CN106886780A - Special vehicle extracting method in a kind of low altitude remote sensing image - Google Patents
Special vehicle extracting method in a kind of low altitude remote sensing image Download PDFInfo
- Publication number
- CN106886780A CN106886780A CN201510907603.4A CN201510907603A CN106886780A CN 106886780 A CN106886780 A CN 106886780A CN 201510907603 A CN201510907603 A CN 201510907603A CN 106886780 A CN106886780 A CN 106886780A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- extraction
- special vehicle
- special
- feature
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30248—Vehicle exterior or interior
Abstract
Special vehicle extracting method, comprises the following steps in a kind of low altitude remote sensing image：(1) it is different with the when phase character of common vehicle according to special vehicle, electromagnetism POP and space characteristics are distributed as the distinguishing characteristic of the two；(2) region where determining vehicle；(3) vehicle extraction：According to the region that vehicle is likely to occur, electromagnetic spectrum feature, spatial distribution characteristic and changes with time feature according to vehicle carry out vehicle extraction with the FX modules of ENVI；(4) after vehicle extraction is finished, extracting again for special vehicle is carried out as according to differentiation using the difference on the reflectivity of POP.
Description
Technical field
The present invention relates to Remote Sensing Image Processing Technology field, and in particular to special vehicle extracting method in a kind of low altitude remote sensing image.
Background technology
In high-resolution aviation remote sensing image application, the extraction to special vehicle entity is particularly important, and traditional distant
Sense Extraction of Image method is less for the Study on Extraction Method of small range entity, and the resolution requirement for remote sensing image is higher, carries
The precision for taking also has much room for improvement.
The content of the invention
It is an object of the invention to provide being special vehicle extracting method in a kind of low altitude remote sensing image, to the side of existing extraction
Method is improved, and is using providing more accurate solution in aviation remote sensing image.
Technical scheme is as follows：
Special vehicle extracting method in a kind of low altitude remote sensing image, it is characterised in that：
(1) it is different with the when phase character of common vehicle according to special vehicle, using electromagnetism POP and space characteristics distribution both
Distinguishing characteristic；
(2) region where determining vehicle：Suitable threshold range is selected, first the natural landscape in significantly image is separated
Out；The wherein obvious building site of feature is extracted using small multi-scale segmentation；Secondly continuity had according to road, connected
The similar feature of continuous property, form, the differentiation of the places of cultural interest is carried out with reference to spectral reflectivity；
(3) vehicle extraction：According to the region that vehicle is likely to occur, the electromagnetic spectrum feature according to vehicle, spatial distribution characteristic
With changes with time feature, vehicle extraction is carried out with the FX modules of ENVI；
(4) special vehicle based on classification results is extracted again：After vehicle extraction is finished, made with the difference on the reflectivity of POP
It is to carry out extracting again for special vehicle according to differentiation.
The present invention solves the method that special vehicle is differentiated in aviation image.Propose and special vehicle is extracted from common vehicle
Specific strategy.Due to the diversity of special vehicle, there is certain error in the precision of differentiation.The flow of extraction needs for a small amount of
Manual intervention, it is impossible to realize full-automatic extraction.But the present invention provides possible ways for the concrete application of aviation image, when
In the case that effect property and definition are allowed, for aviation image provides referential opinion as advanced reconnaissance means.
Specific embodiment
(1) determine the difference of special vehicle and common vehicle and contact：
Special vehicle is similar to common vehicle extraction, but has its particular point, and method of discrimination has more to enter in traditional extraction process
The development of step.Identical point is similar target morphology, and difference is color, is compared with common vehicle, the standard differentiated by color
True property is relatively low, and special vehicle, can also be directly in types such as meadow, unused lands in addition to being advanced on road in addition
Ground motion, so the extraction on road is similar to vehicle, but target in earth's surface needs to be additionally carried out treatment.Extract stream
Journey generally carries out classification extraction by conventional method to aviation image first, and the region being then likely to occur to special vehicle is carried out
Emphasis is extracted, and with reference to specific color and texture, is extracted again step by step, finally reaches preferable extraction effect, is reached point
Separate out the purpose of weaponry in aviation remote sensing image.
In the last few years, the research for remote sensing image vehicle extraction has a lot, but because the extraction for special vehicle is for shadow
The ageing and quality of picture has requirement higher, so the research of correlation is less.Set up mostly for the extraction of vehicle and extracting road
Carried out on the basis of the information of road, for road spectral color and the difference of vehicle, by the electronics Spectral Characteristic of vehicle, space point
Phase character when Boot is sought peace, extracts to the vehicle on remote sensing image.
Special vehicle on ground etc. is following three aspect with the difference of common vehicle.
First, in electromagnetic spectrum characteristic aspect, because common vehicle and special vehicle material belong to metal class together, vehicle enclosure is big absolutely
Part is metal material, steel plate, carbon fiber, aluminium, reinforced plastics etc., and special vehicle is generally compound shape steel plate, so to electromagnetism
The albedo of POP is similar.But because the color of common vehicle is varied, so not being both in terms of color is made
The reason for maximum different into electromagnetic spectrum feature, can only be with reference to the dark vehicle reflectivity of similar color.
Second, spatial distribution characteristic aspect, common vehicle is more to travel or is stopped in parking lot on road.Travelled on road
Vehicle, headstock direction is mostly parallel with road direction, and in parking lot, parked vehicle is also generally regular directive stop,
And special vehicle except on road travel, beyond parking lot is parked, it is also larger may be on meadow, the type such as unused land
Soil on travel, the soil of the type in addition to forest land, waters, building etc. is likely to special vehicle occur.Several
What is all made up of three rectangles of headstock tailstock vehicle body in shape, and both sides are lower slightly middle high, and special vehicle is likely to be unified high
Spend or be also three-stage, headstock vehicle body tailstock composition is possible to see on the image of 0.5 meter of resolution ratio.
3rd, the phenomenon generation of roof, widow reflections is had on the image that phase characteristic aspect, different time are obtained, produce
The over-exposed figure spot of life.Simultaneously image typically choose sunlight it is strong when shot, it is the same with common vehicle, it is special
Planting vehicle also has heavier shade, and many shade features based on quadrangle, and Enhanced feature extracts point.
In sum, common vehicle essentially consists in phase characteristic aspect with the difference of special vehicle, so electromagnetism POP aspect and sky
Between feature distribution aspect, can partly with reference to extract common vehicle method.
(2) the multiple dimensioned basic land-use style segmentation of object-oriented：
The Remote Sensing Image Segmentation of object-oriented refers to that ground object target is split with spatial framework under different yardsticks.For
The characteristics of in aerial image containing abundant terrestrial object information, the segmentation for be layered substep is more beneficial for experimental result.
Under different scale, different algorithms are selected to be split.
The present invention selects rule-based object-oriented dividing method in large scale, selects suitable threshold range by natural landscape
Separated from significantly image first, because pilot region landform is single and natural landscape, the places of cultural interest have the obvious spy for demarcating
Point, selects simple threshold range and merges yardstick, can reach the purpose for most efficiently, most making victory.Natural landscape only has grassland
And waters, so that two class lands used are determined.
After places of cultural interest boundary is determined, the wherein obvious building site of feature is extracted first using small multi-scale segmentation
Come, according to its special spectral reflectivity, geometry and architectural feature are extracted.Next carries out road extraction, according to road
With the feature such as continuity, continuity, form be similar, the feature extraction based on edge is carried out with reference to spectral reflectivity.Will building
After land used and road extraction are finished, remaining in the places of cultural interest is exactly greenery patches.
(3) vehicle extraction：
After base categories are completed, it is analyzed, sums up vehicle and be only possible to appear in the meadow in natural landscape and the places of cultural interest
In road and greenery patches in.When road on the meadow of natural landscape and the places of cultural interest occurs, vehicle is mainly to travel
State occurs, its spatial distribution be mainly with form a team or it is scattered it is individual occur, and when appearance on the greenery patches of the places of cultural interest,
Mostly it is then resting state, spatial distribution is generally that parked side by side or majority occur in groups.
The region being likely to occur according to vehicle is carried according to the electromagnetic spectrum feature of vehicle, spatial distribution characteristic and changes with time feature
Take.For the form of expression in different land-use styles, choose different methods and extracted.The present invention uses the FX of ENVI
Module is extracted.
(4) special vehicle based on classification results is extracted again：
After vehicle extraction is finished, carry out the extraction of special vehicle, invention with the difference on the reflectivity of POP can as according to
According to differentiation.
The general not independent action of special vehicle, the Group activity of at least more than three, so either under transport condition or
All it is occur in groups under person's resting state.
On grassland, because special vehicle volume and weight is all larger, in the process of moving, the airborne dust that vehicle is trailed is higher,
Can clearly be displayed on image, so when being extracted to driving vehicle, this distinguishes common in-vehicle as key character
And special vehicle.
The special vehicle classification that will be extracted, can determine traveling and stop two states, in natural landscape appearance and the places of cultural interest
The special vehicle occurred on road is all the vehicle for having task, and the armed vehicle stopped at the places of cultural interest is generally without task status car
.
Claims (1)
1. special vehicle extracting method in a kind of low altitude remote sensing image, it is characterised in that；
(1) it is different with the when phase character of common vehicle according to special vehicle, using electromagnetism POP and space characteristics distribution both
Distinguishing characteristic；
(2) region where determining vehicle：Suitable threshold range is selected, first the natural landscape in significantly image is separated
Out；The wherein obvious building site of feature is extracted using small multi-scale segmentation；Secondly continuity had according to road, connected
The similar feature of continuous property, form, the differentiation of the places of cultural interest is carried out with reference to spectral reflectivity；
(3) vehicle extraction：According to the region that vehicle is likely to occur, the electromagnetic spectrum feature according to vehicle, spatial distribution characteristic
With changes with time feature, vehicle extraction is carried out with the FX modules of ENVI；
(4) special vehicle based on classification results is extracted again：After vehicle extraction is finished, made with the difference on the reflectivity of POP
It is to carry out extracting again for special vehicle according to differentiation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510907603.4A CN106886780A (en) | 2015-12-10 | 2015-12-10 | Special vehicle extracting method in a kind of low altitude remote sensing image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510907603.4A CN106886780A (en) | 2015-12-10 | 2015-12-10 | Special vehicle extracting method in a kind of low altitude remote sensing image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106886780A true CN106886780A (en) | 2017-06-23 |
Family
ID=59173487
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510907603.4A Pending CN106886780A (en) | 2015-12-10 | 2015-12-10 | Special vehicle extracting method in a kind of low altitude remote sensing image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106886780A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104217196A (en) * | 2014-08-26 | 2014-12-17 | 武汉大学 | A method for detecting automatically a circular oil tank with a remote sensing image |
| CN104537348A (en) * | 2014-12-23 | 2015-04-22 | 博康智能网络科技股份有限公司 | Special vehicle recognition method and system |
-
2015
- 2015-12-10 CN CN201510907603.4A patent/CN106886780A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104217196A (en) * | 2014-08-26 | 2014-12-17 | 武汉大学 | A method for detecting automatically a circular oil tank with a remote sensing image |
| CN104537348A (en) * | 2014-12-23 | 2015-04-22 | 博康智能网络科技股份有限公司 | Special vehicle recognition method and system |
Non-Patent Citations (2)
| Title |
|---|
| 刘珠妹 等.: "高分辨率卫星影像车辆检测研究进展", 《遥感技术与应用》 * |
| 王浩: "基于面向对象的高分辨率遥感影像车辆提取方法研究", 《中国优秀硕士学位论文全文数据库（电子期刊）信息科技辑》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Guo et al. | Lidar boosts 3D ecological observations and modelings: A review and perspective | |
| Reese et al. | C-correction of optical satellite data over alpine vegetation areas: A comparison of sampling strategies for determining the empirical c-parameter | |
| Langevin et al. | Typology of dust particles collected by the COSIMA mass spectrometer in the inner coma of 67P/Churyumov Gerasimenko | |
| CN105335966B (en) | Multiscale morphology image division method based on local homogeney index | |
| CN104361338B (en) | A kind of peat bog information extracting method based on ENVISAT ASAR, Landsat TM and dem data | |
| CN104899562B (en) | Radar remote sensing image culture's recognizer based on Texture Segmentation fusion | |
| CN110645961B (en) | Forest resource dynamic change detection method based on remote sensing and NDVI | |
| CN103839267B (en) | Building extracting method based on morphological building indexes | |
| CN109164459A (en) | A kind of method that combination laser radar and high-spectral data classify to forest species | |
| Xu et al. | An improved approach to estimate ratoon rice aboveground biomass by integrating UAV-based spectral, textural and structural features | |
| CN103824077A (en) | Urban impervious layer rate information extraction method based on multi-source remote sensing data | |
| CN106228130B (en) | Remote sensing image cloud detection method of optic based on fuzzy autoencoder network | |
| CN104463164A (en) | Tree canopy structure information extraction method based on rib method and crown height ratio | |
| Mbaabu et al. | Quantification of carbon stock to understand two different forest management regimes in Kayar Khola watershed, Chitwan, Nepal | |
| CN104103070B (en) | Landing point selecting method based on optical images | |
| CN107194876A (en) | A kind of large-scale wild animal population quantity investigation method based on unmanned plane | |
| Gu et al. | Predicting forest height using the GOST, Landsat 7 ETM+, and airborne LiDAR for sloping terrains in the Greater Khingan Mountains of China | |
| CN106340005A (en) | High-resolution remote sensing image unsupervised segmentation method based on scale parameter automatic optimization | |
| CN103035006A (en) | High-resolution aerial image partition method based on LEGION and under assisting of LiDAR | |
| Xu et al. | Spatial pattern analysis of Haloxylon ammodendron using UAV imagery-A case study in the Gurbantunggut Desert | |
| Hu et al. | Integrating CART algorithm and multi-source remote sensing data to estimate sub-pixel impervious surface coverage: a case study from Beijing Municipality, China | |
| Wang et al. | Building heights estimation using ZY3 data—A case study of Shanghai, China | |
| CN106886780A (en) | Special vehicle extracting method in a kind of low altitude remote sensing image | |
| Tuomi et al. | Identification and analysis of flash cells in thunderstorms | |
| Abbate et al. | Relations between morphological settings and vegetation covers in a medium relief landscape of Central Italy |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170623 |