CN111178175A - Automatic building information extraction method and system based on high-view satellite image - Google Patents
Automatic building information extraction method and system based on high-view satellite image Download PDFInfo
- Publication number
- CN111178175A CN111178175A CN201911288831.2A CN201911288831A CN111178175A CN 111178175 A CN111178175 A CN 111178175A CN 201911288831 A CN201911288831 A CN 201911288831A CN 111178175 A CN111178175 A CN 111178175A
- Authority
- CN
- China
- Prior art keywords
- view satellite
- image
- building
- reflectivity data
- satellite
- 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
- 238000000605 extraction Methods 0.000 title claims abstract description 46
- 238000002310 reflectometry Methods 0.000 claims abstract description 59
- 230000011218 segmentation Effects 0.000 claims abstract description 44
- 238000000034 method Methods 0.000 claims abstract description 30
- 238000012937 correction Methods 0.000 claims abstract description 20
- 238000004364 calculation method Methods 0.000 claims abstract description 14
- 230000005855 radiation Effects 0.000 claims abstract description 11
- 230000004927 fusion Effects 0.000 claims abstract description 8
- 238000005070 sampling Methods 0.000 claims abstract description 8
- 238000001914 filtration Methods 0.000 claims description 18
- 230000003595 spectral effect Effects 0.000 claims description 16
- 238000012216 screening Methods 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 6
- 238000003708 edge detection Methods 0.000 claims description 5
- 238000013459 approach Methods 0.000 claims description 3
- 235000009508 confectionery Nutrition 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000006116 polymerization reaction Methods 0.000 claims description 3
- 238000011524 similarity measure Methods 0.000 claims description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 3
- 238000001228 spectrum Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 230000007547 defect Effects 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/176—Urban or other man-made structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Processing (AREA)
Abstract
The invention discloses an automatic building information extraction method and system based on high-view satellite images, wherein the method comprises the following steps: (1) generating high-view satellite reflectivity data L according to a calibration formula for performing radiation correction and atmospheric correction on the high-view satellite dataλ(ii) a Performing orthorectification on the high-view satellite reflectivity data by using a high-precision DEM (digital elevation model) to obtain a high-precision orthorectified image, and performing NND (non-uniform density) fusion on the high-precision orthorectified image to generate high-view satellite reflectivity data of 4 wave bands; (2) performing image multi-scale mean shift segmentation on the reflectivity data of the high view satellite with 4 wave bands to obtain each segmented characteristic unit; (3) sampling each feature unit after segmentationDynamic selection and feature calculation. The invention can realize high-precision and high-efficiency building information extraction.
Description
Technical Field
The invention belongs to the technical field of satellite remote sensing, and particularly relates to an automatic building information extraction method and system based on high-view satellite images.
Background
The high-resolution remote sensing observes the ground in a very fine mode, the obtained image not only has certain spectral information, but also can clearly express the shape, spatial distribution and spatial structure characteristics of a ground object target, the method has the characteristic of map integration, and the method is widely applied to the fields of urban planning, functional area division, three-dimensional modeling and the like. The identification and extraction of the artificial ground object target are the basis of application, and the building is one of the most important artificial ground object targets and is also the important content on the high-resolution remote sensing image. The research on the automatic and high-precision extraction of buildings on the high-resolution remote sensing images has great practical significance for various applications. At present, scholars at home and abroad put forward a large number of methods for extracting buildings, which can be mainly divided into the following three types: (1) extracting the outline of the building by using laser radar (LIDAR), InSAR, DSM and DEM for assistance; (2) extracting a stereopair through photogrammetry data with a larger overlapping area by adopting ultrahigh resolution images such as aerial photogrammetry and the like, determining ground points, and extracting a building outline according to height characteristic information of the building; (3) building extraction is carried out by utilizing a single-scene remote sensing image and assuming that a building has a regular shape, and combining spatial relations among ground objects, such as adjacent building shadows and the like;
the traditional method for extracting the crop information has the defects that:
(1) due to the characteristics of buildings, most house buildings are adjacent to roads, a large number of house edges are parallel to the roads, information of the roads and the house edges in the images after segmentation is easy to be mixed, and the boundaries of the roads and the house edges are difficult to distinguish by using a traditional method. Meanwhile, the diversity of the roof materials also causes the spectral characteristics and the texture characteristics of the tops of different buildings to have larger differences. Therefore, the building extraction by using the building characteristics is greatly limited.
(2) The method has the advantages that the accuracy of the remote sensing images utilizing LIDAR, InSAR and ultrahigh resolution is high, the method can be used in the field of three-dimensional modeling, but the accuracy requirement on auxiliary data with height information such as DSM/DEM is high, and the data meeting the conditions are difficult to obtain under general conditions. Meanwhile, the high-spatial-resolution remote sensing image improves the information content of the ground objects to a certain extent, but also increases noise information and phenomena of 'same-spectrum foreign matters' and 'same-object different-spectrum', and increases the difficulty of high-precision building extraction to a certain extent.
(3) The remote sensing image segmentation scale is difficult to grasp. When the remote sensing image is segmented, the segmentation scale is not standard, and different segmentation scales can generate different extraction results. When the segmentation method is suitable for a segmentation scale, the method is difficult to ensure that the method has good applicability on the whole image.
Disclosure of Invention
The technical problem solved by the invention is as follows: the method and the system overcome the defects of the prior art, provide the automatic building information extraction method and the system based on the high-view satellite image, solve the problem that the building extraction is less performed based on the high-view satellite remote sensing data at present, and can realize the high-precision and high-efficiency building information extraction.
The purpose of the invention is realized by the following technical scheme: an automatic building information extraction method based on high-view satellite images comprises the following steps: (1) generating high-view satellite reflectivity data L according to a calibration formula for performing radiation correction and atmospheric correction on the high-view satellite dataλ(ii) a Performing orthorectification on the high-view satellite reflectivity data by using a high-precision DEM (digital elevation model) to obtain a high-precision orthorectified image, and performing NND (non-uniform density) fusion on the high-precision orthorectified image to generate high-view satellite reflectivity data of 4 wave bands; (2) performing image multi-scale mean shift segmentation on the reflectivity data of the high view satellite with 4 wave bands to obtain each segmented characteristic unit; (3) and automatically selecting a sample and calculating the characteristics of each divided characteristic unit.
In the above automatic building information extraction method based on high view satellite images, in step (1), the calibration formula is:
where QCAL is the original quantized DN value, LλFor high view satellite reflectivity data, LMINλHigh aspect satellite reflectivity data for QCAL ═ QCALMIN, LMAXλQCALMAX/QCALMIN are maximum/minimum quantized scaled pixel values for the radiance value when QCAL is QCALMAX.
In the automatic building information extraction method based on the high view satellite image, in the step (2), the image multi-scale mean shift segmentation is performed on the reflectivity data of the high view satellite with 4 wave bands to obtain each segmented characteristic unit, and the method comprises the following steps: (21) converting the reflectivity data of the high view satellite with 4 wave bands from RGB (red, green and blue) chromaticity space to LUV characteristic space to obtain an LUV image; (22) determining a kernel function on the LUV image; (23) performing mean shift filtering on the processed reflectivity data of the high view satellite with 4 wave bands by using the determined kernel function and the bandwidth; (24) determining merging sequences of different scales, and performing iterative operation on the reflectivity data of the high view satellite of 4 wave bands after mean shift filtering to obtain each segmented characteristic unit.
In the automatic building information extraction method based on the high-view satellite image, in the step (3), the automatic sample selection and the characteristic calculation of each segmented characteristic unit comprise the following steps: (31) calculating the spectral characteristics and the shape characteristics of each segmented characteristic unit; the spectral characteristics comprise spectral mean, variance and entropy; the shape characteristics comprise shape compactness, area and perimeter; (32) according to the fact that the building has a regular shape, vegetation and objects with obvious spectral feature differences of water bodies are removed, building samples are selected from feature units automatically and preliminarily, and spectral mean, variance and texture features of the building samples are calculated; screening a representative building template sample in a polymerization mode; (33) constructing a corresponding template according to the typical shape of a representative building template sample, performing convolution operation on the whole scene image, and extracting a building area; (34) and aiming at the area of the building, carrying out building edge detection, refinement and screening to approach the real edge of the building, thereby realizing the extraction of the space outline of the building.
In the automatic building information extraction method based on the high view satellite image, in step (21), after the color image corresponding to the reflectivity data of the high view satellite of 4 wave bands is mapped to the feature space L, the position of each pixel in the image, namely the spatial information (X, Y), is combined, and the value of each pixel in the 5-dimensional feature space, namely (X, Y, L) can be obtained*，U*，V*) (ii) a Wherein L denotes the brightness of the image, U*And V*Respectively, the color differences are indicated.
In the automatic building information extraction method based on the high view satellite image, in the step (22), the D dimension European space R is subjected todOne characteristic data point x in the vector graph is identified by a column vector, and the modulus of the x is | | x | | sweet calculation2＝xTx and R are identification real number fields; if a function K: rd→ R there is a profile function k: [0, ∞]→ R, i.e. k (x) ═ ck(||x||2) (ii) a Wherein, ck0 is a normalization constant and satisfies: (1) k is non-negative; (2) k is non-increasing; (3) k is piecewise continuous, and { [ integral ] (r) dr < ∞, then the function K (x) is called the kernel function;
for a given RdN sampling points in space xiI is more than or equal to 1 and less than or equal to n, and the sum density of the density function obtained by using the kernel function K (x) is as follows:
wherein, ω (x)i) ≧ 0 is the weight of the sample point, satisfying Σ w (x)i) 1, abbreviated as ωiThe kernel function K (x) determines the sampling point xiA similarity measure with the kernel center point x.
In the automatic building information extraction method based on the high-view satellite image, in step (24), merging sequences { M (M) with different scales are determined1,M2,M3,…,MnPerforming iterative operation on the basis of filtering, and performing iterative operation on the filterStoring the result of the same scale, and when the scale M is reachednAnd stopping iteration, and finally realizing multi-scale segmentation.
An automatic building information extraction system based on high view satellite images comprises: a data preprocessing module for generating high-view satellite reflectivity data L according to a calibration formula for performing radiation correction and atmospheric correction on the high-view satellite dataλ(ii) a Performing orthorectification on the high-view satellite reflectivity data by using a high-precision DEM (digital elevation model) to obtain a high-precision orthorectified image, and performing NND (non-uniform density) fusion on the high-precision orthorectified image to generate high-view satellite reflectivity data of 4 wave bands; the image multi-scale mean shift segmentation module is used for carrying out image multi-scale mean shift segmentation on the reflectivity data of the high view satellite with 4 wave bands to obtain each segmented characteristic unit; and the sample automatic selection and feature calculation module is used for automatically selecting samples and calculating features of each segmented feature unit.
In the above automatic building information extraction system based on high view satellite images, the calibration formula is:
where QCAL is the original quantized DN value, LλFor high view satellite reflectivity data, LMINλHigh aspect satellite reflectivity data for QCAL ═ QCALMIN, LMAXλQCALMAX/QCALMIN are maximum/minimum quantized scaled pixel values for the radiance value when QCAL is QCALMAX.
In the above automatic building information extraction system based on high view satellite images, the image multi-scale mean shift segmentation module includes: the LUV image construction module is used for converting the reflectivity data of the high-view satellite with 4 wave bands from RGB (red, green and blue) chromaticity space to LUV characteristic space to obtain an LUV image; a kernel function determining module for determining a kernel function on the LUV image; the mean filtering module is used for carrying out mean drift filtering on the processed reflectivity data of the high view satellite with 4 wave bands by utilizing the determined kernel function and the determined bandwidth; and the multi-scale segmentation module is used for determining merging sequences of different scales and performing iterative operation on the reflectivity data of the high view satellite of 4 wave bands after the mean shift filtering to obtain each segmented characteristic unit.
Compared with the prior art, the invention has the following beneficial effects:
(1) the invention can realize high-precision and high-efficiency building information extraction on the basis of adopting high-view satellite remote sensing data with 0.5m resolution and improving the automatic extraction difficulty of the building.
(2) The invention utilizes the multi-drift-mean multi-scale segmentation technology, can solve the problem that the segmentation scale is difficult to determine when the scale segmentation technology is adopted at present, and the segmentation method is faster and more efficient.
(3) The invention comprehensively considers the spectrum, shape and texture information of the top of the building, accurately constructs the building template through special sample selection and characteristic calculation, and improves the accuracy of building information extraction.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to refer to like parts throughout the drawings. In the drawings:
fig. 1 is a flowchart of an automated building information extraction method based on high-view satellite images according to an embodiment of the present invention;
fig. 2 is a flowchart of multi-scale mean-shift segmentation of an image according to an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art. It should be noted that the embodiments and features of the embodiments may be combined with each other without conflict. The present invention will be described in detail below with reference to the embodiments with reference to the attached drawings.
Fig. 1 is a flowchart of an automated building information extraction method based on high-view satellite images according to an embodiment of the present invention. As shown in fig. 1, the method comprises the steps of:
(1) generating high-view satellite reflectivity data L according to a calibration formula for performing radiation correction and atmospheric correction on the high-view satellite dataλ；
Performing orthorectification on the high-view satellite reflectivity data by using a high-precision DEM (digital elevation model) to obtain a high-precision orthorectified image, and performing NND (non-uniform density) fusion on the high-precision orthorectified image to generate high-view satellite reflectivity data of 4 wave bands;
(2) performing image multi-scale mean shift segmentation on the reflectivity data of the high view satellite with 4 wave bands to obtain each segmented characteristic unit;
(3) and automatically selecting a sample and calculating the characteristics of each divided characteristic unit.
Specifically, the method comprises the following steps of data preprocessing, image multi-scale mean shift segmentation, automatic sample selection and feature calculation, building template construction, building contour extraction and the like:
(1) data preprocessing: including radiation correction and orthofusion.
(a) Radiometric calibration and atmospheric correction: in order to eliminate the influence of the atmosphere, radiation correction and atmosphere correction are required to be carried out on the high-scene satellite data to generate reflectivity data. The scaling formula is as follows:
where QCAL is the original quantized DN value, LλAs radiance/reflectance values, LMINλRadiance/reflectance values at QCALMIN, LMAXλQCALMAX/QCALMIN are maximum/minimum quantized scaled pixel values for the radiance value when QCAL is QCALMAX. In general, QCALMIN is 0. Value of gain in the formula
Offset-LMINλ；
(b) Orthofusion: because the height information of the building exists, the ground target can be deformed in the imaging process, in order to improve the extraction precision of the outline information of the building, firstly, the high-precision DEM is used for performing orthorectification on the reflectivity data of the high-view satellite to obtain a high-precision orthoimage, and the high-precision orthoimage is subjected to NND fusion on the multispectral and panchromatic images to generate a 0.5m high-precision four-waveband orthoproduct;
(2) image multi-scale mean shift segmentation: the method is used for realizing rapid and steady multi-scale segmentation of high-scene data and providing sample information for building templates constructed later, and the flow is shown in fig. 2 and mainly comprises the following 4 steps:
(a) LUV image construction: and (3) carrying out LUV image construction processing on the orthographic fused high view satellite image, namely converting the high view satellite reflectivity data of 4 wave bands from RGB (red, green and blue) chromaticity space to LUV characteristic space, thereby better realizing the separation of the characteristic space. When the color image is mapped to the feature space L, the position of the pixel in the image, i.e. the spatial information (X, Y), is combined, and the value of each pixel in the 5-dimensional feature space, i.e. (X, Y, L), can be obtained*，U*，V*). Wherein L denotes the brightness of the image, U*And V*Respectively, the color differences are indicated. On the basis, points with similar Euclidean distances between the space and the color can be classified into one class by adopting a clustering algorithm, so that the color image segmentation is realized.
(b) Determining kernel function k (x) and bandwidth Hi: the kernel function and the corresponding parameters are determined on the LUV image. For d dimension Euclidean space RdOne characteristic data point x in the method is marked with a column vector reticle, and the mode of x is | | x | | sweet calculation2＝xTAnd x, R scale real number domain. If a function K: rd→ R there is a profile function k: [0, ∞]→ R, i.e. k (x) ═ ck(||x||2)
Wherein, ck0 is a normalization constant and satisfies: (1) k is non-negative; (2) k is not increasing, i.e. asFruit a<b, k (a) is not less than k (b); (3) k is piecewise continuous, and ^ K (r) dr < ∞, then the function K (x) is called a kernel function. For a given RdN sampling points in space xiI is more than or equal to 1 and less than or equal to n, the sum density estimation formula of the density function is as follows by using a kernel function K (x) and a positive definite dXd bandwidth matrix Hi:
wherein, ω (x)i) ≧ 0 is the weight of the sample point, satisfying Σ w (x)i) 1, abbreviated as ωiThe kernel function K (x) determines the sampling point xiThe similarity measure with the kernel center point x, and the bandwidth matrix Hi determines the image range of the kernel function. Intuitively, density function estimation
Is the result of a weighted summation of the kernel functions from each sampling point.
(c) And (3) average filtering based on space and color domain: using a determined kernel function K (x) and a bandwidth HiAnd performing mean shift filtering on the processed high-resolution high-view satellite image remote sensing data with the resolution of 0.5 m.
(d) Multi-scale segmentation: determining merging sequences of different scales { M1,M2,M3,…,MnPerforming iterative operation on the basis of filtering, storing results of different scales, and when the scale M is reachednAnd stopping iteration, and finally realizing multi-scale segmentation.
(3) And automatically selecting a sample and calculating characteristics. Compared with pixels, the feature primitives after scale division have objectified features such as shapes, and the like, and the shape features can be described and expressed through some indexes, such as features of spectral mean, variance, entropy, shape compactness, area, perimeter and the like. And performing feature calculation on the building by using the features, and extracting the spatial outline of the building by edge detection, refinement and screening.
It mainly comprises the following steps:
(a) calculating the spectrum and shape characteristics of each divided characteristic unit, and calculating the characteristics of the spectrum mean value, variance, entropy, shape compactness, area, perimeter and the like of each characteristic unit;
(b) according to the fact that the building has a regular shape, objects with obvious spectral characteristic differences such as vegetation and water bodies are removed, building samples are selected from characteristic units automatically and preliminarily, and the spectral mean, the variance, the texture and other characteristics of the building samples are calculated; then screening out the building template sample with the last representation in a polymerization mode according to the comprehensive threshold values of the shape characteristics and the texture characteristics;
(c) and constructing a corresponding template according to the typical shape of the building, wherein the template has the characteristics of spectrum and shape at the same time, and performing convolution operation on the whole scene image to extract a building area.
(d) And aiming at the extracted building area, carrying out building edge detection, refinement and screening to approach the real edge of the building, thereby realizing the extraction of the space outline of the building.
The embodiment also provides an automatic building information extraction system based on high view satellite images, which comprises: a data preprocessing module for generating high-view satellite reflectivity data L according to a calibration formula for performing radiation correction and atmospheric correction on the high-view satellite dataλ(ii) a Performing orthorectification on the high-view satellite reflectivity data by using a high-precision DEM (digital elevation model) to obtain a high-precision orthorectified image, and performing NND (non-uniform density) fusion on the high-precision orthorectified image to generate high-view satellite reflectivity data of 4 wave bands; the image multi-scale mean shift segmentation module is used for carrying out image multi-scale mean shift segmentation on the reflectivity data of the high view satellite with 4 wave bands to obtain each segmented characteristic unit; and the sample automatic selection and feature calculation module is used for automatically selecting samples and calculating features of each segmented feature unit.
In the above embodiment, the scaling formula is:
where QCAL is the original quantized DN value, LλFor high view satellite reflectivity data, LMINλWhen QCAL is QCALMINOf high view satellites, LMAXλQCALMAX/QCALMIN are maximum/minimum quantized scaled pixel values for the radiance value when QCAL is QCALMAX.
In the above embodiment, the image multi-scale mean shift segmentation module includes: the LUV image construction module is used for converting the reflectivity data of the high-view satellite with 4 wave bands from RGB (red, green and blue) chromaticity space to LUV characteristic space to obtain an LUV image; a kernel function determining module for determining a kernel function on the LUV image; the mean filtering module is used for carrying out mean drift filtering on the processed reflectivity data of the high view satellite with 4 wave bands by utilizing the determined kernel function and the determined bandwidth; and the multi-scale segmentation module is used for determining merging sequences of different scales and performing iterative operation on the reflectivity data of the high view satellite of 4 wave bands after the mean shift filtering to obtain each segmented characteristic unit.
The method comprises the steps of firstly preprocessing high-view satellite data such as radiation correction and orthorectification, removing influences of atmosphere and terrain on imaging, obtaining a high-precision 0.5m high-view satellite orthoreflectance product, carrying out multi-scale segmentation on the reflectance product by using a remote sensing multi-scale mean shift segmentation technology to generate multi-scale segmentation results, and then carrying out sample selection and feature calculation on the segmented product to realize rapid and high-precision extraction of building information based on high-view satellite remote sensing data.
When the method is applied to the extraction of the high-view satellite remote sensing data building, only data preprocessing work such as radiation correction, orthorectification and the like is needed to be carried out on the high-view satellite remote sensing image; after determining a kernel function and bandwidth of the high view satellite remote sensing data drifting mean value segmentation, carrying out multi-scale segmentation on the high view satellite remote sensing data to obtain characteristic elements; and then, building templates are constructed by combining the spectral characteristics and the textural characteristics of the buildings, the building templates are applied to extracting the building outlines, and finally, the extracted building outlines are subjected to edge detection, refinement and screening, so that the rapid and high-precision extraction of the building outlines is realized. Compared with the traditional building outline information extraction method, the main advantages and improvements of the invention mainly comprise the following aspects:
firstly, on the basis of adopting high-view satellite remote sensing data with 0.5m resolution and improving the automatic extraction difficulty of the building, the high-precision and high-efficiency building information extraction can be realized.
Secondly, by utilizing a multi-drift-mean multi-scale segmentation technology, the problem that the segmentation scale is difficult to determine when the scale segmentation technology is adopted at present can be solved, and the segmentation method is quicker and more efficient.
And thirdly, the spectrum, shape and texture information of the top of the building are comprehensively considered, and a building template is accurately constructed through special sample selection and characteristic calculation, so that the accuracy of building information extraction is improved.
Although the present invention has been described with reference to the preferred embodiments, it is not intended to limit the present invention, and those skilled in the art can make variations and modifications of the present invention without departing from the spirit and scope of the present invention by using the methods and technical contents disclosed above.
Claims (10)
1. An automatic building information extraction method based on high view satellite images is characterized by comprising the following steps:
(1) generating high-view satellite reflectivity data L according to a calibration formula for performing radiation correction and atmospheric correction on the high-view satellite dataλ；
Performing orthorectification on the high-view satellite reflectivity data by using a high-precision DEM (digital elevation model) to obtain a high-precision orthorectified image, and performing NND (non-uniform density) fusion on the high-precision orthorectified image to generate high-view satellite reflectivity data of 4 wave bands;
(2) performing image multi-scale mean shift segmentation on the reflectivity data of the high view satellite with 4 wave bands to obtain each segmented characteristic unit;
(3) and automatically selecting a sample and calculating the characteristics of each divided characteristic unit.
2. The method for extracting information of an automated building based on high-vision satellite images as claimed in claim 1, wherein: in step (1), the scaling formula is:
where QCAL is the original quantized DN value, LλFor high view satellite reflectivity data, LMINλHigh aspect satellite reflectivity data for QCAL ═ QCALMIN, LMAXλQCALMAX/QCALMIN are maximum/minimum quantized scaled pixel values for the radiance value when QCAL is QCALMAX.
3. The method for extracting information of an automated building based on high-vision satellite images as claimed in claim 1, wherein: in the step (2), the image multi-scale mean shift segmentation is performed on the reflectivity data of the high view satellite in 4 wave bands to obtain each segmented feature unit, and the method comprises the following steps:
(21) converting the reflectivity data of the high view satellite with 4 wave bands from RGB (red, green and blue) chromaticity space to LUV characteristic space to obtain an LUV image;
(22) determining a kernel function on the LUV image;
(23) performing mean shift filtering on the processed reflectivity data of the high view satellite with 4 wave bands by using the determined kernel function and the bandwidth;
(24) determining merging sequences of different scales, and performing iterative operation on the reflectivity data of the high view satellite of 4 wave bands after mean shift filtering to obtain each segmented characteristic unit.
4. The method for extracting information of an automated building based on high-vision satellite images as claimed in claim 3, wherein: in the step (3), the automatic sample selection and feature calculation for each segmented feature unit comprises the following steps:
(31) calculating the spectral characteristics and the shape characteristics of each segmented characteristic unit; the spectral characteristics comprise spectral mean, variance and entropy; the shape characteristics comprise shape compactness, area and perimeter;
(32) according to the fact that the building has a regular shape, vegetation and objects with obvious spectral feature differences of water bodies are removed, building samples are selected from feature units automatically and preliminarily, and spectral mean, variance and texture features of the building samples are calculated; screening a representative building template sample in a polymerization mode;
(33) constructing a corresponding template according to the typical shape of a representative building template sample, performing convolution operation on the whole scene image, and extracting a building area;
(34) and aiming at the area of the building, carrying out building edge detection, refinement and screening to approach the real edge of the building, thereby realizing the extraction of the space outline of the building.
5. The method for extracting information of an automated building based on high-vision satellite images as claimed in claim 3, wherein: in step (21), after the color image corresponding to the reflectivity data of the high view satellite in 4 bands is mapped to the feature space L, the position of the pixel in the image, i.e. the spatial information (X, Y), is combined, so as to obtain the value of each pixel in the 5-dimensional feature space, i.e. (X, Y, L), i.e. the spatial information (X, Y)*，U*，V*) (ii) a Wherein L denotes the brightness of the image, U*And V*Respectively, the color differences are indicated.
6. The method for extracting information of an automated building based on high-vision satellite images as claimed in claim 3, wherein: in step (22), for d dimension Euclidean space RdOne characteristic data point x in the vector graph is identified by a column vector, and the modulus of the x is | | x | | sweet calculation2＝xTx and R are identification real number fields; if a function K: rd→ R there is a contour function k: [0, ∞]→ R, i.e. k (x) ═ ck(||x||2) (ii) a Wherein, ck0 is a normalization constant and satisfies: (1) k is non-negative; (2) k is non-increasing; (3) k is piecewise continuous, and { [ integral ] (r) dr < ∞, then the function K (x) is called the kernel function;
for a given RdN sampling points in space xiI is more than or equal to 1 and less than or equal to n, and the sum density of the density function obtained by using the kernel function K (x) is as follows:
wherein, ω (x)i) ≧ 0 is the weight of the sample point, satisfying Σ w (x)i) 1, abbreviated as ωiThe kernel function K (x) determines the sampling point xiA similarity measure with the kernel center point x.
7. The method for extracting information of an automated building based on high-vision satellite images as claimed in claim 3, wherein: in step (24), different scale merge sequences { M } are determined1，M2，M3，...，MnPerforming iterative operation on the basis of filtering, storing results of different scales, and when the scale M is reachednAnd stopping iteration, and finally realizing multi-scale segmentation.
8. An automatic building information extraction system based on high view satellite images is characterized by comprising:
a data preprocessing module for generating high-view satellite reflectivity data L according to a calibration formula for performing radiation correction and atmospheric correction on the high-view satellite dataλ(ii) a Performing orthorectification on the high-view satellite reflectivity data by using a high-precision DEM (digital elevation model) to obtain a high-precision orthorectified image, and performing NND (non-uniform density) fusion on the high-precision orthorectified image to generate high-view satellite reflectivity data of 4 wave bands;
the image multi-scale mean shift segmentation module is used for carrying out image multi-scale mean shift segmentation on the reflectivity data of the high view satellite with 4 wave bands to obtain each segmented characteristic unit;
and the sample automatic selection and feature calculation module is used for automatically selecting samples and calculating features of each segmented feature unit.
9. The automated building information extraction system based on high-vision satellite imagery according to claim 8, wherein: the scaling formula is:
where QCAL is the original quantized DN value, LλFor high view satellite reflectivity data, LMINλHigh aspect satellite reflectivity data for QCAL ═ QCALMIN, LMAXλQCALMAX/QCALMIN are maximum/minimum quantized scaled pixel values for the radiance value when QCAL is QCALMAX.
10. The automated building information extraction system based on high-vision satellite imagery according to claim 8, wherein: the image multi-scale mean shift segmentation module comprises:
the LUV image construction module is used for converting the reflectivity data of the high-view satellite with 4 wave bands from RGB (red, green and blue) chromaticity space to LUV characteristic space to obtain an LUV image;
a kernel function determining module for determining a kernel function on the LUV image;
the mean filtering module is used for carrying out mean drift filtering on the processed reflectivity data of the high view satellite with 4 wave bands by utilizing the determined kernel function and the determined bandwidth;
and the multi-scale segmentation module is used for determining merging sequences of different scales and performing iterative operation on the reflectivity data of the high view satellite of 4 wave bands after the mean shift filtering to obtain each segmented characteristic unit.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911288831.2A CN111178175A (en) | 2019-12-12 | 2019-12-12 | Automatic building information extraction method and system based on high-view satellite image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911288831.2A CN111178175A (en) | 2019-12-12 | 2019-12-12 | Automatic building information extraction method and system based on high-view satellite image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111178175A true CN111178175A (en) | 2020-05-19 |
Family
ID=70657216
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911288831.2A Pending CN111178175A (en) | 2019-12-12 | 2019-12-12 | Automatic building information extraction method and system based on high-view satellite image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111178175A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116258958A (en) * | 2022-12-22 | 2023-06-13 | 二十一世纪空间技术应用股份有限公司 | Building extraction method and device for homologous high-resolution images and DSM data |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102254303A (en) * | 2011-06-13 | 2011-11-23 | 河海大学 | Methods for segmenting and searching remote sensing image |
| CN107194405A (en) * | 2017-05-03 | 2017-09-22 | 湖北省电力勘测设计院 | A kind of method that interactive semi-automatic high-resolution remote sensing image building is extracted |
| CN108596103A (en) * | 2018-04-26 | 2018-09-28 | 吉林大学 | High resolution ratio satellite remote-sensing image building extracting method based on optimal spectrum Index selection |
| CN109784320A (en) * | 2019-03-25 | 2019-05-21 | 中国科学院地理科学与资源研究所 | Ginseng industrialized agriculture domain determines method |
| CN109816707A (en) * | 2018-12-25 | 2019-05-28 | 中铁第四勘察设计院集团有限公司 | A kind of field of opencast mining information extracting method based on high-resolution satellite image |
| WO2019157348A1 (en) * | 2018-02-09 | 2019-08-15 | The Board Of Trustees Of The University Of Illinois | A system and method to fuse multiple sources of optical data to generate a high-resolution, frequent and cloud-/gap-free surface reflectance product |
-
2019
- 2019-12-12 CN CN201911288831.2A patent/CN111178175A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102254303A (en) * | 2011-06-13 | 2011-11-23 | 河海大学 | Methods for segmenting and searching remote sensing image |
| CN107194405A (en) * | 2017-05-03 | 2017-09-22 | 湖北省电力勘测设计院 | A kind of method that interactive semi-automatic high-resolution remote sensing image building is extracted |
| WO2019157348A1 (en) * | 2018-02-09 | 2019-08-15 | The Board Of Trustees Of The University Of Illinois | A system and method to fuse multiple sources of optical data to generate a high-resolution, frequent and cloud-/gap-free surface reflectance product |
| CN108596103A (en) * | 2018-04-26 | 2018-09-28 | 吉林大学 | High resolution ratio satellite remote-sensing image building extracting method based on optimal spectrum Index selection |
| CN109816707A (en) * | 2018-12-25 | 2019-05-28 | 中铁第四勘察设计院集团有限公司 | A kind of field of opencast mining information extracting method based on high-resolution satellite image |
| CN109784320A (en) * | 2019-03-25 | 2019-05-21 | 中国科学院地理科学与资源研究所 | Ginseng industrialized agriculture domain determines method |
Non-Patent Citations (2)
| Title |
|---|
| 林祚顶等: "《江淮丘陵区花山流域面源监测与污染物输出规律研究》", 31 December 2018 * |
| 沈占锋等: "遥感影像均值漂移分割算法的并行化实现", 《哈尔滨工业大学学报》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116258958A (en) * | 2022-12-22 | 2023-06-13 | 二十一世纪空间技术应用股份有限公司 | Building extraction method and device for homologous high-resolution images and DSM data |
| CN116258958B (en) * | 2022-12-22 | 2023-12-05 | 二十一世纪空间技术应用股份有限公司 | Building extraction method and device for homologous high-resolution images and DSM data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108573276B (en) | Change detection method based on high-resolution remote sensing image | |
| CN108596103B (en) | High-resolution remote sensing image building extraction method based on optimal spectral index selection | |
| Adeline et al. | Shadow detection in very high spatial resolution aerial images: A comparative study | |
| CN107610164B (en) | High-resolution four-number image registration method based on multi-feature mixing | |
| CN101901343B (en) | Remote sensing image road extracting method based on stereo constraint | |
| CN111814771B (en) | Image processing method and device | |
| CN111553922B (en) | Automatic cloud detection method for satellite remote sensing image | |
| CN107688777B (en) | Urban green land extraction method for collaborative multi-source remote sensing image | |
| CN109829423B (en) | Infrared imaging detection method for frozen lake | |
| CN105405138B (en) | Waterborne target tracking based on conspicuousness detection | |
| CN112287904B (en) | Airport target identification method and device based on satellite images | |
| CN112016478B (en) | Complex scene recognition method and system based on multispectral image fusion | |
| CN113255452A (en) | Extraction method and extraction system of target water body | |
| CN111008664A (en) | Hyperspectral sea ice detection method based on space-spectrum combined characteristics | |
| CN110738693B (en) | Multi-angle image registration method for ground-based imaging radar | |
| CN107103295B (en) | Optical remote sensing image cloud detection method | |
| Chaloeivoot et al. | Building detection from terrestrial images | |
| CN113705433A (en) | Power line detection method based on visible light aerial image | |
| CN111178175A (en) | Automatic building information extraction method and system based on high-view satellite image | |
| CN113902759B (en) | Space-spectrum information combined satellite-borne hyperspectral image segmentation and clustering method | |
| CN115457325A (en) | Object and spectrum rule oriented multi-spectral remote sensing image unsupervised classification method | |
| CN114565653A (en) | Heterogeneous remote sensing image matching method with rotation change and scale difference | |
| CN114792322A (en) | Method for detecting cloud and cloud shadow of mountain domestic high-resolution satellite image | |
| CN114445394A (en) | Method and system for detecting surface change | |
| Ye et al. | Research on Automatic Highway Extraction Technology Based on Spectral Information of Remote Sensing Images. |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210702 Address after: 310012 Room 101, Fudi 4 management room, Xixi National Wetland Park, Xihu District, Hangzhou City, Zhejiang Province Applicant after: Siwei Gaojing satellite remote sensing Co.,Ltd. Address before: 100094, Beijing, Yongfeng Haidian District industrial base, No. 5 East Feng Feng Road Applicant before: CHINA CENTRE FOR RESOURCES SATELLITE DATA AND APPLICATION |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200519 |