An object-oriented classification method and device based on scale set

Abstract

The present invention provides an object-oriented classification method and device based on a scale set, which relates to the technical field of remote sensing image classification. The method and device include: obtaining block image data of a remote sensing image to be classified, and segmenting the block image data to obtain segmented image data, wherein there is an overlapping area between any two adjacent block image data; seamlessly splicing the segmented image data to obtain spliced ​​image data; merging hierarchical regions of the spliced ​​image data, and constructing a low-level scale set, and determining level 0 scale segmentation data in the low-level scale set; constructing a high-level scale set based on the level 0 scale segmentation data; obtaining a scale parameter input by a user, and classifying the remote sensing image to be classified based on the scale parameter and the high-level scale set to obtain a classification result. The method solves the technical problem of low classification accuracy and classification efficiency of existing object-oriented classification methods.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing image classification, and in particular to an object-oriented classification method and device based on a scale set. Background Art

[0002] Currently, traditional remote sensing image classification primarily utilizes machine learning classification algorithms based on pixels and objects. Compared to using only spectral information, many researchers have improved the accuracy of pixel-based classification by adding data such as NDVI, NDWI, and DEM. However, due to the "same object, different spectrum" and "same spectrum, different objects" phenomena, adjacent pixels often have different spectra, resulting in classification results prone to problems such as salt and pepper noise and poor visual quality, often leading to lower classification accuracy. Object-oriented classification processing is based on the image segmentation object, which fully utilizes the spectral information, spatial structure, texture information, location information, and topological relationships of remote sensing images, overcoming these issues, such as "salt and pepper noise," "same object, different spectrum," and "same spectrum, different objects," and achieving higher classification accuracy.

[0003] However, existing object-oriented classification methods have problems such as optimal scale stability and over-segmentation, under-segmentation and spatial autocorrelation uncertainty of local objects.

[0004] For the above problems, no effective solutions have been proposed yet. Summary of the Invention

[0005] In view of this, an object of the present invention is to provide an object-oriented classification method and apparatus based on a scale set, so as to alleviate the technical problems of low classification accuracy and classification efficiency of existing object-oriented classification methods.

[0006] In a first aspect, an embodiment of the present invention provides an object-oriented classification method based on a scale set, comprising: obtaining block image data of a remote sensing image to be classified, and segmenting the block image data to obtain segmented image data, wherein there is an overlapping area between any two adjacent block image data; seamlessly splicing the segmented image data to obtain spliced ​​image data; hierarchical region merging of the spliced ​​image data, and constructing a low-level scale set, and determining level 0 scale segmentation data in the low-level scale set; constructing a high-level scale set based on the level 0 scale segmentation data; obtaining a scale parameter input by a user, and classifying the remote sensing image to be classified based on the scale parameter and the high-level scale set to obtain a classification result.

[0007] Furthermore, in a first determining step, target image data is determined based on the segmented image data and adjacent segmented image data of the segmented image data, wherein the target image data is the segmented image data with the smallest merge cost value between the adjacent segmented image data and the segmented image data; in a second determining step, a node is constructed based on each segmented image data and the corresponding target image data, and an initial node is determined, and the initial node is inserted into an initial red-black tree, wherein the initial node is an initial node containing the same segmented image data; in a constructing step, the node with the smallest merge cost value among the initial nodes is determined as the target node, and the initial nodes other than the target node are removed from the initial red-black tree to obtain an intermediate red-black tree; in a third determining step, a final node is determined based on the intermediate red-black tree, and the final node is deleted from the intermediate red-black tree point, obtaining a target red-black tree, wherein the final node is a node at the bottom layer of the intermediate red-black tree and has the smallest merge cost value; determining the segmented image data other than the target image data in the adjacent segmented image data as the adjacent segmented image data, and determining the target red-black tree as the initial red-black tree, repeating the first determining step, the second determining step and the constructing step for a first preset number of times, wherein the first preset number of times is the number of the segmented image data; merging the segmented image data corresponding to each final node obtained in the first preset number of times to obtain the merged spliced ​​image data; determining a first cumulative merge cost value of each segmented image data based on the merge cost value corresponding to each final node obtained in the first preset number of times, and constructing the low-level scale set based on the first cumulative merge cost value.

[0008] Furthermore, determining the level 0 scale segmentation data in the low-level scale set includes: sorting the first cumulative merge cost values ​​and determining a target cumulative merge cost value corresponding to a preset ratio; merging the segmented image data that is less than the target cumulative merge cost value to obtain the level 0 scale segmentation data.

[0009] Furthermore, based on the level 0 scale segmentation data, a high-level scale set is constructed, including: classifying the level 0 scale segmentation data according to a preset classification algorithm to obtain an initial classification result of the segmented image data contained in the level 0 scale segmentation data; determining the level 0 scale segmentation data as the segmented image data, and repeating the first determination step, the second determination step and the construction step a first preset number of times, wherein the second preset number of times is the number of segmented image data contained in the level 0 scale segmentation data; determining a second cumulative merged cost value of the segmented image data contained in the level 0 scale segmentation data based on the merged cost value corresponding to each final node obtained in the second preset number of times, and constructing the high-level scale set based on the second cumulative merged cost value.

[0010] Furthermore, if the classification type for classifying the level 0 scale segmentation data is land cover classification, the level 0 scale segmentation data is classified based on a preset classification algorithm to obtain an initial classification result of the segmented image data contained in the level 0 scale segmentation data, including: based on the land cover classification, obtaining feature data of the level 0 scale segmentation data, wherein the feature data includes: spectral features, texture features, shape features and index features; based on the preset classification algorithm and the feature data, the level 0 scale segmentation data is classified into land cover classes to obtain the initial classification result.

[0011] Furthermore, before obtaining the block image data of the remote sensing image to be classified, the method includes: obtaining the remote sensing image to be classified, and segmenting the remote sensing image to be classified according to a preset size to obtain initial block image data; based on the four-dimensional range of the initial block image data, determining the overlapping area between the initial block image data and the adjacent initial block image data of the initial block image data; and constructing the block image data based on the overlapping area and the initial block image data.

[0012] Furthermore, based on the scale parameter and the high-level scale set, the remote sensing image to be classified is classified to obtain a classification result, including: based on the scale parameter and the high-level scale set, the level 0 scale segmentation data is merged to obtain a merged result; based on the merged result and the initial classification result, the classification result is determined.

[0013] In a second aspect, an embodiment of the present invention further provides an object-oriented classification device based on a scale set, comprising: an acquisition unit, a splicing unit, a first construction unit, a second construction unit and a classification unit, wherein the acquisition unit is used to acquire block image data of the remote sensing image to be classified, and segment the block image data to obtain segmented image data, wherein there is an overlapping area between any two adjacent block image data; the splicing unit is used to seamlessly splice the segmented image data to obtain spliced ​​image data; the first construction unit is used to perform hierarchical region merging on the spliced ​​image data, and construct a low-level scale set, and determine the level 0 scale segmentation data in the low-level scale set; the second construction unit is used to construct a high-level scale set based on the level 0 scale segmentation data; the classification unit is used to obtain the scale parameter input by the user, and classify the remote sensing image to be classified based on the scale parameter and the high-level scale set to obtain a classification result.

[0014] In a third aspect, an embodiment of the present invention further provides an electronic device comprising a memory and a processor, wherein the memory is used to store a program that supports the processor to execute the method described in the first aspect above, and the processor is configured to execute the program stored in the memory.

[0015] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored.

[0016] In an embodiment of the present invention, segmented image data of a remote sensing image to be classified is obtained, and the segmented image data is segmented to obtain segmented image data, wherein there is an overlapping area between any two adjacent segmented image data; the segmented image data is seamlessly spliced ​​to obtain spliced ​​image data; hierarchical region merging is performed on the spliced ​​image data, and a low-level scale set is constructed, and level 0 scale segmented data in the low-level scale set is determined; a high-level scale set is constructed based on the level 0 scale segmented data; a scale parameter input by a user is obtained, and the remote sensing image to be classified is classified based on the scale parameter and the high-level scale set to obtain a classification result, thereby achieving the purpose of classifying remote sensing images using a two-level scale set, thereby solving the technical problem of low classification accuracy and classification efficiency of the existing object-oriented classification method, thereby achieving the technical effect of improving the classification accuracy and classification efficiency of the object-oriented classification method.

[0017] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or understood by practicing the present invention. The purposes and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0018] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 A flowchart of an object-oriented classification method based on a scale set provided by an embodiment of the present invention;

[0021] Figure 2 A schematic diagram of stitching image data provided by an embodiment of the present invention;

[0022] Figure 3 A schematic diagram of an object-oriented classification device based on a scale set provided by an embodiment of the present invention;

[0023] Figure 4 A schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0025] Example 1:

[0026] According to an embodiment of the present invention, an embodiment of an object-oriented classification method based on a scale set is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0027] Figure 1 is a flow chart of an object-oriented classification method based on a scale set according to an embodiment of the present invention. Figure 1 As shown, the method includes the following steps:

[0028] Step S102, obtaining block image data of the remote sensing image to be classified, and segmenting the block image data to obtain segmented image data, wherein there is an overlapping area between any two adjacent block image data;

[0029] Step S104, seamlessly stitching the segmented image data to obtain stitched image data;

[0030] Step S106, performing hierarchical region merging on the stitched image data, constructing a low-level scale set, and determining level 0 scale segmentation data in the low-level scale set;

[0031] Step S108, constructing a high-level scale set based on the level 0 scale segmentation data;

[0032] Step S110 , obtaining a scale parameter input by a user, and classifying the remote sensing image to be classified based on the scale parameter and the high-level scale set to obtain a classification result.

[0033] In an embodiment of the present invention, segmented image data of a remote sensing image to be classified is obtained, and the segmented image data is segmented to obtain segmented image data, wherein there is an overlapping area between any two adjacent segmented image data; the segmented image data is seamlessly spliced ​​to obtain spliced ​​image data; hierarchical region merging is performed on the spliced ​​image data, and a low-level scale set is constructed, and level 0 scale segmented data in the low-level scale set is determined; a high-level scale set is constructed based on the level 0 scale segmented data; a scale parameter input by a user is obtained, and the remote sensing image to be classified is classified based on the scale parameter and the high-level scale set to obtain a classification result, thereby achieving the purpose of classifying remote sensing images using a two-level scale set, thereby solving the technical problem of low classification accuracy and classification efficiency of the existing object-oriented classification method, thereby achieving the technical effect of improving the classification accuracy and classification efficiency of the object-oriented classification method.

[0034] In an embodiment of the present invention, before obtaining the block image data of the remote sensing image to be classified, the method includes the following steps:

[0035] Step S11, obtaining the remote sensing image to be classified, and segmenting the remote sensing image to be classified according to a preset size to obtain initial block image data;

[0036] Step S12, determining an overlapping area between the initial block image data and adjacent initial block image data of the initial block image data based on the boundary range of the initial block image data;

[0037] Step S13: constructing the block image data based on the overlapping area and the initial block image data.

[0038] In an embodiment of the present invention, after obtaining the remote sensing image to be classified, it is necessary to perform block processing on the remote sensing image to be classified. Taking into account the computer memory size and the efficiency of single-step operation, the remote sensing image to be classified is divided into blocks according to preset sizes (such as 512x1024, 512x512, 1024x1024, etc.) to obtain initial block image data. In the embodiment of the present invention, the blocks are divided according to the size of 512x1024.

[0039] Since the subsequent steps require seamless splicing of the segmentation results, after obtaining the initial segmented image data, it is necessary to determine the n-pixel width outside the four-dimensional range of each block of initial segmented image data as the overlapping area between the initial segmented image data and the adjacent initial segmented image data, thereby constructing the segmented image data.

[0040] When necessary, the general n value follows the following calculation formula:

[0041]

[0042] S represents the expected value of the average area of ​​the block image data, n represents the width of the overlapping area, and the units of S and n are both pixels. In the embodiment of the present invention, the value of S is 100.

[0043] The following is a detailed description of step S104.

[0044] After obtaining the segmented image data, the four boundaries of the previous segmented image data are used as the basis, and the four adjacent segmented image data are obtained for splicing. First, a row or a column of data in the non-overlapping area at the four boundaries is obtained, and then a row or a column of data corresponding to the overlapping area of ​​the latter segmented image data is obtained. The two sets of data are compared. When the current data and the latter data are inconsistent, the former data value is modified to be consistent with the latter data value until the entire row or column is modified, thereby achieving seamless splicing between the two data. Then, this data is used as the front loop to splice the latter data, and finally a seamless splicing segmentation result of the entire scene image is achieved to obtain the spliced ​​image data.

[0045] In this embodiment of the present invention, step S106 includes the following steps:

[0046] A first determining step is to determine target image data based on the segmented image data and adjacent segmented image data of the segmented image data, wherein the target image data is the segmented image data with the smallest merging cost between the adjacent segmented image data and the segmented image data;

[0047] A second determining step is to construct a node based on each segmented image data and the corresponding target image data, determine an initial node, and insert the initial node into an initial red-black tree, wherein the initial node is an initial node containing the same segmented image data;

[0048] A construction step of determining the node with the smallest merge cost value among the initial nodes as the target node, and removing the initial nodes other than the target node from the initial red-black tree to obtain an intermediate red-black tree;

[0049] A third determining step is to determine a final node based on the intermediate red-black tree, and delete the final node from the intermediate red-black tree to obtain a target red-black tree, wherein the final node is a node at the bottom of the intermediate red-black tree with the smallest merge cost;

[0050] Determining the segmented image data other than the target image data in the adjacent segmented image data as the adjacent segmented image data, and determining the target red-black tree as the initial red-black tree, and repeatedly performing the first determining step, the second determining step, and the constructing step a first preset number of times, wherein the first preset number of times is the number of the segmented image data;

[0051] Merging the segmented image data corresponding to each final node obtained in the first preset number of times to obtain the merged spliced ​​image data;

[0052] Based on the merge cost values ​​corresponding to the final nodes obtained in the first preset number of times, a first cumulative merge cost value of each segmented image data is determined, and the low-level scale set is constructed based on the first cumulative merge cost value.

[0053] In the embodiment of the present invention, first, a merging cost between each segmented image data and its adjacent segmented image data is calculated.

[0054] It should be noted that the calculation formula for the merger cost is as follows:

[0055] C i，j =w c ×Δh c +w s ×Δh s ;

[0056] w c and w s is the coefficient for adjusting the weight of spectral and spatial features, and w c +w s =1,Δh c represents the amount of increase in spectral heterogeneity due to region merging and is defined as;

[0057]

[0058] Where n represents the area of ​​the region, w c represents the weight of the image c band, σ c It represents the standard deviation of the region in the c-band, i and j represent two adjacent regions respectively, and i∪j represents the region obtained by merging regions i and j.

[0059] Δh s It represents the increase in shape heterogeneity caused by region merging, and is composed of two aspects: the shape smoothness and shape compactness of the region.

[0060] Δh s =w smooth ×Δh smooth +w compt ×Δh compt ;

[0061] in:

[0062]

[0063]

[0064] where wcompt +w smooth =1, L represents the perimeter of the area, and b represents the perimeter of the minimum circumscribed rectangle.

[0065] Then, adjacent segmented image data (ie, target image data) with the minimum merging cost corresponding to each segmented image data is determined.

[0066] Loop through each segmented image data and find the segmented image data with the lowest merging cost among all adjacent segmented image data. This creates a segmented image data pair, indicating that this pair of segmented image data is the most suitable for merging. This segmented image data pair forms the initial node and is generated or inserted into the red-black tree to obtain the initial red-black data.

[0067] When constructing node A later, if the segmented image data in the segmented image data pair already exists in the red-black tree node B, the size of the merged cost value of the segmented image data pair is used to determine whether to rebuild the node: when the merged cost value of node A is less than the merged cost value in node B, delete node B in the red-black tree and insert node A into the red-black tree, finally completing the construction of the intermediate red-black tree.

[0068] Specifically, such as Figure 2 As shown, if the merged cost value of segmented image data 6 and segmented image data 7 (node ​​B) is the smallest among the merged cost values ​​of segmented image data 6 and segmented image data 1, 2, 3, 5, 7, 9, 10, and 11, such as 1, then the merged cost value of segmented image data 7 and segmented image data 2, 3, 4, 6, 8, 10, 11, and 12 is calculated to be the smallest among the merged cost values ​​of segmented image data 7 and segmented image data 8 (node ​​A), such as 0.5, then the segmented image data 6 and segmented image data 7 (node ​​B) are deleted, and the segmented image data 7 and segmented image data 8 (node ​​A) are inserted into the red-black tree.

[0069] Then, find the node with the smallest merge cost value at the bottom of the intermediate red-black tree (i.e., the final node), record the segmented image data and the corresponding merge cost value corresponding to the final node, delete the final node from the intermediate red-black tree, and delete the adjacent information of the segmented image data corresponding to the final node.

[0070] Next, the first determining step, the second determining step and the constructing step are repeatedly performed a first preset number of times.

[0071] Finally, merging the segmented image data corresponding to each final node obtained in the first preset number of times to obtain the merged spliced ​​image data;

[0072] Based on the merge cost values ​​corresponding to the final nodes obtained in the first preset number of times, a first cumulative merge cost value of each segmented image data is determined, and the low-level scale set is constructed based on the first cumulative merge cost value.

[0073] In this embodiment of the present invention, step S106 further includes the following steps:

[0074] sorting the first cumulative combined cost values ​​and determining a target cumulative combined cost value corresponding to a preset ratio;

[0075] The segmented image data having a value less than the target cumulative merging cost is merged to obtain the level 0 scale segmented data.

[0076] In the embodiment of the present invention, the first cumulative combined cost values ​​are first sorted, and then the sorting results are intercepted according to a preset ratio to determine a target cumulative combined cost value corresponding to the preset ratio.

[0077] It should be noted that the calculation formula of the preset ratio is as follows:

[0078]

[0079] in, w and H represent the width and height of the remote sensing image to be classified, N represents the number of segmented image data, and S is the expected value of the average area of ​​the segmented image data.

[0080] Then, the segmented image data that is less than the target cumulative merging cost value is merged to obtain the level 0 scale segmentation data.

[0081] In this embodiment of the present invention, step S108 includes the following steps:

[0082] Classifying the level 0 scale segmentation data based on a preset classification algorithm to obtain an initial classification result of the segmented image data contained in the level 0 scale segmentation data;

[0083] Determining the level 0 scale segmentation data as the segmented image data, and repeating the first determining step, the second determining step, and the constructing step a second preset number of times, wherein the second preset number of times is the number of segmented image data included in the level 0 scale segmentation data;

[0084] Based on the merge cost values ​​corresponding to the final nodes obtained in the second preset number of times, a second cumulative merge cost value of the segmented image data included in the level 0 scale segmentation data is determined, and the high-level scale set is constructed based on the second cumulative merge cost value.

[0085] In the embodiment of the present invention, first, the level 0 scale segmentation data is classified based on a preset classification algorithm to obtain an initial classification result of the segmented image data contained in the level 0 scale segmentation data.

[0086] The following example takes the land cover classification of level 0 scale segmentation data as an example:

[0087] First, sample selection is performed based on the level 0 segmentation data. In this implementation, samples are selected based on land cover type, which includes five types: buildings, roads, cultivated land, forest land, and water bodies. Feature data includes four categories: spectral features, texture features, shape features, and index features. Spectral features include mean, standard deviation, and histogram; texture features include mean, variance, homogeneity, contrast, dissimilarity, entropy, angular second moment, and correlation; shape features include area, length, shape index, rectangle metric, proportion, density, and circle features; and index features include NDVI and NDWI.

[0088] Then, based on the preset classification algorithm and feature data, the land cover classification is performed on the level 0 scale segmentation data to obtain the initial classification results.

[0089] After obtaining the initial classification result, determining the level 0 scale segmentation data as the segmented image data, and repeating the first determining step, the second determining step, and the constructing step a second preset number of times, wherein the second preset number of times is the number of segmented image data included in the level 0 scale segmentation data;

[0090] Based on the merge cost values ​​corresponding to the final nodes obtained in the second preset number of times, a second cumulative merge cost value of the segmented image data included in the level 0 scale segmentation data is determined, and a high-level scale set is constructed based on the second cumulative merge cost value.

[0091] In this embodiment of the present invention, step S110 includes the following steps:

[0092] Merging the level 0 scale segmentation data based on the scale parameter and the high-level scale set to obtain a merged result;

[0093] The classification result is determined based on the combined result and the initial classification result.

[0094] After obtaining the scale parameter input by the user, the level 0 scale segmentation data is merged upward according to the scale parameter input by the user and the high-level scale set to obtain merged data, and then the classification result is determined according to the merged data and the initial classification result.

[0095] In an embodiment of the present invention, when performing object-oriented classification on remote sensing images, when the scale is selected from bottom to top, region merging transmits category information, and the category information of the merged object is determined by the area size of the object category at the lowest scale, thereby correcting the problem of incorrect classification of level 0 objects.

[0096] During the region merging process, a certain scale can ensure the integrity and purity of the object patch, so the problem of the optimal scale can be solved through interaction.

[0097] Region merging itself is based on the characteristics of adjacent objects for merging judgment. Therefore, region merging can be used to equivalently replace the application of spatial autocorrelation features in classification, avoiding the problem of strong uncertainty in spatial autocorrelation caused by different spatial distributions of sample data, thereby improving the accuracy of classification results.

[0098] Example 2:

[0099] An embodiment of the present invention also provides an object-oriented classification device based on a scale set, which is used to execute the object-oriented classification method based on a scale set provided in the above content of the embodiment of the present invention. The following is a detailed introduction to the device provided by the embodiment of the present invention.

[0100] like Figure 3 As shown, Figure 3 Schematic diagram of the above-mentioned object-oriented classification device based on scale sets, which includes: an acquisition unit 10, a splicing unit 20, a first construction unit 30, a second construction unit 40 and a classification unit 50.

[0101] The acquisition unit 10 is used to acquire block image data of the remote sensing image to be classified, and segment the block image data to obtain segmented image data, wherein there is an overlapping area between any two adjacent block image data;

[0102] The stitching unit 20 is used to seamlessly stitch the segmented image data to obtain stitched image data;

[0103] The first construction unit 30 is configured to perform hierarchical region merging on the stitched image data, construct a low-level scale set, and determine level 0 scale segmentation data in the low-level scale set;

[0104] The second construction unit 40 is configured to construct a high-level scale set based on the level 0 scale segmentation data;

[0105] The classification unit 50 is configured to obtain a scale parameter input by a user, and classify the remote sensing image to be classified based on the scale parameter and the high-level scale set to obtain a classification result.

[0106] In an embodiment of the present invention, segmented image data of a remote sensing image to be classified is obtained, and the segmented image data is segmented to obtain segmented image data, wherein there is an overlapping area between any two adjacent segmented image data; the segmented image data is seamlessly spliced ​​to obtain spliced ​​image data; hierarchical region merging is performed on the spliced ​​image data, and a low-level scale set is constructed, and level 0 scale segmented data in the low-level scale set is determined; a high-level scale set is constructed based on the level 0 scale segmented data; a scale parameter input by a user is obtained, and the remote sensing image to be classified is classified based on the scale parameter and the high-level scale set to obtain a classification result, thereby achieving the purpose of classifying remote sensing images using a two-level scale set, thereby solving the technical problem of low classification accuracy and classification efficiency of the existing object-oriented classification method, thereby achieving the technical effect of improving the classification accuracy and classification efficiency of the object-oriented classification method.

[0107] Example 3:

[0108] An embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein the memory is used to store a program that supports the processor to execute the method described in the above embodiment 1, and the processor is configured to execute the program stored in the memory.

[0109] See also Figure 4 An embodiment of the present invention further provides an electronic device 100, comprising: a processor 60, a memory 61, a bus 62 and a communication interface 63, wherein the processor 60, the communication interface 63 and the memory 61 are connected via the bus 62; the processor 60 is used to execute an executable module stored in the memory 61, such as a computer program.

[0110] The memory 61 may include high-speed random access memory (RAM) and may also include non-volatile memory, such as at least one disk storage. The system network element communicates with at least one other network element via at least one communication interface 63 (which may be wired or wireless), and may utilize the Internet, a wide area network, a local area network, a metropolitan area network, or the like.

[0111] The bus 62 may be an ISA bus, a PCI bus, or an EISA bus. The bus may be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, Figure 4 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0112] Among them, the memory 61 is used to store programs, and the processor 60 executes the program after receiving the execution instruction. The method executed by the device for flow process definition disclosed in any embodiment of the above-mentioned embodiment of the present invention can be applied to the processor 60 or implemented by the processor 60.

[0113] The processor 60 may be an integrated circuit chip with signal processing capabilities. During implementation, the steps of the above method may be performed by hardware integrated logic circuits or software instructions within the processor 60. The processor 60 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It may implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of the present invention may be directly executed by a hardware decoding processor or by a combination of hardware and software modules within the decoding processor. The software modules may be located in storage media well-known in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or the like. The storage medium is located in the memory 61 , and the processor 60 reads the information in the memory 61 and completes the steps of the above method in combination with its hardware.

[0114] Example 4:

[0115] An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method described in the above embodiment 1 are executed.

[0116] In addition, in the description of the embodiments of the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; they may refer to mechanical connections or electrical connections; they may refer to direct connections or indirect connections through an intermediate medium; and they may refer to internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on the specific circumstances.

[0117] In the description of the present invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0118] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. There may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed may be through some communication interface, indirect coupling or communication connection of devices or units, which may be electrical, mechanical or other forms.

[0119] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0120] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0121] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the above-described embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-described embodiments within the technical scope disclosed by the present invention, or replace some of the technical features therein with equivalents. Such modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.

Claims

1. An object-oriented classification method based on a scale set, characterized in that: include: Obtaining block image data of the remote sensing image to be classified, and segmenting the block image data to obtain segmented image data, wherein there is an overlapping area between any two adjacent block image data; seamlessly splicing the segmented image data to obtain spliced ​​image data; Performing hierarchical region merging on the stitched image data, constructing a low-level scale set, and determining level 0 scale segmentation data in the low-level scale set; Based on the level 0 scale segmentation data, a high-level scale set is constructed; Obtaining a scale parameter input by a user, and classifying the remote sensing image to be classified based on the scale parameter and the high-level scale set to obtain a classification result; Perform hierarchical region merging on the stitched image data and construct a low-level scale set, including: A first determining step is to determine target image data based on the segmented image data and adjacent segmented image data of the segmented image data, wherein the target image data is the segmented image data with the smallest merging cost between the adjacent segmented image data and the segmented image data; A second determining step is to construct a node based on each segmented image data and the corresponding target image data, determine an initial node, and insert the initial node into an initial red-black tree, wherein the initial node is an initial node containing the same segmented image data; A construction step of determining the node with the smallest merge cost value among the initial nodes as the target node, and removing the initial nodes other than the target node from the initial red-black tree to obtain an intermediate red-black tree; A third determining step is to determine a final node based on the intermediate red-black tree, and delete the final node from the intermediate red-black tree to obtain a target red-black tree, wherein the final node is a node at the bottom of the intermediate red-black tree with the smallest merge cost; Determining the segmented image data other than the target image data in the adjacent segmented image data as the adjacent segmented image data, and determining the target red-black tree as the initial red-black tree, and repeatedly performing the first determining step, the second determining step, and the constructing step a first preset number of times, wherein the first preset number of times is the number of the segmented image data; Merging the segmented image data corresponding to each final node obtained in the first preset number of times to obtain merged stitched image data; Determining a first cumulative merging cost value for each segmented image data based on the merging cost value corresponding to each final node obtained in the first preset number of times, and constructing the low-level scale set based on the first cumulative merging cost value; Determining the level 0 scale segmentation data in the low-level scale set includes: sorting the first cumulative combined cost values ​​and determining a target cumulative combined cost value corresponding to a preset ratio; Merging the segmented image data smaller than the target cumulative merging cost value to obtain the level 0 scale segmentation data; Based on the level 0 scale segmentation data, a high-level scale set is constructed, including: Classifying the level 0 scale segmentation data based on a preset classification algorithm to obtain an initial classification result of the segmented image data contained in the level 0 scale segmentation data; Determining the level 0 scale segmentation data as the segmented image data, and repeating the first determining step, the second determining step, and the constructing step a second preset number of times, wherein the second preset number of times is the number of segmented image data included in the level 0 scale segmentation data; Based on the merge cost values ​​corresponding to the final nodes obtained in the second preset number of times, a second cumulative merge cost value of the segmented image data included in the level 0 scale segmentation data is determined, and the high-level scale set is constructed based on the second cumulative merge cost value.

2. The method according to claim 1, characterized in that If the classification type for classifying the level 0 scale segmentation data is land cover classification, the level 0 scale segmentation data is classified based on a preset classification algorithm to obtain an initial classification result of the segmented image data contained in the level 0 scale segmentation data, including: Based on the land cover classification, acquiring feature data of the level 0 scale segmentation data, wherein the feature data includes: spectral features, texture features, shape features and index features; Based on the preset classification algorithm and the feature data, land cover classification is performed on the level 0 scale segmentation data to obtain the initial classification result.

3. The method according to claim 1, characterized in that Before obtaining the block image data of the remote sensing image to be classified, the method includes: Acquire the remote sensing image to be classified, and segment the remote sensing image to be classified according to a preset size to obtain initial block image data; Determining, based on the boundaries of the initial block image data, an overlapping area between the initial block image data and adjacent initial block image data of the initial block image data; The block image data is constructed based on the overlapping area and the initial block image data.

4. The method according to claim 2, characterized in that Classifying the remote sensing image to be classified based on the scale parameter and the high-level scale set to obtain a classification result, including: Merging the level 0 scale segmentation data based on the scale parameter and the high-level scale set to obtain a merged result; The classification result is determined based on the combined result and the initial classification result.

5. An object-oriented classification device based on a scale set, characterized in that: include: Acquisition unit, splicing unit, first building unit, second building unit and classification unit, wherein, The acquisition unit is used to acquire block image data of the remote sensing image to be classified, and segment the block image data to obtain segmented image data, wherein there is an overlapping area between any two adjacent block image data; The stitching unit is configured to seamlessly stitch the segmented image data to obtain stitched image data; The first construction unit is configured to perform hierarchical region merging on the stitched image data, construct a low-level scale set, and determine level 0 scale segmentation data in the low-level scale set; The second construction unit is configured to construct a high-level scale set based on the level 0 scale segmentation data; The classification unit is configured to obtain a scale parameter input by a user, and classify the remote sensing image to be classified based on the scale parameter and the high-level scale set to obtain a classification result; The first building block comprises: a first determining unit configured to determine target image data based on the segmented image data and adjacent segmented image data of the segmented image data, wherein the target image data is segmented image data having the smallest merging cost with the segmented image data among the adjacent segmented image data; A second determining unit is configured to construct a node based on each segmented image data and the corresponding target image data, determine an initial node, and insert the initial node into an initial red-black tree, wherein the initial node is an initial node containing the same segmented image data; A construction unit is configured to determine a node with the smallest merge cost among the initial nodes as a target node, and remove the initial nodes other than the target node from the initial red-black tree to obtain an intermediate red-black tree; a third determining unit, configured to determine a final node based on the intermediate red-black tree, and delete the final node from the intermediate red-black tree to obtain a target red-black tree, wherein the final node is a node at the bottom of the intermediate red-black tree with the smallest merge cost; a first executing unit, configured to determine the segmented image data other than the target image data in the adjacent segmented image data as the adjacent segmented image data, and to determine the target red-black tree as the initial red-black tree, and to repeatedly execute the first determining unit, the second determining unit, and the constructing unit a first preset number of times, wherein the first preset number of times is the number of the segmented image data; a merging execution unit, configured to merge the segmented image data corresponding to each final node obtained in the first preset number of times to obtain merged spliced ​​image data; A construction execution unit is configured to determine a first cumulative merging cost value of each segmented image data based on the merging cost value corresponding to each final node obtained in the first preset number of times, and to construct the low-level scale set based on the first cumulative merging cost value; In the process of determining the level 0 scale segmentation data in the low-level scale set, the first construction unit is further configured to: sort the first cumulative merging cost values ​​and determine a target cumulative merging cost value corresponding to a preset ratio; merge the segmented image data that is less than the target cumulative merging cost value to obtain the level 0 scale segmentation data; The second construction unit is further used to: classify the level 0 scale segmentation data based on a preset classification algorithm to obtain an initial classification result of the segmented image data contained in the level 0 scale segmentation data; determine the level 0 scale segmentation data as the segmented image data, and repeatedly execute the first determination unit, the second determination unit, and the construction unit a second preset number of times, wherein the second preset number of times is the number of segmented image data contained in the level 0 scale segmentation data; determine a second cumulative merged cost value of the segmented image data contained in the level 0 scale segmentation data based on the merged cost value corresponding to each final node obtained in the second preset number of times, and construct the high-level scale set based on the second cumulative merged cost value.

6. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory is used to store a program for supporting the processor to execute the method according to any one of claims 1 to 4, and the processor is configured to execute the program stored in the memory.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are executed.

Images