CN107507202A - A kind of vegetation rotary island towards high-resolution remote sensing image automates extracting method - Google Patents
A kind of vegetation rotary island towards high-resolution remote sensing image automates extracting method Download PDFInfo
- Publication number
- CN107507202A CN107507202A CN201710900510.8A CN201710900510A CN107507202A CN 107507202 A CN107507202 A CN 107507202A CN 201710900510 A CN201710900510 A CN 201710900510A CN 107507202 A CN107507202 A CN 107507202A
- Authority
- CN
- China
- Prior art keywords
- msub
- mrow
- image
- vegetation
- roundabout
- 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
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000000605 extraction Methods 0.000 claims abstract description 38
- 238000012706 support-vector machine Methods 0.000 claims abstract description 17
- 238000003709 image segmentation Methods 0.000 claims abstract description 15
- 230000011218 segmentation Effects 0.000 claims abstract description 15
- 230000004927 fusion Effects 0.000 claims abstract description 14
- 238000004422 calculation algorithm Methods 0.000 claims description 24
- 238000012937 correction Methods 0.000 claims description 13
- 230000000694 effects Effects 0.000 claims description 10
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 238000005520 cutting process Methods 0.000 claims description 8
- 238000012216 screening Methods 0.000 claims description 8
- 230000000007 visual effect Effects 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 claims description 5
- 238000003708 edge detection Methods 0.000 claims description 2
- 230000003595 spectral effect Effects 0.000 claims description 2
- 241001269238 Data Species 0.000 abstract 1
- 238000012545 processing Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- 238000001514 detection method Methods 0.000 description 2
- 230000005484 gravity Effects 0.000 description 2
- 239000004575 stone Substances 0.000 description 2
- 235000014443 Pyrus communis Nutrition 0.000 description 1
- 238000012952 Resampling Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000034373 developmental growth involved in morphogenesis Effects 0.000 description 1
- 238000001983 electron spin resonance imaging Methods 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000002310 reflectometry Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 239000011800 void material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/80—Geometric correction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20021—Dividing image into blocks, subimages or windows
-
- 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
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
-
- 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
- G06T2207/20112—Image segmentation details
- G06T2207/20152—Watershed segmentation
-
- 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
-
- 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
- G06T2207/30181—Earth observation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of vegetation rotary island towards high-resolution remote sensing image to automate extracting method, including four steps：The image segmentation of object-oriented and rule-based extraction, the classification of road based on SVMs, the extraction of road cavity, examine rotary island block and the topological relation in cavity.The image segmentation of the object-oriented of the present invention has two steps of segmentation and fusion, and segmentation has both of which：Edge and intensity, prioritizing selection edge pattern；Fusion has both of which：Full λ patterns or quick λ patterns, the full λ patterns of prioritizing selection；The classification of road based on SVMs of the present invention has four kinds of functions alternative for kernel function：Linearly, multinomial, RBF, S-shaped, prioritizing selection RBF.The present invention can effectively extract the rotary island comprising vegetation, have the characteristics that high-accuracy, strong adaptability, automation, independent of other assistance datas.
Description
Technical Field
The invention belongs to the technical field of visual processing, and relates to a method for automatically extracting a vegetation rotary island from a high-resolution multiband remote sensing image.
Background
The rotary island is a traffic facility which is usually located in a section with low traffic pressure and sufficient positions, and because the rotary island converts a junction of vehicles into a driving point, the rotary island can reduce the driving speed of the vehicles and reduce the occurrence of collision accidents, thereby improving the traffic quality. The radius of the rotary island is generally 12 to 30 meters, and varies with road grades, and some rotary islands can even become landmarks or leisure squares. The rotary island is often planted with vegetation, which is generally evergreen plants, flowers or low shrubs, and should have a sharp edge and a height that does not affect the driving of the driver.
The method can be used for extracting the vegetation rotary island to know the traffic condition, help municipal planning, perfect an electronic map for assisting manual or automatic driving, or measure the prosperity degree of a region, and has certain significance for urban traffic, design and humanistic research. The research on extracting the roundabout from the remote sensing image is less, and in 2009, the method for extracting the roundabout by taking a geographic space database as priori knowledge is proposed by Ravanbakhsh, so that the research on extracting the roundabout is developed. However, the existing methods for extracting the rotary island still have many disadvantages, such as low extraction accuracy, poor utilization of vegetation information in remote sensing images, or the need to rely on other auxiliary data.
Therefore, it is a difficult problem to overcome in the art to propose a method for automatically extracting a vegetation rotary island from a high-resolution multiband remote sensing image.
Disclosure of Invention
In order to solve the technical problem, the invention provides a high-resolution remote sensing image-oriented automatic extraction method of a vegetation rotary island.
The technical scheme of the invention is as follows: an algorithm for automatically extracting a vegetation rotary island from a high-resolution multiband remote sensing image comprises the following steps:
step 1: executing the flow A and the flow B in parallel;
scheme A: carrying out image segmentation on an original image, and extracting a roundabout and a roundabout block;
and (B) a process: carrying out road classification on the original image, and extracting road cavities and cavity blocks;
step 2: checking the topological relation between the ring island blocks and the holes;
and step 3: and acquiring a roundabout extraction result.
In the invention, two watershed algorithms and two fusion modes can be selected for image segmentation, and four kernel functions can be selected for classification by a support vector machine, so that the method is flexible; the method has no higher requirement on the processed image, can adapt to high-resolution multiband remote sensing images with different sizes, different cutting, different resolutions and different waveband configurations, and has strong adaptability and robustness; the extraction result of the invention has higher accuracy and higher coincidence rate with the result of manual visual interpretation; the data processing time of the invention is short, and a large amount of roundabout extractions can be processed in a short time; the invention fully utilizes the information of the image, fully considers the vegetation information of the ground objects, does not use other auxiliary knowledge or data, and has strong independence; the method can still carry out rough topological relation judgment under the condition of non-ideal data condition or error caused by the algorithm, has certain fault-tolerant rate, and ensures the reliability of the extraction result. The invention improves and innovates the algorithm for extracting the land feature of the roundabout from the remote sensing image, reduces the cost for the automatic identification and extraction task of the roundabout, increases the stability and the accuracy rate, and has positive promotion effect on the related research of the extraction of the remote sensing land feature.
Drawings
FIG. 1 is a simplified flow diagram of an embodiment of the present invention;
FIG. 2 is a simplified diagram of a pretreatment process according to an embodiment of the present invention;
FIG. 3 is a simplified process diagram of object-oriented image segmentation according to an embodiment of the present invention;
FIG. 4 is a simplified rule-based extraction flow diagram according to an embodiment of the present invention;
FIG. 5 is a simplified flowchart of a road classification based on a support vector machine according to an embodiment of the present invention;
FIG. 6 is a simplified flowchart of hole detection according to an embodiment of the present invention;
Detailed Description
In order to facilitate the understanding and implementation of the present invention for those of ordinary skill in the art, the present invention is further described in detail with reference to the accompanying drawings and examples, it is to be understood that the embodiments described herein are merely illustrative and explanatory of the present invention and are not restrictive thereof.
The invention relates to an algorithm for automatically extracting a vegetation rotary island from a high-resolution multiband remote sensing image, which comprises the following four steps: the method comprises the steps of object-oriented image segmentation and rule-based extraction, road classification based on a support vector machine, road void extraction and detection of topological relation between ring island blocks and voids. Before the first step, there may be appropriate image pre-processing; after the fourth step, the extraction results can be compared with results of manual visual interpretation to evaluate the outcome of the automated algorithm. The first step may be in parallel with the second and third steps, the second step having to precede the third step and the fourth step having to be performed after the first, second and third steps have been completed, as shown in fig. 1.
The preprocessing refers to processing such as geometric correction, radiometric calibration, sky-ground map registration, cutting and the like on the original remote sensing image, so that the quality of the original remote sensing image is improved as much as possible, and the accuracy of the result of the method is expected to be improved from a data source. In order, there is no specific precedence between geometric correction and radiometric calibration, and the registration of the celestial map and the terrestrial map must be performed after the first two are completed, and the cropping must be performed after the first three are completed, as shown in fig. 2. Geometric correction means eliminating or reducing geometric deformation of the original image; radiometric calibration is to convert dimensionless DN value recorded by the sensor into atmospheric top layer radiance or reflectivity with actual physical meaning; the heaven and earth map registration is to register the geometric correction and radiometric calibration image with a standard and orthoscopic world map; the cutting is a process of cutting a large remote sensing image into a small remote sensing image containing one or more complete rotary islands in order to reduce processing time. The whole preprocessing is not necessary, but the proper preprocessing can ensure the accuracy of the extraction result of the invention in a certain range and can reduce the processing time of the data.
The specific implementation scheme of the embodiment is as follows:
and (3) geometric correction: in ENVI5.3sp software, high-resolution second-order original image data and metadata thereof (other high-resolution multiband images can also be selected), an RPCOrthoresis Workflow tool under Orthovisual under Geometric Correction is selected, the size of an output pixel is selected to be 4m, a resampling mode is selected to be a cubic equation, and Geometric Correction can be carried out according to set parameters after setting of an output path and the like is completed.
Radiation calibration: in ENVI5.3sp software, opening high-score second-order original image data or data after geometric Correction, selecting a Radiometric Calibration tool under Radiometric Correction, setting a Calibration type as Radiance data Radiance, automatically setting a data type required by a FLAASH atmospheric Correction tool by the ENVI software, wherein the storage sequence is BIL or BIP, the data type is FLOAT, setting a unit adjustment coefficient of Radiance data to be 0.1, and implementing Radiometric Calibration according to set parameters after finishing related setting.
The heaven and earth map registration can use ArcMap10.2 in ArcGIS 10.2 software to call Georeference function, and several pairs of obvious artificial objects are selected to be registered by clicking (standard world map http:// www.scgis.net.cn/imap/iMapServer/defaultRest/services/newtienditudom/WMS) aiming at the area shot by the image of the orthoscopic standard world map.
Cutting can be completed by using a resize function in the envi5.3sp software and specifying the position of a back corner point of cutting and the size of cutting (such as coordinates 114.468 and 30.471 at the top left corner, and the size is 400 pixels by 400 pixels).
The first formal step of the invention: and (3) object-oriented image segmentation, namely segmenting the whole high-resolution multiband remote sensing image by using a watershed algorithm, wherein the segmentation result is an image block, and extracting and processing the image block based on rules aiming at a plurality of image blocks. The watershed algorithm is derived from the concept of hydrology, that is, as the water level rises, water bodies in different elevation areas are converged, and watersheds are built at the convergence positions of the water bodies in the different elevation areas. In the field of digital image processing, a watershed algorithm is a commonly used image segmentation algorithm, and an ideal segmentation result should be consistent with an object in the real world, that is, under an ideal condition, one segmentation block should correspond to an entity in the real world, which is an embodiment of an object-oriented idea. Watershed segmentation includes two steps of segmentation and fusion, as shown in fig. 3.
The segmentation has two optional fusion modes: the edge mode is to perform edge detection of a sobel operator on the original image, then form a gradient map and perform segmentation of a watershed algorithm, and the intensity mode is to select a partial wave band from the original image, take a mean value in a certain range to generate an intensity map and perform segmentation of the watershed algorithm.
The fusion is to judge each divided image block according to the correlation degree between the blocks and then to perform fusion in order to avoid the problem of excessive division. Fusion has two fusion modes of choice: full λ mode or fast λ mode; the full lambda mode formula is as follows, if the merging cost t of two pixelsi,jIf the value is less than the threshold value, merging:
wherein: o isiIs the area i in the image;
|Oii is the area of region i;
μiis the mean of region i;
||μi-μji is the Euclidean distance of the spectral values of the region i and the region j;
is OiAnd OjThe length of the common boundary;
the fast lambda mode formula is as follows, the smaller the lambda value is, the smaller the Euclidean color distance is, the larger the public boundary length is, the more the combination is needed, if the lambda value is smaller than the threshold value, the combination is needed;
wherein: n1 is the number of pixels in region 1;
e is the euclidean distance of the region 1 and the region 2;
l is the common boundary length of region 1 and region 2.
The first step of rule-based extraction is to consider the obvious characteristics of the vegetation rotary island simultaneously for the segmentation result: size, circularity and vegetation index set up the three's comprehensive consideration as the rule, promptly: the size of the circle is within the range of the upper limit and the lower limit of the circle area with the radius of the roundabout specified by the traffic department of the country and the region to which the image belongs; the roundness (the formula is as follows) is within a certain range close to 1; the vegetation index uses NDVI (normalized differential growth index, formula below) and the NDVI value of the image block where the vegetation ring island is located should not be less than the first 40% of the NDVI value of the whole image, as shown in fig. 4. After the three attributes of the rule are set, screening all image blocks according to the rule; the image blocks left after screening should be as many as one and only one, namely, the ring island blocks.
Wherein: r is the roundness of the region;
s is the area of the region;
c is the perimeter of the region;
wherein: NDVI is the normalized vegetation index;
eps is a constant small enough to avoid denominators other than 0;
b1 is the infrared band;
b2 is the near infrared band;
the specific implementation scheme of the embodiment is as follows:
selecting a Rule Based Feature Extraction Workflow tool in ENVI for the preprocessed image, checking a Normalized difference option, setting a third wave band (red wave band) of the image as band 1, and setting a fourth wave band (infrared wave band) of the image as band 2, thereby participating in the calculation of NDVI. Next, entering image segmentation, where it is required to set a watershed algorithm mode and parameters of image segmentation, and a fusion mode and parameters, respectively, which can be set as: "edge mode, degree 50, all λ mode, 0".
The rules are then set. Because the rules in the invention comprise three items of size, roundness and vegetation index, the rules are created firstly, then the created rules are clicked, three attributes are created, and the attributes can be respectively set as follows: "spatial attribute-size: 452 + 2826; spatial property-roundness: 0.5-1.2; band attribute-normalized index (NDVI): 60% -100% ". The method is operated according to rules, screening is carried out, the screened result is checked, and two aspects are mainly concerned: whether a roundabout is left after rule-based screening; and screening whether redundant non-circular island blocks exist. If the former is yes and the latter is no, object-oriented image segmentation and rule-based extraction are completed. The extracted result can be converted into evf and then converted into shp file for subsequent operation. It should be noted that the shape data format (shp file) of the ESRI is not the only feasible data format of the present invention, and varies depending on the specific operating software and production requirements.
The road classification based on the Support Vector Machine is to use a Support Vector Machine (SVM) algorithm to classify images after selecting a proper number of road sample points. The SVM is a common algorithm for pattern recognition, and the idea is as follows: for the linear inseparable condition, the nonlinear mapping algorithm is used to convert the linear inseparable sample of the low-dimensional input space into the high-dimensional feature space (hyperplane) to make the linear inseparable sample, so that the linear analysis of the nonlinear feature of the sample by the linear algorithm in the high-dimensional feature space becomes possible.
The road sample points can be manually selected in ENVI software by a mouse to form ROI (regions of interest) to obtain road pixel samples, and classification samples can also be designated by other modes. The classified kernel function has four alternatives, namely linear, polynomial, radial basis function and sigmoid (formula is shown below), and the classified kernel function is output as a classification result containing the banded entity. It is suggested that the radial basis function can be chosen to be a kernel function if there are no special requirements, as it works better in most cases. Three parameters required by the SVM algorithm are also set: pyramid level, penalty parameter, classification probability threshold (as shown in fig. 5). The pyramid level number is the number of hierarchical processing levels applied in the support vector machine to avoid classification or over-classification. The multiplication parameters allow a certain degree of misjudgment, theoretically allow some training points to be in the hyperplane of the wrong side, and therefore the support vector machine model has strong fault tolerance and flexibility. The classification probability threshold represents the confidence that the closest partial points represent the class in the hyperplane, so a higher threshold results in fewer closest points being classified and an increase in unclassified portions in the image. In short, the classification probability threshold may control the degree of classification.
In the present invention, a satisfactory SVM-based road classification result is a continuous strip-like entity, i.e., a strip-like road, in which there should be a large hole close to a circle. Since the large hole should be due to the roundabout, the spatial position of the hole and roundabout should substantially coincide. Slight errors are difficult to avoid, such as fine broken holes and uneven edges of road extraction results caused by shadow of vehicles or road trees, and the phenomenon that some gaps or superposition exists on the edges of the holes and the roundabout. The present invention is tolerant to these unavoidable errors, which are reflected in the next two steps.
Linearity: k (x)i,xj)＝xi Txj
Polynomial expression: k (x)i,xj)＝(gxi Txj+r)d
Radial basis function: k (x)i,xj)＝exp(-g||xi-xj||2)
S-shaped: k (x)i,xj)＝tanh(gxi Txj+r)
Wherein: g is the gamma parameter required in functions other than the linear kernel function;
d is a parameter required for the polynomial kernel;
r is the bias term parameter required for the polynomial and sigmoid kernel functions.
The specific implementation scheme of the embodiment is as follows:
and opening the image to be classified, selecting a SupportVector Machine tool of Supervised under Classication in the ENVI, and performing SVM Classification. The following settings may be made: selecting a radial basis function (radialbasifunction), setting gamma to 0.25, a penalty parameter (penalty parameter) to 100, a pyramid level (pyramid level) to 0, and a classification probability threshold (classification probability threshold) to 0.7. It should be noted that the classification effect of these settings varies from image to image, and sometimes may require multiple appropriate adjustments to achieve the desired effect.
And extracting road holes, namely extracting only holes in the strip road, and regarding the largest hole as a hole caused by the roundabout. This is because it is normally difficult for the SVM-based classification effect of roads to produce extraction errors with radii in excess of 12 meters, i.e., errors larger in area than islands.
The hole extraction comprises four steps, each step is completed by a GIS (geographic information system) tool, and the three steps are conversion; firstly, converting polygons into points, namely converting strip-shaped road polygons into central points, wherein the attributes of the central points comprise a number FID, a Shape, a type _ Name, a type _ number, a Class _ ID, a block number part, a Length, an Area and a source number ORIG _ FID; secondly, converting polygons into lines, namely converting strip-shaped road polygons into boundary lines, wherein the boundary lines only have a small number of attributes, and the boundary line attributes comprise a number FID, a Shape, a type _ Name, a type _ number, a type _ ID, a block number part, a Length, an Area and a source element ID; thirdly, the conversion line is a polygon, namely, a polygon is generated according to the boundary line, and the polygon attribute comprises a number FID, a Shape, a source element ID, a type _ number Class _ ID, a block number part, a Length, an Area and a source number ORIG _ FID; adding the attributes of the points generated in the first step into the polygon generated in the third step according to the ID or other marks of each polygon, wherein if the source polygon does not exist, the additional attributes are null; fourthly, a polygon with the type _ number Class _ ID, the block number Parts, the Length, the Area and the source element ID all being 0 is a hole; finally, the holes are sorted according to area, and the largest road hole, namely the road hole caused by the roundabout, is selected and called a hole block (the flow is shown in fig. 6).
That is, the extraction of the road holes utilizes the phenomena of loss and retention of attributes in various conversions of the GIS, and the holes are screened out; since the empty hole is not a single polygon entity at first, after three-step conversion, a polygon having the same shape as the empty hole is generated from the boundary of the empty hole, and there is naturally a difference in the integrity and the number of items in the attribute compared to all the polygons before conversion.
The specific implementation scheme of the embodiment is as follows:
in ArcMap10.2, "feature to points", "polygon to lines", and "feature to polygons" can be used sequentially to complete the conversion of polygons into points, polygons into lines, and polygons into polygons. It should be emphasized that, in the third step of conversion operation, the points converted in the first step are used as label features, and preservettibutes option is selected, so that the attributes obtained in the first step can be added to the polygon generated in the third step, and then all holes can be found. Finally, all the polygons shapefile generated in the third step are selected, the polygons with the lost attributes such as classname, area and original fid are found by utilizing attribute-based query, and then the sorting button is clicked to sort according to the size. And selecting the polygon with the largest area, and storing the polygon with the largest area in a shp format, wherein the polygon is called a hole block.
The topological relation of the ring island blocks and the cavities is checked, the purpose is to hopefully tolerate the incomplete overlapping phenomenon of the ring island blocks and the cavities caused by a plurality of errors, and allow gaps or overlapping of the ring island blocks and the boundaries of the cavities, namely, the phenomenon that the relation of roads and the ring island blocks is not the inclusion relation under an ideal state is contained, but at the same time, the topological relation of the ring island blocks and the roads is still checked roughly, and a basis is provided for judging whether the ring island blocks are the ring islands. The specific method comprises the steps of respectively calculating the central points of the roundabout block and the hollow, measuring the distance between the two central points, comparing the distance with a certain threshold, judging that the extracted roundabout block is the roundabout if the distance is smaller than the threshold, and extracting inefficacy if the distance is larger than the threshold.
The specific implementation scheme of the embodiment is as follows:
in arcmap10.2, the center of gravity, or called center, of the ring island block and the hole block, shape file extracted in the previous step is obtained by using "feature to points". The center of gravity is calculated by taking the mean of the two-dimensional coordinates of all the boundary points of the entity. The two centers are displayed in different ways, such as red triangles and black dots. The range of the two points was measured again using the Measure tool of ArcMap. A rough threshold can be set with reference to the size and resolution of the remote sensing image, in combination with the general width of the road. If the original remote sensing image is 7300 × 6908 pixels, the cut remote sensing image is 400 × 400 pixels, each pixel represents 4 × 4 meters of actual ground area (the ground resolution is 4m × 4m), the region where the to-be-judged roundabout is yellow stone in Hubei province of China, the road width is about 8-30 meters, and the threshold value can be set to be 5 meters by combining the factors. And comparing the distance with a threshold value (5m), making a corresponding judgment, and roughly judging whether the approximate topological relation between the roundabout block and the road cavity is coincident or not, namely whether the extracted roundabout block can be regarded as a real roundabout or not. The specific value of the threshold is not limited too much, as long as the characteristic of judging the extraction effect of the roundabout from the distance between two points can be embodied, and the extraction effect of the roundabout can be roughly judged.
The best method for testing the extraction effect of the roundabout is to count and comprehensively analyze the distances between a plurality of extracted roundabout blocks and a cavity block for many times and compare the plurality of extracted roundabout blocks with the effect of manual visual interpretation. Table 1 shows the statistical results of distances and the comparison results with visual interpretation in experiments of 10 islands in Wuhan, Huangshi and Ezhou, Hubei province, China. Table 1 shows that, in the case of a ground resolution of 4m × 4m or 3.7m × 3.7m, the distance between the two central points is less than 4m, i.e. about one pixel side length, the rough topology judgment can play a good role, and the rough topology judgment shows robustness to different data sources, and the extraction effect can be considered as relatively precise. Table 1 also shows that the production accuracy (the percentage of the number of pixels labeled as islands by both automation and visual interpretation to the number of pixels labeled as islands by visual interpretation) was over 85% in 10 experiments, with a reasonably good extraction accuracy, and with a significant accuracy advantage over existing algorithms. Table 2 additionally shows the positions, area sizes, circularities, and NDVI of 10 roundabouts, the roundabouts are substantially within an expected range, and meet the requirements of the transportation department, the circularities are substantially between 0.4 and 1.0, and meet the geometric rules and the concept of the present invention, and the NDVI basically has a higher numerical value in the whole or cut image where the roundabouts are located, so that it can be seen that the concept of rule-based screening is very correct and effective. The results can prove that the algorithm result for automatically extracting the vegetation rotary island from the high-resolution multiband remote sensing image can be used and accurate, the scheme for automatically extracting the vegetation rotary island from the high-resolution multiband remote sensing image is feasible, and the method has the advantages of high accuracy, strong adaptability, automation and no dependence on other auxiliary data.
| GTR(m) | Distance(m) | Producer Accuracy | |
| 1 | 3.7 | 3.036 | 91.67％ |
| 2 | 3.7 | 3.271 | 86.02％ |
| 3 | 3.7 | 3.409 | 85.71％ |
| 4 | 3.7 | 2.308 | 92.63％ |
| 5 | 3.7 | 3.821 | 87.37％ |
| 6 | 3.7 | 3.286 | 90.32％ |
| 7 | 4 | 0.858 | 95.47％ |
| 8 | 4 | 2.461 | 94.06％ |
| 9 | 4 | 2.604 | 96.67％ |
| 10 | 4 | 2.486 | 94.74％ |
TABLE 1 ground resolution (GTR), center-point Distance (Distance) and production Accuracy (Producer Accuracy) for ten experiments
| ID | Coordinates | Rd.Name | Size(m2) | Roundness | NDVI |
| 1 | 114.368,30.571 | Oriole road | 555.4803 | 0.584201 | 0.044186 |
| 2 | 114.377,30.569 | Along lake | 1353.093 | 0.473994 | 0.122496 |
| 3 | 114.255,30.626 | Evergreen road | 897.3143 | 0.710775 | 0.072343 |
| 4 | 115.059,30.255 | Yellow stone road | 1623.712 | 0.737449 | 0.094100 |
| 5 | 114.378,30.580 | Pear garden road | 1285.934 | 0.816796 | 0.117352 |
| 6 | 115.097,30.211 | Yiyang Lu | 1373.633 | 0.764884 | 0.035440 |
| 7 | 114.825,30.417 | Four seas road | 3926.975 | 0.908119 | 0.229562 |
| 8 | 114.920,30.398 | Singer road | 1600.000 | 0.810996 | 0.239780 |
| 9 | 114.882,30.391 | Dongfu Wulu (Dongfu Wulu) | 2432.000 | 0.855590 | 0.207698 |
| 10 | 114.882,30.367 | Guanliulu (a Chinese character of 'guanliulu') | 1104.000 | 0.820611 | 0.166888 |
Table 2 roundabout Coordinates (Coordinates), name (rd. name), Size (Size), Roundness (Roundness) and vegetation index (NDVI) of ten experiments
It should be understood that parts of the specification not set forth in detail are well within the prior art.
It should be understood that the above description of the preferred embodiments is given for clarity and not for any purpose of limitation, and that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (8)
1. A vegetation rotary island automatic extraction method facing a high-resolution remote sensing image is characterized by comprising the following steps:
step 1: executing the flow A and the flow B in parallel;
scheme A: carrying out image segmentation on an original image, and extracting a roundabout and a roundabout block;
and (B) a process: carrying out road classification on the original image, and extracting road cavities and cavity blocks;
step 2: checking the topological relation between the ring island blocks and the holes;
and step 3: and acquiring a roundabout extraction result.
2. The automatic extraction method of the vegetation rotary island facing to the high-resolution remote sensing image according to claim 1, characterized in that: in the step 1, carrying out image segmentation on the whole high-resolution multiband remote sensing image by using a watershed algorithm, wherein the segmentation result is an image block;
the watershed algorithm comprises two steps of segmentation and fusion;
there are two fusion modes of segmentation: edge or strength; the edge mode is to perform edge detection of sobel operator on the original image, then form a gradient image and perform segmentation of watershed algorithm; the intensity mode is that aiming at the original image, part of wave bands are selected from the original image, and an average value in a certain range is taken to generate an intensity map;
there are two fusion modes of fusion: full λ mode or fast λ mode;
the full λ mode formula is as follows:
<mrow> <msub> <mi>t</mi> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> </msub> <mo>=</mo> <mfrac> <mrow> <mfrac> <mrow> <mo>|</mo> <msub> <mi>O</mi> <mi>i</mi> </msub> <mo>|</mo> <mo>&CenterDot;</mo> <mo>|</mo> <msub> <mi>O</mi> <mi>j</mi> </msub> <mo>|</mo> </mrow> <mrow> <mo>|</mo> <msub> <mi>O</mi> <mi>i</mi> </msub> <mo>|</mo> <mo>+</mo> <mo>|</mo> <msub> <mi>O</mi> <mi>j</mi> </msub> <mo>|</mo> </mrow> </mfrac> <mo>&CenterDot;</mo> <mo>|</mo> <mo>|</mo> <msub> <mi>&mu;</mi> <mi>i</mi> </msub> <mo>-</mo> <msub> <mi>&mu;</mi> <mi>j</mi> </msub> <mo>|</mo> <msup> <mo>|</mo> <mn>2</mn> </msup> </mrow> <mrow> <mi>l</mi> <mi>e</mi> <mi>n</mi> <mi>g</mi> <mi>t</mi> <mi>h</mi> <mrow> <mo>(</mo> <mo>&part;</mo> <mo>(</mo> <mrow> <msub> <mi>O</mi> <mi>i</mi> </msub> <mo>,</mo> <msub> <mi>O</mi> <mi>j</mi> </msub> </mrow> <mo>)</mo> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>;</mo> </mrow>
wherein: o isiIs the area i in the image; i OiI is the area of region i; mu.siIs the mean of region i; | mu |i-μjI is the Euclidean distance of the spectral values of the region i and the region j;is OiAnd OjThe length of the common boundary; if the merging cost t of two pixelsi，jIf the sum is less than the threshold value, merging;
the fast λ mode formula is as follows:
<mrow> <mi>L</mi> <mi>a</mi> <mi>m</mi> <mi>b</mi> <mi>d</mi> <mi>a</mi> <mo>=</mo> <mo>&lsqb;</mo> <mfrac> <mrow> <msub> <mi>N</mi> <mn>1</mn> </msub> <mo>&CenterDot;</mo> <msub> <mi>N</mi> <mn>2</mn> </msub> </mrow> <mrow> <msub> <mi>N</mi> <mn>1</mn> </msub> <mo>+</mo> <msub> <mi>N</mi> <mn>2</mn> </msub> </mrow> </mfrac> <mo>&rsqb;</mo> <mfrac> <mi>E</mi> <mi>L</mi> </mfrac> <mo>;</mo> </mrow>
wherein: n1 is the number of pixels in region 1; e is the euclidean distance of the region 1 and the region 2; l is the common boundary length of region 1 and region 2; the smaller the lambda value is, the smaller the Euclidean color distance is, the larger the common boundary length is, the more should be merged, and if the lambda value is smaller than the threshold value, the merging is performed.
3. The automatic extraction method of the vegetation rotary island facing to the high-resolution remote sensing image according to claim 1, characterized in that: in the step 1, based on the image segmentation result, the size, the roundness and the vegetation index of the image block are considered at the same time, and the setting rule of the three are considered comprehensively: the size of the circle is within the range of the upper limit and the lower limit of the circle area with the radius of the roundabout specified by the traffic department of the country and the region to which the image belongs; the roundness R is within a predetermined range close to 1; the vegetation index is normalized vegetation index NDVI and the NDVI value of the image block where the vegetation rotary island is located is not lower than the first 40 percent of the NDVI value of the whole image; screening all image blocks according to a rule; the screening result is a roundabout block;
wherein,r is the roundness of the region, S is the area of the region, C is the perimeter of the region;NDVI is the normalized vegetation index, eps is a constant small enough to avoid denominators other than 0, b1 is the infrared band, b2 is the near infrared band.
4. The automatic extraction method of the vegetation rotary island facing to the high-resolution remote sensing image according to claim 1, characterized in that: in step 1, after selecting a proper number of road sample points, using a support vector machine algorithm to classify images, wherein classified kernel functions comprise linear kernel functions, polynomial kernel functions, radial basis kernel functions and S-shaped kernel functions, and output a classification result containing a banded entity, wherein the banded entity is a banded road and should contain a cavity which accords with the size of a roundabout;
linear kernel function: k (x)i，xj)＝xi Txj，
Polynomial kernel function: k (x)i，xj)＝(gxi Txj+r)d，
Radial basis kernel function: k (x)i，xj)＝exp(-g||xi-xj||2)，
Sigmoid kernel function: k (x)i，xj)＝tan h(gxi Txj+r)，
Wherein: g is a parameter required in functions other than the linear kernel function; d is a parameter required for the polynomial kernel; r is the bias term parameter required for the polynomial and sigmoid kernel functions.
5. The automatic extraction method of the vegetation rotary island facing to the high-resolution remote sensing image according to claim 1, characterized in that: in the step 1, only extracting the holes in the strip road, and regarding the largest hole as a hole caused by a roundabout;
the cavity extraction comprises four steps:
the first step is as follows: converting the polygon into points;
converting a strip-shaped road polygon into a central point, wherein the central point attribute comprises a number FID, a Shape, a type _ Name, a type _ number Class _ ID, a block number part, a Length, an Area and a source number ORIG _ FID;
the second step is that: converting the polygon into a line;
converting a strip-shaped road polygon into a boundary line, wherein the boundary line attribute comprises a number FID, a Shape, a type _ Name, a type _ number, a Class _ ID, a block number part, a Length, an Area and a source element ID;
the third step: the conversion line is a polygon;
generating a polygon according to the boundary line, wherein the polygon attribute comprises a number FID, a Shape, a number file FID _ shp, a type _ number Class _ ID, a block number part, a Length, an Area and a source element ID; adding the attributes of the points generated in the first step into the polygon generated in the third step according to the ID or other marks of each polygon, wherein if the source polygon does not exist, the additional attributes are null;
the fourth step: a polygon with type _ number Class _ ID, block number Parts, Length, Area and source element ID all being 0 is a hole; the holes are sorted according to area, and the largest hole block is selected.
6. The automatic extraction method of the vegetation rotary island facing to the high-resolution remote sensing image according to claim 1, characterized in that: in step 2, the central points of the roundabout block and the hollow block are respectively calculated, the distance between the two central points is measured, the distance is compared with a certain threshold, if the distance is smaller than the threshold, the extracted roundabout block is judged to be the roundabout, and if the distance is larger than the threshold, the extraction is invalid.
7. The method for automatically extracting the vegetation rotary island facing the high-resolution remote sensing image according to any one of claims 1 to 6, wherein the method comprises the following steps: in the step 1, firstly, preprocessing is carried out on an original image, and then image segmentation or road classification is carried out;
the pretreatment specifically comprises the following substeps:
step A1: judging whether to carry out geometric correction on the original image or not;
if so, performing geometric correction on the original image, and then performing radiometric calibration;
if not, firstly carrying out radiometric calibration and then carrying out geometric correction on the original image;
step A2: registering the geometrically corrected and radiometric images with a standard, orthoscopic world map;
step A3: and cutting the processed remote sensing image into a plurality of small remote sensing images containing one or more complete rotary islands.
8. The method for automatically extracting the vegetation rotary island facing the high-resolution remote sensing image according to any one of claims 1 to 6, wherein the method comprises the following steps: and (3) counting and comprehensively analyzing the distances between the plurality of extracted roundabout blocks and the cavity block for a plurality of times according to the roundabout extraction result obtained in the step (3), and comparing the plurality of extracted roundabout blocks with the effect of artificial visual interpretation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710900510.8A CN107507202A (en) | 2017-09-28 | 2017-09-28 | A kind of vegetation rotary island towards high-resolution remote sensing image automates extracting method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710900510.8A CN107507202A (en) | 2017-09-28 | 2017-09-28 | A kind of vegetation rotary island towards high-resolution remote sensing image automates extracting method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107507202A true CN107507202A (en) | 2017-12-22 |
Family
ID=60700087
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710900510.8A Pending CN107507202A (en) | 2017-09-28 | 2017-09-28 | A kind of vegetation rotary island towards high-resolution remote sensing image automates extracting method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107507202A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109241210A (en) * | 2018-07-09 | 2019-01-18 | 广东省农业科学院农业经济与农村发展研究所 | A kind of general banking process of Regional agriculture planning spatial data |
| CN112036233A (en) * | 2020-07-13 | 2020-12-04 | 国家海洋信息中心 | Ice lake extraction method integrating threshold segmentation and watershed transformation algorithms |
| CN112559787A (en) * | 2020-12-31 | 2021-03-26 | 广州智云尚大数据科技有限公司 | Telemetry data processing method and system based on block chain and cloud platform |
| CN112597939A (en) * | 2020-12-29 | 2021-04-02 | 中国科学院上海高等研究院 | Surface water body classification extraction method, system, equipment and computer storage medium |
| WO2021243576A1 (en) * | 2020-06-02 | 2021-12-09 | Huawei Technologies Co., Ltd. | Apparatus, system and method for detecting speed bumps and potholes on a road |
| CN118097069A (en) * | 2024-04-29 | 2024-05-28 | 中铁四局集团有限公司 | Vegetation filtering method and device for mountain terrain grid model and computer equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102013015A (en) * | 2010-12-02 | 2011-04-13 | 南京大学 | Object-oriented remote sensing image coastline extraction method |
| CN103034863A (en) * | 2012-12-24 | 2013-04-10 | 重庆市勘测院 | Remote-sensing image road acquisition method combined with kernel Fisher and multi-scale extraction |
| CN106778605A (en) * | 2016-12-14 | 2017-05-31 | 武汉大学 | Remote sensing image road net extraction method under navigation data auxiliary |
-
2017
- 2017-09-28 CN CN201710900510.8A patent/CN107507202A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102013015A (en) * | 2010-12-02 | 2011-04-13 | 南京大学 | Object-oriented remote sensing image coastline extraction method |
| CN103034863A (en) * | 2012-12-24 | 2013-04-10 | 重庆市勘测院 | Remote-sensing image road acquisition method combined with kernel Fisher and multi-scale extraction |
| CN106778605A (en) * | 2016-12-14 | 2017-05-31 | 武汉大学 | Remote sensing image road net extraction method under navigation data auxiliary |
Non-Patent Citations (2)
| Title |
|---|
| XINYI LI 等: "Automatic Object-Oriented Roundabouts Extrction from High Resolution Multispectral Images", 《INTERNATIONAL ARCHIVES OF THE PHOTOGRAMMETRY, REMOTE SENSING AND SPATIAL INFORMATION SCIENCES》 * |
| 张明华 等: "基于CBERS影像的厦门环岛路沙滩信息提取研究", 《厦门理工学院学报》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109241210A (en) * | 2018-07-09 | 2019-01-18 | 广东省农业科学院农业经济与农村发展研究所 | A kind of general banking process of Regional agriculture planning spatial data |
| WO2021243576A1 (en) * | 2020-06-02 | 2021-12-09 | Huawei Technologies Co., Ltd. | Apparatus, system and method for detecting speed bumps and potholes on a road |
| CN112036233A (en) * | 2020-07-13 | 2020-12-04 | 国家海洋信息中心 | Ice lake extraction method integrating threshold segmentation and watershed transformation algorithms |
| CN112036233B (en) * | 2020-07-13 | 2023-04-28 | 国家海洋信息中心 | Ice lake extraction method integrating threshold segmentation and watershed transformation algorithm |
| CN112597939A (en) * | 2020-12-29 | 2021-04-02 | 中国科学院上海高等研究院 | Surface water body classification extraction method, system, equipment and computer storage medium |
| CN112597939B (en) * | 2020-12-29 | 2023-11-07 | 中国科学院上海高等研究院 | Surface water body classification extraction method, system, equipment and computer storage medium |
| CN112559787A (en) * | 2020-12-31 | 2021-03-26 | 广州智云尚大数据科技有限公司 | Telemetry data processing method and system based on block chain and cloud platform |
| CN118097069A (en) * | 2024-04-29 | 2024-05-28 | 中铁四局集团有限公司 | Vegetation filtering method and device for mountain terrain grid model and computer equipment |
| CN118097069B (en) * | 2024-04-29 | 2024-07-09 | 中铁四局集团有限公司 | Vegetation filtering method and device for mountain terrain grid model and computer equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107507202A (en) | A kind of vegetation rotary island towards high-resolution remote sensing image automates extracting method | |
| Jin et al. | An integrated system for automatic road mapping from high-resolution multi-spectral satellite imagery by information fusion | |
| Zhang et al. | Automated delineation of agricultural field boundaries from Sentinel-2 images using recurrent residual U-Net | |
| Ünsalan et al. | A system to detect houses and residential street networks in multispectral satellite images | |
| CN103400151B (en) | The optical remote sensing image of integration and GIS autoregistration and Clean water withdraw method | |
| Münzinger et al. | Mapping the urban forest in detail: From LiDAR point clouds to 3D tree models | |
| Matkan et al. | Road extraction from lidar data using support vector machine classification | |
| CN109389163B (en) | Unmanned aerial vehicle image classification system and method based on topographic map | |
| Sun et al. | Information fusion for rural land-use classification with high-resolution satellite imagery | |
| CN109146889A (en) | A kind of field boundary extracting method based on high-resolution remote sensing image | |
| CN107330422A (en) | A kind of method for carrying out mima type microrelief classification to semiarid zone based on high accuracy number elevation model | |
| CN111028255A (en) | Farmland area pre-screening method and device based on prior information and deep learning | |
| CN104778679A (en) | Gaofen-1 satellite data-based control point graphic element rapid-matching method | |
| Ok et al. | 2-D delineation of individual citrus trees from UAV-based dense photogrammetric surface models | |
| Bengoufa et al. | Rocky shoreline extraction using a deep learning model and object-based image analysis | |
| Helmholz et al. | Semi-automatic verification of cropland and grassland using very high resolution mono-temporal satellite images | |
| CN115272860A (en) | Method and system for determining rice planting area, electronic device and storage medium | |
| Jing et al. | Island road centerline extraction based on a multiscale united feature | |
| CN111882573A (en) | Cultivated land plot extraction method and system based on high-resolution image data | |
| CN116612383A (en) | Landslide identification method and device based on generation of countermeasure network data expansion strategy | |
| CN115223000A (en) | Method for manufacturing deep learning sample for farmland resource monitoring | |
| Torres et al. | Mountain summit detection with Deep Learning: evaluation and comparison with heuristic methods | |
| Ozdarici-Ok et al. | Automated extraction and validation of Stone Pine (Pinus pinea L.) trees from UAV-based digital surface models | |
| CN113516059A (en) | Solid waste identification method and device, electronic device and storage medium | |
| Mohammadi et al. | Land cover mapping using a novel combination model of satellite imageries: case study of a part of the Cameron Highlands, Pahang, Malaysia. |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171222 |