CN105761250A - Building extraction method based on fuzzy scene segmentation - Google Patents
Building extraction method based on fuzzy scene segmentation Download PDFInfo
- Publication number
- CN105761250A CN105761250A CN201610070777.4A CN201610070777A CN105761250A CN 105761250 A CN105761250 A CN 105761250A CN 201610070777 A CN201610070777 A CN 201610070777A CN 105761250 A CN105761250 A CN 105761250A
- Authority
- CN
- China
- Prior art keywords
- image
- building
- method based
- fuzzy
- fuzzy scene
- 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
- G06T2207/10036—Multispectral image; Hyperspectral image
-
- 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
- G06T2207/10041—Panchromatic image
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a building extraction method based on fuzzy scene segmentation. The method comprises the following steps: the step 1, performing image fusion of a panchromatic image with high spatial resolution and a corresponding multispectral image; the step 2, performing shadow extraction of the fused image, and obtaining a shadow mask; the step 3, performing segmentation of the fused image in the step 1 through adoption of the segmentation algorithm; the step 4, generating a fuzzy scene through adoption of the sun illumination angle [Lambda] in a remote-sensing image, the shadow mask obtained in the step 2 and the segmentation result obtained in the step 3; the step 5, performing segmentation of the fuzzy scene generated in the step 4 through adoption of a segmentation algorithm B, and obtaining a segmentation result SegmB; and the step 6, processing the segmentation result SegmB obtained by the step 5, and obtaining a building. The building extraction method based on fuzzy scene segmentation solves the problem that the building extraction is not high in accuracy in a remote-sensing image, achieves a complete automation effect and may be used for remote-sensing image drawing and obtaining and automatic update of data of a geographic information system.
Description
Technical field
The present invention relates to a kind of field of remote sensing image processing, specifically a kind of building extracting method based on fuzzy scene cut.
Background technology
Building is one of main geographic element in city, is the important content of various cities thematic map, and the extraction of research building is significant to integrated survey urban geographic information environment.Along with the fast development of high-resolution remote sensing image acquiring technology, the process of remote sensing image, analysis and application have had better data source, its digital product then had more extensively, deeper into application.The progress all obtained in various degree of the aspects such as computer image processing technology, pattern recognition, artificial intelligence, provides possibility for extracting the effective information in huge image data efficiently.But the extraction of building information is more much more difficult than the acquisition of other information such as road, water body, and main cause is as follows:
(1) remote sensing image of data source mainly two dimension, in most cases lacks direct three-dimensional data;
(2) different remote sensing image Chang Yinwei spectral region, resolution, the difference of the factor such as several picture and image-forming condition of sensor and have bigger difference;
(3) its outward appearance showed and grain details etc. of different types of building is ever-changing, shows widely different on remote sensing images, and unified building model storehouse is difficult to set up, and this makes automatically extracting of information become extremely difficult;
(4) complexity of scene residing for building, time as relatively low in contrast, house mutually block, the shade of building self and be in the shade etc. of other atural object, so it is comparatively difficult to think automatically to extract sharply marginated building from background.
Summary of the invention
The invention provides a kind of building extracting method based on fuzzy scene cut, the problem that in current remote sensing image, building extracts difficulty can be overcome, make full use of the shadow character of remote sensing image, utilize multiple partitioning algorithm, the building target in remote sensing image can be detected, and retaining its edge exactly, it is not necessary to manual intervention, automaticity is high.
Target by realizing the present invention be the technical scheme is that method comprises the following steps:
Step 1: the high spatial resolution panchromatic image P_image of areal and the multispectral image M_image of correspondence thereof is carried out visual fusion process, the fusion evaluation PM_image after being processed;
Step 2: the fusion evaluation PM_image in step 1 is carried out shadow extraction, obtains shadow mask S_mask;
Step 3: utilize partitioning algorithm A that the fusion evaluation PM_image in step 1 is split, obtain segmentation result SegmA;
Step 4: utilize the segmentation result SegmA obtained in the shadow mask S_mask obtained in the solar illumination angle λ in remote sensing image, step 2 and step 3 to generate fuzzy scene FScence;
Step 5: utilize partitioning algorithm B that the fuzzy scene FScence generated in step 4 is split, obtain segmentation result SegmB;
Step 6: utilize pruning method that the segmentation result SegmB obtained in step 5 is carried out post processing and obtain building.
Shadow extraction in described step 2 is calculated by below equation:
Wherein, SB and IB represents the fusion evaluation PM_image saturation component being converted in HIS color space and strength component respectively, and normal represents normalized function, and BW_otsu represents the automatic selected threshold of Da-Jin algorithm and carries out binary conversion treatment function.
Partitioning algorithm A in described step 3 is less divided algorithm, it is possible to selected marker watershed or mean shift algorithm, to guarantee that building is farthest retained, it is prevented that building missing inspection.
The generation method of the fuzzy scene FScence in described step 4 is: for the object S in given shadow mask S_mask, a fuzzy value is distributed for the SegmA object around the object S along solar illumination angle λ direction, ranging for [0,1], value is directly proportional to the distance to object S.
Partitioning algorithm B in described step 5 is the partitioning algorithm that operational efficiency is high and edge laminating degree is high, it is possible to select the super-pixel partitioning algorithm based on entropy rate.
Pruning method in described step 6 is: using the length-width ratio of building, rectangular degree and top homogeneity as constraints, screens qualified cutting object as the building extracted.
The invention has the beneficial effects as follows: solve building in remote sensing image and extract the problem that accuracy is not high, reach full automatic effect.May be used for remote sensing image drawing, the data acquisition of GIS-Geographic Information System and automatically update.
Accompanying drawing explanation
Fig. 1 is the overall process flow figure of the present invention.
Detailed description of the invention
The specific embodiment of the present invention is described in detail below in conjunction with accompanying drawing.
In step 101, the high spatial resolution panchromatic image P_image of areal and the multispectral image M_image of correspondence thereof is carried out visual fusion process, obtains fusion evaluation PM_image.
In step 102, below equation calculate the shadow mask S_mask in extraction step 101:
Wherein, SB and IB represents the fusion evaluation PM_image saturation component being converted in HIS color space and strength component respectively, and normal represents normalized function, and BW_otsu represents the automatic selected threshold of Da-Jin algorithm and carries out binary conversion treatment function.
In step 103, farthest retained in order to ensure building, it is prevented that building missing inspection, utilize mean shift algorithm that the fusion evaluation PM_image in step 101 is carried out less divided, obtain segmentation result SegmA.
In step 104, the solar illumination angle λ in fusion evaluation PM_image is estimated.
In step 105, the segmentation result SegmA obtained in the shadow mask S_mask obtained in the solar illumination angle λ in remote sensing image, step 2 and step 3 is utilized to generate fuzzy scene FScence, method particularly includes:
For obtaining the object S in shadow mask S_mask in given step 102, distributing a fuzzy value for the SegmA object around the object S along solar illumination angle λ direction, range for [0,1], value is directly proportional to the distance to object S.
In step 106, in order to ensure higher operational efficiency and edge laminating degree, utilize the fuzzy scene FScence based on generating in the super-pixel partitioning algorithm step 105 of entropy rate to split, obtain segmentation result SegmB.
In step 107, using the length-width ratio of building, rectangular degree and top homogeneity as constraints, in the segmentation result SegmB obtained in screening step 106, qualified cutting object is as the building extracted.
The visualization of the building extracted in step 108, step 107.
Claims (6)
1. the building extracting method based on fuzzy scene cut, it is characterised in that comprise the following steps:
Step 1: the high spatial resolution panchromatic image P_image of areal and the multispectral image M_image of correspondence thereof is carried out visual fusion process, the fusion evaluation PM_image after being processed;
Step 2: the fusion evaluation PM_image in step 1 is carried out shadow extraction, obtains shadow mask S_mask;
Step 3: utilize partitioning algorithm A that the fusion evaluation PM_image in step 1 is split, obtain segmentation result SegmA;
Step 4: utilize the segmentation result SegmA obtained in the shadow mask S_mask obtained in the solar illumination angle λ in remote sensing image, step 2 and step 3 to generate fuzzy scene FScence;
Step 5: utilize partitioning algorithm B that the fuzzy scene FScence generated in step 4 is split, obtain segmentation result SegmB;
Step 6: utilize pruning method that the segmentation result SegmB obtained in step 5 is carried out post processing and obtain building.
2. a kind of building extracting method based on fuzzy scene cut according to claim 1, it is characterised in that the shadow extraction in step 2 is calculated by below equation:
Wherein, SB and IB represents the fusion evaluation PM_image saturation component being converted in HIS color space and strength component respectively, and normal represents normalized function, and BW_otsu represents the automatic selected threshold of Da-Jin algorithm and carries out binary conversion treatment function.
3. a kind of building extracting method based on fuzzy scene cut according to claim 1, it is characterized in that the partitioning algorithm A in step 3 is less divided algorithm, can selected marker watershed or mean shift algorithm, to guarantee that building is farthest retained, it is prevented that building missing inspection.
4. a kind of building extracting method based on fuzzy scene cut according to claim 1, it is characterized in that the generation method of the fuzzy scene FScence in step 4 is: for the object S in given shadow mask S_mask, a fuzzy value is distributed for the SegmA object around the object S along solar illumination angle λ direction, range for [0,1], value is directly proportional to the distance to object S.
5. a kind of building extracting method based on fuzzy scene cut according to claim 1, it is characterised in that the partitioning algorithm B in step 5 is the partitioning algorithm that operational efficiency is high and edge laminating degree is high, it is possible to select the super-pixel partitioning algorithm based on entropy rate.
6. a kind of building extracting method based on fuzzy scene cut according to claim 1, the pruning method that it is characterized in that in step 6 is: using the length-width ratio of building, rectangular degree and top homogeneity as constraints, screens qualified cutting object as the building extracted.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610070777.4A CN105761250A (en) | 2016-02-01 | 2016-02-01 | Building extraction method based on fuzzy scene segmentation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610070777.4A CN105761250A (en) | 2016-02-01 | 2016-02-01 | Building extraction method based on fuzzy scene segmentation |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105761250A true CN105761250A (en) | 2016-07-13 |
Family
ID=56343020
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610070777.4A Pending CN105761250A (en) | 2016-02-01 | 2016-02-01 | Building extraction method based on fuzzy scene segmentation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105761250A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108509926A (en) * | 2018-04-08 | 2018-09-07 | 福建师范大学 | A kind of building extracting method based on two-way color notation conversion space |
CN108596088A (en) * | 2018-04-23 | 2018-09-28 | 福建师范大学 | A kind of building analyte detection method for panchromatic remote sensing image |
CN114581709A (en) * | 2022-03-02 | 2022-06-03 | 深圳硅基智能科技有限公司 | Model training, method, apparatus, and medium for recognizing target in medical image |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2003153252A (en) * | 2001-11-13 | 2003-05-23 | Toshiba Corp | Remote supervisory system and image processing apparatus |
CN104778714A (en) * | 2015-05-04 | 2015-07-15 | 福建师范大学 | High spatial resolution remote sensing image building detection method based on shadow and segmentation |
CN104794456A (en) * | 2015-05-04 | 2015-07-22 | 福建师范大学 | Remote-sensing image building detecting method based on colors and shadows |
-
2016
- 2016-02-01 CN CN201610070777.4A patent/CN105761250A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2003153252A (en) * | 2001-11-13 | 2003-05-23 | Toshiba Corp | Remote supervisory system and image processing apparatus |
CN104778714A (en) * | 2015-05-04 | 2015-07-15 | 福建师范大学 | High spatial resolution remote sensing image building detection method based on shadow and segmentation |
CN104794456A (en) * | 2015-05-04 | 2015-07-22 | 福建师范大学 | Remote-sensing image building detecting method based on colors and shadows |
Non-Patent Citations (3)
Title |
---|
于东方: "基于阴影的遥感图像中建筑物自动提取方法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
周亚男等: "阴影辅助下的面向对象城市建筑物提取", 《地理与地理信息科学》 * |
杨俊等: "一种高分辨率遥感影像阴影去除方法", 《武汉大学学报﹒信息科学版》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108509926A (en) * | 2018-04-08 | 2018-09-07 | 福建师范大学 | A kind of building extracting method based on two-way color notation conversion space |
CN108509926B (en) * | 2018-04-08 | 2021-06-01 | 福建师范大学 | Building extraction method based on bidirectional color space transformation |
CN108596088A (en) * | 2018-04-23 | 2018-09-28 | 福建师范大学 | A kind of building analyte detection method for panchromatic remote sensing image |
CN108596088B (en) * | 2018-04-23 | 2021-04-20 | 福建师范大学 | Building detection method for panchromatic remote sensing image |
CN114581709A (en) * | 2022-03-02 | 2022-06-03 | 深圳硅基智能科技有限公司 | Model training, method, apparatus, and medium for recognizing target in medical image |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Rottensteiner et al. | The ISPRS benchmark on urban object classification and 3D building reconstruction | |
Ramiya et al. | Segmentation based building detection approach from LiDAR point cloud | |
Yu et al. | Automated derivation of urban building density information using airborne LiDAR data and object-based method | |
CN102521273B (en) | High-resolution remote-sensing multifunctional urban land spatial information generation method | |
Hidayati et al. | Developing an extraction method of urban built-up area based on remote sensing imagery transformation index | |
Cai et al. | Study on shadow detection method on high resolution remote sensing image based on HIS space transformation and NDVI index | |
CN104850853B (en) | City rapid extracting method and device | |
Montesano et al. | Spaceborne potential for examining taiga–tundra ecotone form and vulnerability | |
Berveglieri et al. | Identification of successional stages and cover changes of tropical forest based on digital surface model analysis | |
CN105761250A (en) | Building extraction method based on fuzzy scene segmentation | |
CN107463944B (en) | A kind of road information extracting method using multidate High Resolution SAR Images | |
CN109345450B (en) | Image mosaic method using geographic entity element information | |
CN109635715A (en) | A kind of remote sensing images building extracting method | |
Song et al. | An unsupervised, open-source workflow for 2d and 3d building mapping from airborne lidar data | |
Yadav et al. | Supervised learning based greenery region detection using unnamed aerial vehicle for smart city application | |
Sari et al. | Quality analysis of single tree object with obia and vegetation index from lapan surveillance aircraft multispectral data in urban area | |
CN114758087B (en) | Method and device for constructing urban information model | |
Tian et al. | A process-oriented method for rapid acquisition of canopy height model from RGB point cloud in semiarid region | |
Zhu | A pipeline of 3D scene reconstruction from point clouds | |
Ratnaparkhi et al. | Classification of land use and land cover using remotely sensed data for Parbhani City, Maharashtra, India | |
Liu et al. | A Novel Impervious Surface Extraction Method Based on Automatically Generating Training Samples From Multisource Remote Sensing Products: A Case Study of Wuhan City, China | |
Muller | Key issues in image understanding in remote sensing | |
LOTFIAN | Urban climate modeling: case study of Milan city | |
Motayyeb et al. | Enhancing contrast of images to improve geometric accuracy of a UAV photogrammetry project | |
CN105740825B (en) | It is a kind of for mixing the large format remote sensing image building extracting method of scene |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160713 |