CN106778547B - Method and device for constructing remote sensing image ground feature classification reference library - Google Patents
Method and device for constructing remote sensing image ground feature classification reference library Download PDFInfo
- Publication number
- CN106778547B CN106778547B CN201611078318.7A CN201611078318A CN106778547B CN 106778547 B CN106778547 B CN 106778547B CN 201611078318 A CN201611078318 A CN 201611078318A CN 106778547 B CN106778547 B CN 106778547B
- Authority
- CN
- China
- Prior art keywords
- feature
- ground
- ground object
- image block
- determining
- 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.)
- Active
Links
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
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Library & Information Science (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
- Processing Or Creating Images (AREA)
Abstract
The invention provides a method and a device for constructing a remote sensing image ground feature classification reference library, and belongs to the technical field of remote sensing. The method comprises the following steps: acquiring a preset number of interest points; screening all interest points according to the land classification standard and the surface feature types of all the interest points; based on the geographical coordinate information of each screened interest point, taking each screened interest point as a center, and intercepting an image block with a preset size from the remote sensing image; and determining the image blocks and the ground object types which are put in storage based on the image blocks and the ground object types corresponding to the screened interest points. The invention determines the image block and the ground object type to be put in storage based on the image block and the ground object type corresponding to each point of interest after screening. Because the image blocks can be intercepted according to a large number of interest points of different ground feature types, more images and ground feature types can be obtained. In addition, the ground object layering mechanism of the land classification standard is used for reference, so that the requirements of diversity, layering and comprehensiveness of ground object types of the reference library can be met.
Description
Technical Field
The invention relates to the technical field of remote sensing, in particular to a method and a device for constructing a remote sensing image ground feature classification reference library.
Background
With the development of remote sensing technology, more and more occasions need to analyze the ground feature image information by constructing a remote sensing image ground feature classification reference library. The remote sensing image feature classification reference library is difficult to construct due to the influences of various factors such as the complexity of the feature of the remote sensing image, the limitation of the image source, the lack of the feature position and the like. Therefore, how to effectively construct the remote sensing image ground feature classification reference library is a key problem. The existing method for constructing the remote sensing image ground object classification reference library mainly selects a plurality of image blocks from a remote sensing image in a manual selection mode and determines the ground object types corresponding to the selected image blocks, thereby realizing the construction of the remote sensing image ground object classification reference library.
In the process of implementing the invention, the prior art is found to have at least the following problems: when the remote sensing image ground feature classification database is constructed in a manual selection mode, the problems that the selection of ground features is more blindness, the manual search efficiency is lower and the like exist, so that the number of images and the types of the ground features in the remote sensing image ground feature classification reference database are relatively small, and the selected types of the ground features lack representativeness and reference value. In addition, the construction efficiency of the reference library is low.
Disclosure of Invention
The present invention provides a method that overcomes, or at least partially solves, the above problems.
According to one aspect of the invention, a method for constructing a remote sensing image ground feature classification reference library is provided, and the method comprises the following steps:
acquiring a preset number of interest points, wherein each interest point comprises a ground object name and geographical coordinate information;
screening all interest points according to the land classification standard and the surface feature types of all the interest points;
based on the geographical coordinate information of each screened interest point, taking each screened interest point as a center, and intercepting an image block with a preset size from the remote sensing image;
and determining the image blocks and the ground object types which are put in storage based on the image blocks and the ground object types corresponding to the screened interest points.
According to another aspect of the present invention, there is provided an apparatus for constructing a reference library of ground feature classification of remote sensing images, the apparatus comprising:
the system comprises an acquisition module, a display module and a display module, wherein the acquisition module is used for acquiring a preset number of interest points, and each interest point at least comprises a ground feature type and geographic coordinate information;
the screening module is used for screening all interest points according to the land classification standard and the ground feature types of all the interest points;
the intercepting module is used for intercepting image blocks with preset sizes from the remote sensing image by taking each screened interest point as a center based on the geographical coordinate information of each screened interest point;
and the determining module is used for determining the image blocks and the ground object types which are put in storage based on the image blocks and the ground object types corresponding to the screened interest points.
The beneficial effect that technical scheme that this application provided brought is:
and screening all interest points according to the land classification standard and the ground feature types of all the interest points by acquiring a preset number of interest points. Based on the geographical coordinate information of each screened interest point, taking each screened interest point as a center, and intercepting an image block with a preset size from the remote sensing image. And determining the image blocks and the ground object types which are put in storage based on the image blocks and the ground object types corresponding to the screened interest points. Because the image blocks can be intercepted according to a large number of interest points with different ground object types, more images and ground object types can be obtained, and the overall construction efficiency is higher. In addition, as the ground feature layering mechanism of the land classification standard is used for reference, each ground feature type is strictly graded, and a tree-shaped library building mode is formed, so that the requirements of diversity, hierarchy and comprehensiveness of the ground feature types in the subsequent building of a reference library can be further met. Meanwhile, the ground feature classification in the subsequent construction of the reference library has great practical significance and research value.
Drawings
Fig. 1 is a schematic flow chart of a method for constructing a remote sensing image ground feature classification reference library according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart of a method for constructing a remote sensing image ground feature classification reference library according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of an image block set of a feature type according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of an image block set of a feature type according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of an image block set of a feature type according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of an image block set of a feature type according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of a device for constructing a remote sensing image ground object classification reference library according to an embodiment of the present invention;
fig. 8 is a schematic structural diagram of a determining module according to an embodiment of the present invention.
Detailed Description
The following detailed description of embodiments of the present invention is provided in connection with the accompanying drawings and examples. The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention.
With the development of remote sensing technology, more and more occasions need to analyze the ground feature image information by constructing a remote sensing image ground feature classification reference library. The remote sensing image feature classification reference library is difficult to construct due to the influences of various factors such as the complexity of the feature of the remote sensing image, the limitation of the image source, the lack of the feature position and the like. At the present research stage, a remote sensing image reference library is relatively lacked, and a remote sensing image ground feature classification reference library for remote sensing image classification is more lacked. Therefore, how to construct the remote sensing image ground feature classification reference library is a key problem.
The existing method for constructing the remote sensing image ground feature classification reference library mainly selects a plurality of image blocks from a remote sensing image in a manual selection mode and determines ground feature categories corresponding to the selected image blocks, thereby realizing the construction of the remote sensing image ground feature classification reference library. When the remote sensing image ground feature classification database is constructed in a manual selection mode, the problems that the selection of ground features is more blindness, the manual search efficiency is lower and the like exist, so that the number of images and the types of the ground features in the remote sensing image ground feature classification reference database are relatively small, and the selected types of the ground features lack representativeness and reference value. In addition, the construction efficiency of the reference library is low.
For example, the current reference library UC-Merced commonly used for remote sensing image classification is constructed by collecting classified remote sensing images with spatial resolution of 0.3 m, including only 21 ground object types and total 2100 ground utilization sizes of 256 × 256 from the urban area of the national map of the USGS (United States Geological Survey).
Aiming at the problems in the prior art, the embodiment provides a method for constructing a remote sensing image ground feature classification reference library. It should be noted that the remote sensing images used in the construction of the remote sensing image ground feature classification reference library in the present embodiment and the subsequent embodiments are mainly derived from crowd source osm (open Street map) geographic data. In which a concept of a crowd-sourced, rational data generation and development to a crowd source is combined with conventional data. The crowd-sourced geographic data is open geospatial data which is acquired volunteers and provided to the public or related institutions through the Internet. It has the advantages of large data volume, rich information, low cost, strong current situation and the like. Compared with the traditional geographic data, the method has great advantages, thereby having great development potential and application value. Referring to fig. 1, the method for constructing the remote sensing image ground feature classification reference library comprises the following steps: 101. acquiring a preset number of interest points; 102. screening all interest points according to the land classification standard and the surface feature types of all the interest points; 103. based on the geographical coordinate information of each screened interest point, taking each screened interest point as a center, and intercepting an image block with a preset size from the remote sensing image; 104. and determining the image blocks and the ground object types which are put in storage based on the image blocks and the ground object types corresponding to the screened interest points. In step 101, each interest point at least includes a feature type and geographic coordinate information, which is not specifically limited in this embodiment.
According to the method provided by the embodiment of the invention, all interest points are screened according to the land classification standard and the ground feature types of all the interest points by acquiring the preset number of interest points. Based on the geographical coordinate information of each screened interest point, taking each screened interest point as a center, and intercepting an image block with a preset size from the remote sensing image. And determining the image blocks and the ground object types which are put in storage based on the image blocks and the ground object types corresponding to the screened interest points. Because the image blocks can be intercepted according to a large number of interest points with different ground object types, more images and ground object types can be obtained, and the overall construction efficiency is higher. In addition, as the ground feature layering mechanism of the land classification standard is used for reference, each ground feature type is strictly graded, and a tree-shaped library building mode is formed, so that the requirements of diversity, hierarchy and comprehensiveness of the ground feature types in the subsequent building of a reference library can be further met. Meanwhile, the ground feature classification in the subsequent construction of the reference library has great practical significance and research value.
As an alternative embodiment, all the interest points are screened according to the land classification criteria and the surface feature types of all the interest points, which includes:
determining the ground feature types overlapped between the ground feature types of the land classification standard and the ground feature types of all the interest points;
and taking the interest points corresponding to the overlapped ground feature types as the screened interest points.
As an optional embodiment, determining the image block and the surface feature type to be put in storage based on the image block and the surface feature type corresponding to each point of interest after screening, includes:
for any point of interest after screening and any image block corresponding to any point of interest, matching the ground feature type of any point of interest with the actual ground feature in any image block;
when the two are matched, determining the feature type corresponding to any image block according to the feature type of each interest point in any image block;
and classifying and warehousing any image block according to the ground object type corresponding to any image block.
As an optional embodiment, determining the feature type corresponding to any image block according to the feature type of each interest point in any image block includes:
determining a ground object main body corresponding to the ground object type of each interest point in any image block;
determining the ground object main body with the highest recognition degree in all the ground object main bodies;
and taking the ground object type corresponding to the ground object main body with the highest recognition degree as the ground object type corresponding to any image block.
As an alternative embodiment, determining the feature body with the highest recognition degree in all the feature bodies includes:
determining the maximum ground object main body area according to the ground object main body area of each ground object main body, and taking the ground object main body corresponding to the maximum ground object main body area as the ground object main body with the highest recognition degree;
or, according to the span range of each ground feature body in any image block, determining the ground feature body with the largest span range as the ground feature body with the highest recognition degree.
All the above-mentioned optional technical solutions can be combined arbitrarily to form the optional embodiments of the present invention, and are not described herein again.
Based on the content provided by the embodiment corresponding to fig. 1, the embodiment of the invention provides a method for constructing a remote sensing image ground feature classification reference library. Referring to fig. 2, the method flow provided by this embodiment includes: 201. acquiring a preset number of interest points; 202. determining the ground feature types overlapped between the ground feature types of the land classification standard and the ground feature types of all the interest points; 203. taking the interest points corresponding to the overlapped ground feature types as the screened interest points; 204. based on the geographical coordinate information of each screened interest point, taking each screened interest point as a center, and intercepting an image block with a preset size from the remote sensing image; 205. and determining the image blocks and the ground object types which are put in storage based on the image blocks and the ground object types corresponding to the screened interest points.
Wherein 201, a preset number of interest points are obtained.
When the remote sensing image ground feature classification reference library is constructed, image blocks need to be intercepted from the remote sensing image, so that a preset number of interest points can be obtained before the reference library is constructed. And intercepting the remote sensing image according to the acquired interest point. Considering the quantity distribution of the interest points in the world, the activity degree of the labels and the resolution of the remote sensing images, when the interest points are obtained, the obtained interest points are mainly located in provinces such as Beijing, Shanghai, hong Kong, Guangzhou, Hainan and Fujian in China; cities in new york, washington, los angeles, chicago, etc., usa; cities such as tokyo, osaka, and shenhu, japan; cities such as paris, nice, etc. france; cities such as Ottawa, Toronto, Canada; cities like Russian Moscow, St. Peterburg, and other countries and regions thereof. Of course, the interest points may also be obtained according to actual requirements when obtaining the interest points, which is not specifically limited in this embodiment.
The interest point may at least include a surface feature type and a geographic coordinate. The land feature type is a type divided into land according to factors such as land use, vegetation form on the land and the like. Each large feature type may include many sub-feature types, which is not limited in this embodiment. For example, the land object type "traffic" can be divided according to land use, and the traffic "can be subdivided into" crossroads "," highways "," express lanes "," ports "," overpasses ", and" river bridges ", etc. The land feature type of forest land can be divided according to the vegetation form on the land, and the forest land can be subdivided into forest, seedling, river-protecting forest, forest land and brush forest. In addition, the geographic coordinate information in the interest point is a longitude and latitude value of the interest point, which is not specifically limited in this embodiment. The preset number may be set according to actual requirements, and this embodiment is not particularly limited.
Wherein 202, the ground feature types overlapped between the ground feature types of the land classification standard and the ground feature types of all the interest points are determined.
Through the step 201, after a large number of interest points are obtained, the obtained interest points may be screened in order to obtain interest points with a higher reference value, which is not specifically limited in this embodiment. This step 202 and the following step 203 are mainly detailed processes of screening all the points of interest according to the land classification criteria and the surface feature types of all the points of interest. Of course, other screening methods may also be adopted in the process of screening the interest points, which is not specifically limited in this embodiment.
Considering the reference value of land feature types in China, the land classification standard can be a classification mechanism of land in China, namely, the land in China is divided into a plurality of land feature types according to factors such as land use and the like. For example, one can classify as "residential", "commercial shopping", "industrial manufacturing", "social group infrastructure related", "travel trip", "free", "natural resources related", "unattended or unclassified", and "large-scale crowd gathering", etc. Accordingly, the surface feature types of the points of interest may include: "convenience facility", "land for use in obstructed area", "construction land", "emergency land for use", "road", "historical land for use", "free land for use", "artificial area", "natural area", "office", "public transportation", "railway", "shop", "sport place", and "tourist land for use".
The interest points are different from the standard when the land object types are divided according to the land classification standard, so that the divided land object types are different literally, but actually, a plurality of land object types or sub land object types under the land object types are overlapped. For example, the terrain type "free" in the land classification criteria overlaps with the "free" in the point of interest terrain types, both refer to free land, and both are ground terrain types. The ground feature type "convenience facility" in the land classification criteria overlaps with the "commercial shopping" in the point-of-interest ground feature type, because the sub-ground feature types under the "convenience facility" may include "mall", "supermarket" and "shop" and the like, and the sub-ground feature types under the "commercial shopping" may also include "mall", "supermarket" and "shop" and the like, i.e., the sub-ground feature types of both overlap.
Based on the above, the feature types overlapping between the feature types of the land classification criteria and the feature types of all the interest points can be referred to the following table (1):
watch (1)
In the table entry corresponding to each row and each column in the above table (1), if the contents of the table entry are empty, it indicates that the types of the ground features are not overlapped. If the representation content is a hook, the representation contents indicate that the ground feature types are overlapped. It should be noted that, due to space limitations of the table contents, there may be many overlapped ground object types, and this embodiment is not limited in this respect.
And 203, taking the interest points corresponding to the overlapped surface feature types as the screened interest points.
Through the above step 202, overlapped ground feature types can be obtained. For example, the ground type "plot free" superimposed in table (1) above. In addition, there are also literally non-overlapping geodetic types that contain overlapping sub-geodetic types, such as the geodetic type "natural resources related" of the land classification criteria and the geodetic type "tourist ground" of the point of interest.
And conversely screening the interest points according to the overlapped ground feature types and the ground feature type of each interest point to obtain the screened interest points. Through the step 202 and the step 203, as the ground feature layering mechanism of the land classification standard is used for reference, each ground feature type is strictly graded, and a tree-shaped library building mode is formed, so that the requirements of diversity, layering and comprehensiveness of the ground feature types in the subsequent building of the reference library can be further met. In addition, the ground feature types when the reference library is subsequently established depend on the overlapped ground feature types, so that the ground feature classification when the reference library is subsequently established has greater practical significance and research value.
And 204, based on the geographical coordinate information of each screened interest point, taking each screened interest point as a center, and intercepting an image block with a preset size from the remote sensing image.
Before executing the step, the interest point and the remote sensing image can be superposed based on the geographic coordinate of the interest point and the geographic coordinate information of the remote sensing image, so that the interest point falls into a certain distance position in the remote sensing image. Based on the above contents, the image block with the preset size can be intercepted from the remote sensing image by taking the geographic coordinate of each interest point as the center. The area corresponding to the image block with the predetermined size is not particularly limited in this embodiment. In addition, the image block may be a square, and the shape of the image block is not specifically limited in this embodiment.
And 205, determining the image block and the ground object type which are put in storage based on the image block and the ground object type corresponding to each screened interest point.
Since a large number of image blocks corresponding to the interest points are obtained in the step 204, for convenience of description, the process of this step will be described by taking any selected interest point and any image block corresponding to any selected interest point as examples. The present embodiment does not specifically limit the manner of determining the image block and the surface feature type to be put in storage based on the image block and the surface feature type corresponding to each point of interest after the screening, which includes but is not limited to: matching the surface feature type of any interest point with the surface feature type of any image block; when the two are matched, determining the feature type corresponding to any image block according to the feature type of each interest point in any image block; and classifying and warehousing any image block according to the ground object type corresponding to any image block.
Since there may be an error in the feature type of the interest point, the feature type of any interest point may be matched with the feature type of any image block, which is not specifically limited in this embodiment. For example, if the feature type of the point of interest is "parking lot", and the feature type of the image block obtained according to the image content in the image block is "lawn", it may be determined that the feature type of the point of interest does not match the feature type of the image block. If the feature type of the interest point is 'road', the feature type of the image block obtained according to the image content in the image block is also 'road', so that the feature type of the interest point can be determined to be matched with the feature type of the image block.
When the two images do not match, the corresponding image blocks and the corresponding ground object types are not accurate. Therefore, the image block can be directly deleted, which is not limited in this embodiment. And the image blocks which do not meet the conditions are deleted through the matching result, so that the accuracy and the reliability of the image data can be ensured.
When the two match, the points of interest may be very close when they are acquired in step 201 because different surface feature types may appear intensively. Therefore, the image block intercepted in step 204 may contain other interest points besides the center interest point. For the image block, the feature type of the central interest point is not necessarily the feature type with the highest matching degree with the image block. Therefore, the feature type corresponding to the image block needs to be determined according to the feature type of each interest point in the image block.
The present embodiment does not specifically limit the manner of determining the feature type corresponding to any image block according to the feature type of each interest point in any image block, and includes but is not limited to: determining a ground object main body corresponding to the ground object type of each interest point in any image block; determining the ground object main body with the highest recognition degree in all the ground object main bodies; and taking the ground object type corresponding to the ground object main body with the highest recognition degree as the ground object type corresponding to any image block.
When determining the feature body with the highest recognition degree in all the feature bodies, the following two ways may be included, but are not limited to:
the first method is as follows: and determining the maximum ground object main body area according to the ground object main body area of each ground object main body, and taking the ground object main body corresponding to the maximum ground object main body area as the ground object main body with the highest recognition degree.
The method mainly determines the feature main body with the highest recognition degree according to the occupied image area, namely determines which feature main body in the image block occupies the largest image area, and takes the feature main body occupying the largest image area as the feature main body with the highest recognition degree. For example, if there are two feature bodies in one image block, one is a "road" and the other is a "bus station". If the area of the map width occupied by the road is larger than that occupied by the bus station in the map, the main body of the ground object with the highest recognition degree is the road.
The second method comprises the following steps: and determining the ground object main body with the largest span range according to the span range of each ground object main body in any image block, and taking the ground object main body with the highest identification degree as the ground object main body.
In the first mode, when the feature body with the highest recognition degree is determined, there is a possibility that the difference between the areas of the map frames occupied by the plurality of feature bodies is not large. For the situation, the ground object body with the highest recognition degree can be determined in a sampling mode II. For example, if there are two feature bodies in one image block, "bridge" and "opera house", respectively, and the "bridge" and "opera house" occupy the same area of the image frame in the image block. Wherein, the opera house is positioned at one corner of the image block and only exposes one part of the building body, and the span range is very small. The bridge is positioned from the upper end to the lower end of the image block, the bridge body is positioned in the image block, and the span range is large. Therefore, the feature body with the highest recognition degree in the image block can be determined as the bridge.
After the feature body with the highest recognition degree is determined from all the feature bodies, the feature type corresponding to the feature body with the highest recognition degree can be used as the feature type corresponding to the image block. For example, the "bridge" in the above example can be the type of feature corresponding to the image block. The ground object type of the image block is determined through the two modes, and the reliability and the accuracy of the image data can be further ensured.
After the surface feature type of each image block is determined, the image blocks can be classified and stored according to the surface feature types of the image blocks. For example, the classification-binning partial results may be as shown in fig. 3-6. Fig. 3 is a set of image blocks corresponding to a feature type of "dam", fig. 4 is a set of image blocks corresponding to a feature type of "runway", fig. 5 is a set of image blocks corresponding to a feature type of "city building", and fig. 6 is a set of image blocks corresponding to a feature type of "city road".
According to the method provided by the embodiment of the invention, all interest points are screened according to the land classification standard and the ground feature types of all the interest points by acquiring the preset number of interest points. Based on the geographical coordinate information of each screened interest point, taking each screened interest point as a center, and intercepting an image block with a preset size from the remote sensing image. And determining the image blocks and the ground object types which are put in storage based on the image blocks and the ground object types corresponding to the screened interest points. Because the image blocks can be intercepted according to a large number of interest points with different ground object types, more images and ground object types can be obtained, and the overall construction efficiency is higher. In addition, as the ground feature layering mechanism of the land classification standard is used for reference, each ground feature type is strictly graded, and a tree-shaped library building mode is formed, so that the requirements of diversity, hierarchy and comprehensiveness of the ground feature types in the subsequent building of a reference library can be further met. Meanwhile, the ground feature classification in the subsequent construction of the reference library has great practical significance and research value.
Finally, the ground object type of the interest point can be matched with the actual ground object in the image block, and the unmatched image block is deleted, so that the accuracy and reliability of the data in the remote sensing image ground object classification reference library can be guaranteed. In addition, the block feature type of the image can be determined according to the feature main body with the highest recognition degree in the image block, so that the reliability and the accuracy of the data in the remote sensing image feature classification reference library can be further improved.
The embodiment of the invention provides a device for constructing a remote sensing image ground feature classification reference library, which is used for executing the method for constructing the remote sensing image ground feature classification reference library provided in the embodiment corresponding to the figure 1 or the figure 2. Referring to fig. 7, the apparatus includes:
an obtaining module 701, configured to obtain a preset number of interest points, where each interest point at least includes a surface feature type and a geographic coordinate;
the screening module 702 is configured to screen all the interest points according to the land classification standard and the surface feature types of all the interest points;
an intercepting module 703, configured to intercept, based on the geographical coordinate information of each filtered interest point, an image block with a preset size from the remote sensing image with each filtered interest point as a center;
and the determining module 704 is configured to determine the image block and the ground object type to be put in storage based on the image block and the ground object type corresponding to each screened interest point.
As an alternative embodiment, the screening module 702 is configured to determine the ground feature types overlapping between the ground feature types of the land classification criteria and the ground feature types of all the interest points; and taking the interest points corresponding to the overlapped ground feature types as the screened interest points.
As an alternative embodiment, referring to fig. 8, the determining module 704 includes:
a matching unit 7041, configured to match, for any filtered interest point and any image block corresponding to any interest point, the surface feature type of any interest point with the actual surface feature in any image block;
a determining unit 7042, configured to determine, when the two are matched, a feature type corresponding to any image block according to the feature type of each interest point in any image block;
the warehousing unit 7043 is configured to classify and warehouse any image block according to the ground feature type corresponding to the image block.
As an alternative embodiment, determining unit 7042 includes:
the first determining subunit is used for determining a feature main body corresponding to the feature type of each interest point in any image block;
the second determining subunit is used for determining the feature main body with the highest recognition degree in all the feature main bodies;
and the third determining subunit is used for taking the ground object type corresponding to the ground object main body with the highest recognition degree as the ground object type corresponding to any image block.
As an optional embodiment, the second determining subunit is configured to determine a maximum surface area of the feature body according to the surface area of each feature body, and use the feature body corresponding to the maximum surface area of the feature body as the feature body with the highest recognition degree; or, according to the span range of each ground feature body in any image block, determining the ground feature body with the largest span range as the ground feature body with the highest recognition degree.
According to the device provided by the embodiment of the invention, all interest points are screened according to the land classification standard and the ground feature types of all the interest points by acquiring the preset number of interest points. Based on the geographical coordinate information of each screened interest point, taking each screened interest point as a center, and intercepting an image block with a preset size from the remote sensing image. And determining the image blocks and the ground object types which are put in storage based on the image blocks and the ground object types corresponding to the screened interest points. Because the image blocks can be intercepted according to a large number of interest points with different ground object types, more images and ground object types can be obtained, and the overall construction efficiency is higher. In addition, as the ground feature layering mechanism of the land classification standard is used for reference, each ground feature type is strictly graded, and a tree-shaped library building mode is formed, so that the requirements of diversity, hierarchy and comprehensiveness of the ground feature types in the subsequent building of a reference library can be further met. Meanwhile, the ground feature classification in the subsequent construction of the reference library has great practical significance and research value.
Finally, the method of the present application is only a preferred embodiment and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (8)
1. A method for constructing a remote sensing image ground feature classification reference library is characterized by comprising the following steps:
acquiring a preset number of interest points, wherein each interest point comprises a ground feature type and geographic coordinate information;
screening all interest points according to the land classification standard and the surface feature types of all the interest points;
based on the geographical coordinate information of each screened interest point, taking each screened interest point as a center, and intercepting an image block with a preset size from the remote sensing image;
determining the image blocks and the ground object types which are put in storage based on the image blocks and the ground object types which correspond to the screened interest points;
and screening all interest points according to the land classification standard and the surface feature types of all the interest points, wherein the screening comprises the following steps:
determining the ground feature types overlapped between the ground feature types of the land classification standard and the ground feature types of all the interest points;
taking the interest points corresponding to the overlapped ground feature types as the screened interest points;
the method for determining the image block and the ground object type of the warehouse based on the image block and the ground object type corresponding to each screened interest point comprises the following steps:
for any selected interest point and any image block corresponding to the selected interest point, matching the surface feature type of the interest point with the actual surface feature in the image block;
when the two image blocks are matched, determining the feature type corresponding to any image block according to the feature type of each interest point in any image block;
and classifying and warehousing any image block according to the ground object type corresponding to the image block.
2. The method according to claim 1, wherein the determining the feature type corresponding to any image block according to the feature type of each interest point in any image block comprises:
determining a ground object main body corresponding to the ground object type of each interest point in any image block;
determining the ground object main body with the highest recognition degree in all the ground object main bodies;
and taking the ground object type corresponding to the ground object main body with the highest recognition degree as the ground object type corresponding to any image block.
3. The method according to claim 2, wherein the determining the feature body with the highest recognition degree in all the feature bodies comprises:
determining the maximum ground object main body area according to the ground object main body area of each ground object main body, and taking the ground object main body corresponding to the maximum ground object main body area as the ground object main body with the highest recognition degree;
or, according to the span range of each ground feature body in any image block, determining the ground feature body with the largest span range as the ground feature body with the highest recognition degree.
4. The device for constructing the remote sensing image ground feature classification reference library is characterized by comprising the following components:
the system comprises an acquisition module, a display module and a display module, wherein the acquisition module is used for acquiring a preset number of interest points, and each interest point at least comprises a ground feature type and geographic coordinate information;
the screening module is used for screening all interest points according to the land classification standard and the ground feature types of all the interest points;
the intercepting module is used for intercepting image blocks with preset sizes from the remote sensing image by taking each screened interest point as a center based on the geographical coordinate information of each screened interest point;
and the determining module is used for determining the image blocks and the ground object types which are put in storage based on the image blocks and the ground object types corresponding to the screened interest points.
5. The apparatus of claim 4, wherein the screening module is configured to determine land feature types that overlap between the land feature types of the land classification criteria and the land feature types of all the points of interest; and taking the interest points corresponding to the overlapped ground feature types as the screened interest points.
6. The apparatus of claim 4, wherein the determining module comprises:
the matching unit is used for matching the ground object type of any interest point with the actual ground object in any image block for any screened interest point and any image block corresponding to the interest point;
the determining unit is used for determining the ground object type corresponding to any image block according to the ground object type of each interest point in any image block when the two image blocks are matched;
and the warehousing unit is used for classifying and warehousing any image block according to the ground object type corresponding to the image block.
7. The apparatus of claim 6, wherein the determining unit comprises:
the first determining subunit is used for determining a feature main body corresponding to the feature type of each interest point in any image block;
the second determining subunit is used for determining the feature main body with the highest recognition degree in all the feature main bodies;
and the third determining subunit is used for taking the feature type corresponding to the feature main body with the highest recognition degree as the feature type corresponding to any image block.
8. The apparatus of claim 7, wherein the second determining subunit,
the device comprises a ground object main body, a ground object main body and a control unit, wherein the ground object main body is used for determining the maximum ground object main body area according to the ground object main body area of each ground object main body, and the ground object main body corresponding to the maximum ground object main body area is used as the ground object main body with the highest recognition degree; or, according to the span range of each ground feature body in any image block, determining the ground feature body with the largest span range as the ground feature body with the highest recognition degree.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611078318.7A CN106778547B (en) | 2016-11-29 | 2016-11-29 | Method and device for constructing remote sensing image ground feature classification reference library |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611078318.7A CN106778547B (en) | 2016-11-29 | 2016-11-29 | Method and device for constructing remote sensing image ground feature classification reference library |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106778547A CN106778547A (en) | 2017-05-31 |
| CN106778547B true CN106778547B (en) | 2020-06-30 |
Family
ID=58898091
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611078318.7A Active CN106778547B (en) | 2016-11-29 | 2016-11-29 | Method and device for constructing remote sensing image ground feature classification reference library |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106778547B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109508585A (en) * | 2017-09-15 | 2019-03-22 | 中国科学院城市环境研究所 | A method of urban function region is extracted based on POI and high-resolution remote sensing image |
| CN108053409B (en) * | 2017-12-11 | 2022-05-13 | 中南大学 | Automatic construction method and system for remote sensing image segmentation reference library |
| CN109063754B (en) * | 2018-07-18 | 2020-08-07 | 武汉大学 | Remote sensing image multi-feature joint classification method based on OpenStreetMap |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521598A (en) * | 2011-12-16 | 2012-06-27 | 北京市城市规划设计研究院 | Identification method based on remote-sensing image |
| CN104112136A (en) * | 2013-04-19 | 2014-10-22 | 昆山鸿鹄信息技术服务有限公司 | Image low-level visual feature extraction method |
| CN104915636A (en) * | 2015-04-15 | 2015-09-16 | 北京工业大学 | Remote sensing image road identification method based on multistage frame significant characteristics |
| CN106021499A (en) * | 2016-05-20 | 2016-10-12 | 中国农业大学 | Construction land classification method and device based on geographic information of volunteer |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9020274B2 (en) * | 2009-05-06 | 2015-04-28 | University Of New Brunswick | Method of interest point matching for images |
-
2016
- 2016-11-29 CN CN201611078318.7A patent/CN106778547B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521598A (en) * | 2011-12-16 | 2012-06-27 | 北京市城市规划设计研究院 | Identification method based on remote-sensing image |
| CN104112136A (en) * | 2013-04-19 | 2014-10-22 | 昆山鸿鹄信息技术服务有限公司 | Image low-level visual feature extraction method |
| CN104915636A (en) * | 2015-04-15 | 2015-09-16 | 北京工业大学 | Remote sensing image road identification method based on multistage frame significant characteristics |
| CN106021499A (en) * | 2016-05-20 | 2016-10-12 | 中国农业大学 | Construction land classification method and device based on geographic information of volunteer |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106778547A (en) | 2017-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zheng et al. | Spatial–temporal travel pattern mining using massive taxi trajectory data | |
| CN107229690B (en) | Dynamic High-accuracy map datum processing system and method based on trackside sensor | |
| Zhang et al. | Density and diversity of OpenStreetMap road networks in China | |
| CN106296350A (en) | A kind of visual analyzing city public bicycle system borrows the method for also pattern | |
| CN106897417A (en) | A kind of construction method of the city space holographic map based on the fusion of multi-source big data | |
| US11710074B2 (en) | System for providing corridor metrics for a corridor of a road network | |
| CN105206057A (en) | Detection method and system based on floating car resident trip hot spot regions | |
| CN106778547B (en) | Method and device for constructing remote sensing image ground feature classification reference library | |
| CN107609242A (en) | It is a kind of that selection method is reconnoitred based on the highway of smart mobile phone and digital topography map | |
| CN105844031B (en) | A kind of urban transportation gallery recognition methods based on mobile phone location data | |
| CN108388995A (en) | A kind of method for building up of road asset management system and establish system | |
| CN101425238B (en) | Traffic geographical information system generating method | |
| CN106021499B (en) | Construction land classification method and device based on volunteer geographic information | |
| CN105468595A (en) | Bus route planning method and system | |
| CN103246650B (en) | A kind of road logic model preparation method | |
| CN107943962A (en) | A kind of Gis2.5D maps and preparation method thereof | |
| CN111291146A (en) | Method, device and storage medium for fusing multi-rule-in-one data | |
| Bianchi et al. | An innovational digital tool in GIS procedure: Mapping adriatic coast in abruzzo region to support design of slow mobility routes | |
| Niu | Estimating housing vacancy rate in Qingdao city with NPP-VIIRS nighttime light and geographical national conditions monitoring data | |
| Bartelme | Geographic information systems | |
| Udoh et al. | Production of revised street map of Uyo Urban Area, Nigeria using remote sensing and GIS approach | |
| Ojiako et al. | Application of GIS and remote sensing approach for the analysis of Asaba Urban Street Network of Delta State, Nigeria | |
| Zachar et al. | Cenagis-Als Benchmark-New Proposal for Dense ALS Benchmark Based on the Review of Datasets and Benchmarks for 3d Point Cloud Segmentation | |
| CN112883138B (en) | Urban part searching method based on measurable real scenes | |
| Zhang et al. | Extraction and Digital Translation of Urban Spatial Elements Based on a Series of Historical Maps |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |