504@2.0G 2.0G (2 processor) computer,
6.0GB internal memory, 1GB video card, the software environment of operation is: Matlab R2012a, Windows7 64 bit manipulation system
System.We achieve, with Matlab software, the method that the present invention proposes.In experiment, high-resolution remote sensing image used is all
Intercept from Google Earth.
The present invention is embodied as follows:
1. cross complete dictionary learning: for 20 typical airports at 5 kms from Google Earth, 15
Km and the different visual field height of 25 kms three kinds intercept 20*3=60 panel height resolution remote sensing images, and to its every
45 degree carry out rotation transformation, and last there are 60*8=480 width, this 480 width image composition training set.Extract every
The dense SIFT feature of one width training image, extracting the tile size used is 64*64, slides and is spaced apart 16.
Dense SIFT feature to the whole training images extracted, uses efficient sparse coding algorithm to enter
Row study obtained complete dictionary D.
2. extract dense SIFT feature x of Fig. 2, ask for dense SIFT feature x and crossing on complete dictionary D
Sparse coding s (x, D)；
3. each sparse coding s (x, D) is regarded as a summit, and in its 4 contiguous range, sets up 4 connected graph G,
Set up the condition random field models on 4 connected graph G；
4. the every piece image in pair training set carries out airport mark, forms training set Ground Truth, by
Max-margin algorithm obtains conditional random field models parameter in training set and Ground Truth learning thereof；
5. obtain conditional random field models parameter according to study, LBP algorithm infer each sparse coding s (x, D)
Belong to the probit on airport, thus obtain airport target probability graph, i.e. Fig. 3；
6. couple Fig. 3 carries out Threshold segmentation, obtains airfield detection result Fig. 4.
Claims (2)
1. a high-resolution remote sensing image airfield detection method based on conditional random field models, it is characterised in that bag
Include following steps:
Step 1: extract dense SIFT feature x of high-resolution remote sensing image, to the dense SIFT extracted
Feature employing efficient sparse coding algorithm carries out study and obtained complete dictionary D, asks for dense
SIFT feature x is at the sparse coding s (x, D) crossed on complete dictionary D；
Step 2: with each sparse coding s (x, D) as summit, set up 4 in 4 contiguous range on summit
Connected graph G, set up the condition random field models on 4 connected graph G；
Step 3: specified criteria random field models parameter, is inferred each sparse coding s (x, D) by LBP algorithm
Belong to the probit on airport, obtain airport target probability graph；
Step 4: use adaptive threshold fuzziness method to split airport target probability graph, obtains airport inspection
Survey result.
A kind of high-resolution remote sensing image airport based on conditional random field models the most according to claim 1 is examined
Survey method, it is characterised in that: the conditional random field models parameter of described step 3, by using Max-margin
Algorithm Learning obtains.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310332701.0A CN103413123B (en) | 2013-08-01 | 2013-08-01 | A kind of high-resolution remote sensing image airfield detection method based on conditional random field models |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310332701.0A CN103413123B (en) | 2013-08-01 | 2013-08-01 | A kind of high-resolution remote sensing image airfield detection method based on conditional random field models |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103413123A CN103413123A (en) | 2013-11-27 |
| CN103413123B true CN103413123B (en) | 2016-09-21 |
Family
ID=49606132
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310332701.0A Expired - Fee Related CN103413123B (en) | 2013-08-01 | 2013-08-01 | A kind of high-resolution remote sensing image airfield detection method based on conditional random field models |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103413123B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB2523330A (en) | 2014-02-20 | 2015-08-26 | Nokia Technologies Oy | Method, apparatus and computer program product for segmentation of objects in media content |
| CN106056128B (en) * | 2016-04-20 | 2019-03-22 | 北京航空航天大学 | Remote sensing image classification mask method based on composite diagram condition random field |
| CN106778635A (en) * | 2016-12-19 | 2017-05-31 | 江苏慧眼数据科技股份有限公司 | A kind of human region detection method of view-based access control model conspicuousness |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009148731A1 (en) * | 2008-06-02 | 2009-12-10 | Massachusetts Institute Of Technology | Fast pattern classification based on a sparse transform |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8374442B2 (en) * | 2008-11-19 | 2013-02-12 | Nec Laboratories America, Inc. | Linear spatial pyramid matching using sparse coding |
-
2013
- 2013-08-01 CN CN201310332701.0A patent/CN103413123B/en not_active Expired - Fee Related
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009148731A1 (en) * | 2008-06-02 | 2009-12-10 | Massachusetts Institute Of Technology | Fast pattern classification based on a sparse transform |
Non-Patent Citations (3)
| Title |
|---|
| 基于中层语义表示的图像场景分类研究;解文杰;《中国博士学位论文全文数据库信息科技辑》;20110915(第09期);论文正文第76页第2-4段，第77页第1段、附图5-7 * |
| 基于条件随机场的目标提取;张晓峰;《中国博士学位论文全文数据库信息科技辑》;20130315(第03期);论文正文第19页第4-8段，第20页第8段，第21页第1-3段，第55页第3-6段，第56页第1段，第58页第2-6段，第59页第3段、附图4-4，4-5，4-6 * |
| 基于稀疏表示的高空间分辨率遥感影像纹理描述方法的研;步晓亮;《中国优秀硕士学位论文全文数据库信息科技辑》;20120715(第07期);论文正文第19页第5-9段，第20页第1-10段，第21页第1-9段，第22页第6-10段，第23页第12段 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103413123A (en) | 2013-11-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102831402B (en) | Sparse coding and visual saliency-based method for detecting airport through infrared remote sensing image | |
| Gong et al. | Context-aware convolutional neural network for object detection in VHR remote sensing imagery | |
| Cheng et al. | Accurate urban road centerline extraction from VHR imagery via multiscale segmentation and tensor voting | |
| Deng et al. | Entropy-based window selection for detecting dim and small infrared targets | |
| US10255523B2 (en) | Moving vehicle detection and analysis using low resolution remote sensing imagery | |
| Luo et al. | Deeply supervised convolutional neural network for shadow detection based on a novel aerial shadow imagery dataset | |
| CN102722712B (en) | Multiple-scale high-resolution image object detection method based on continuity | |
| CN107563370B (en) | Visual attention mechanism-based marine infrared target detection method | |
| CN103413123B (en) | A kind of high-resolution remote sensing image airfield detection method based on conditional random field models | |
| Yang et al. | Building detection in high spatial resolution remote sensing imagery with the U-Rotation Detection Network | |
| Li et al. | An aerial image segmentation approach based on enhanced multi-scale convolutional neural network | |
| Liu et al. | Building segmentation from satellite imagery using U-Net with ResNet encoder | |
| Tang et al. | Fast multidirectional vehicle detection on aerial images using region based convolutional neural networks | |
| Han et al. | Targets mask U-Net for wind turbines detection in remote sensing images | |
| Wang et al. | A knowledge-based method for road damage detection using high-resolution remote sensing image | |
| Ma et al. | An effective risk identification method for power fence operation based on neighborhood correlation network and vector calculation | |
| Men et al. | Airport runway semantic segmentation based on DCNN in high spatial resolution remote sensing images | |
| Li et al. | Road network extraction in high-resolution SAR images based CNN features | |
| Zhang et al. | HoughLaneNet: Lane detection with deep hough transform and dynamic convolution | |
| Torrione et al. | Histogram of gradient features for buried threat detection in ground penetrating radar data | |
| Chen et al. | A modified convolutional neural network with transfer learning for road extraction from remote sensing imagery | |
| Jiang et al. | Fast Traffic Accident Identification Method Based on SSD Model | |
| Zhang et al. | A vehicle detection method taking shadow areas into account for high resolution aerial imagery | |
| Sun et al. | Contextual models for automatic building extraction in high resolution remote sensing image using object-based boosting method | |
| Sui et al. | An Automatic integrated image segmentation, registration and change detection method for water-body extraction using HSR images and GIS data |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160921 Termination date: 20190801 |