Landform boundary estimation and classification method based on spatial composition structure information

Abstract

The invention discloses a geomorphic form boundary estimation and classification method based on spatial composition structure information, which relates to the technical field of digital geomorphic form classification, and comprises the following steps: step 1, extracting geomorphic form elements; 2, acquiring space composition structure information to form a two-dimensional sample space; 3, on the basis of the formed two-dimensional sample space, deducting and estimating the landform boundary to obtain a corresponding boundary line and a segmentation region; 4, according to the topological relation of the spatial composition structure information of the landform, carrying out classification and drawing on the areas divided by the boundary lines; according to the method provided by the invention, expert priori knowledge such as difference of spatial composition structure information between landform types and spatial distribution and topological relation of the landform types is fully utilized, the problems of selection of landform factors, optimal analysis scale and the like do not need to be considered, and the overall process is simple, convenient, effective and high in operability.

Description

Technical Field

[0001] The present invention relates to the technical field of digital landform morphology classification, and in particular to a landform morphology boundary estimation and classification method based on spatial composition structure information. Background Art

[0002] Geomorphology is an important description of geomorphic types, and is crucial for revealing the causes and evolution of geomorphology, as well as for geomorphic mapping and other applications. With the continuous enrichment of raster DEM (digital elevation model) data, the digital geomorphic morphological analysis methods based on it have also been continuously developed, which provides a technical basis for the automatic classification of geomorphic morphology.

[0003] At present, there are many methods for automatic classification of landforms. In summary, they are mainly divided into three categories: clustering-based methods, rule-based methods, and typical sample-based methods. These traditional methods can basically meet the automatic classification of basic landform types, namely landform elements, but they perform poorly in the automatic classification of composite landforms with multi-level relationships and strong spatial heterogeneity. On the one hand, it is difficult for traditional methods to effectively establish a reasonable attribute space set or give a clear constraint rule knowledge for automatic classification. On the other hand, traditional methods also have a common defect, that is, they lack sufficient consideration of the spatial composition structure and cannot reflect the spatial hierarchical combination relationship of landforms. Therefore, traditional methods are difficult to apply to the automatic classification of composite landforms. Therefore, a landform boundary estimation and classification method based on spatial composition structure information is urgently needed to overcome the problems of existing methods. Summary of the invention

[0004] The purpose of the present invention is to provide a method for estimating and classifying landform boundaries based on spatial composition structure information, so as to solve the problems existing in the prior art that it is difficult to effectively establish a reasonable attribute space set or provide a clear constraint rule knowledge for automatic classification, and it is difficult to reflect the spatial hierarchical combination relationship of landforms.

[0005] To achieve the above object, the present invention provides a method for estimating and classifying landform boundaries based on spatial composition structure information, comprising the following steps: Step 1: Extract geomorphic elements; Step 2: Obtain spatial composition structure information to form a two-dimensional sample space; Step 3: Based on the two-dimensional sample space formed in step 2, estimate the landform boundary to obtain the corresponding boundary line and segmentation area; Step 4: Classify and map the areas divided by the boundary lines according to the topological relationship of the spatial composition structure information of the landform.

[0006] Preferably, the process of extracting landform elements in step 1 is as follows: S11, setting a number of analysis scale windows of different scales, and then identifying terrain elements through each analysis scale window; S12, constructing a set of element classification results for each position according to the recognition results of the terrain elements in S11; S13. Use the comprehensive evaluation method to determine the final element type at each position based on the principle of the element with the highest frequency of occurrence. The expression is as follows: (1) In the formula, represents the analysis scale, Indicates at location On a single scale of analysis The element type below, Indicates at location The final element type of the .

[0007] Preferably, the process of obtaining the spatial composition structure information in step 2 to form a two-dimensional sample space is as follows: S21. Outline an approximate dividing line at a location where a boundary may exist based on the expert's prior knowledge, establish a buffer zone for the dividing line, and limit the calculation area; S22, obtaining element type and elevation information of each position in the fuzzy boundary buffer; S23, storing the element types at the positions with the same elevation as a sub-information set respectively; S24. Arrange the sub-information sets at each elevation in the order of elevation change to form a two-dimensional sample space.

[0008] Preferably, in step 3, based on the two-dimensional sample space formed in step 2, the process of estimating the landform boundary and obtaining the corresponding boundary line and segmented area is as follows: S31, respectively calculating information entropy for sub-information sets of elements located at the same elevation, so as to transform the two-dimensional information in the two-dimensional sample space into one-dimensional information; S32, setting different elevations as demarcation thresholds, and calculating the second-order information entropy for the information entropy values ​​on both sides of the demarcation threshold; S33, calculating the sum of the second-order information entropy on both sides of the demarcation threshold; S34, traversing and calculating the one-dimensional information generated in S31 in the order of elevation, and taking the elevation line with the maximum sum of the second-order information entropy values ​​on both sides as the corresponding boundary line.

[0009] Preferably, in S31, the expression for calculating the information entropy of the sub-information sets of elements at the same elevation is as follows: (2) In the formula, are the element types located at the same elevation, is the total number of element types at the same elevation, For element type The probability statistics of It represents the information entropy of the elements at the same elevation. Indicates the total number of elements of type M.

[0010] Preferably, in S32, the calculation expressions for the second-order information entropy of the information entropy values ​​on both sides of the demarcation threshold are as follows: (3) In the formula, is the total number of elements on one side of the dividing threshold that make up the information entropy, The information entropy of the elements The probability statistics of The information entropy of the second-order elements on one side of the dividing threshold is represented by It represents the total number of elements whose information entropy is E(M).

[0011] Preferably, the calculation expression of the sum of the second-order information entropy on both sides of the demarcation threshold in S33 is as follows: (4) In the formula, Indicates that it is less than the demarcation threshold The entropy value of Indicates that it is greater than the demarcation threshold The entropy value of The demarcation threshold is In addition, the calculation result of formula (4) can also give the uncertainty spatial distribution of the boundary position estimation results under different threshold conditions.

[0012] Preferably, the topological relationship of the spatial composition structure information of the landform is specifically spatial adjacency and spatial inclusion.

[0013] Therefore, the present invention adopts the above-mentioned landform boundary estimation and classification method based on spatial composition structure information, which has the following beneficial effects: (1) The method proposed in the present invention realizes the classification of geomorphic types from the perspective of the differences in spatial composition structure between composite geomorphic types, based on the consideration of the spatial composition structure of geomorphic types. The whole process does not require the deriving of a large number of geomorphic factors to construct a complete terrain attribute set or knowledge rules, so there is no need to consider a series of issues such as the selection of geomorphic factors and the scale of suitability analysis. Relatively speaking, the method of the present invention is the most simple and effective. (2) The present invention uses the line with the smallest uncertainty as the dividing line between types based on the boundary estimation results, and then converts the range contained in the dividing line into surface data, and assigns morphological types to the surface data based on the prior knowledge of the spatial topological relationship of the composite landform type, and finally obtains a landform morphological type classification map. The overall classification result is more in line with the actual situation; and the classification method of the present invention performs well in the composite landform morphological type classification application, can obtain ideal landform morphological classification results, and can meet the needs of composite landform mapping applications.

[0014] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is an overall flow chart of a landform morphology boundary estimation and classification method based on spatial composition structure information of the present invention; Figure 2 The fuzzy boundary calculation area setting diagram of the embodiment of the present invention, wherein (a) is the set fuzzy boundary map, and (b) is the composition map of the terrain elements in the landform morphological types on both sides covered in the calculation area; Figure 3 A two-dimensional sample space diagram of element types within the fuzzy boundary region of an embodiment of the present invention; Figure 4 A complexity distribution diagram of element types at different elevations according to an embodiment of the present invention; Figure 5 1 is a boundary position estimation diagram under different elevation threshold conditions of an embodiment of the present invention, wherein (a) is a second-order information entropy statistical distribution diagram, and (b) is a boundary position estimation uncertainty spatial distribution diagram; Figure 6 A topological relationship diagram of composite landform types according to an embodiment of the present invention; Figure 7 A distribution map of the research area of ​​an embodiment of the present invention; Figure 8 Result diagrams of the estimated position of the landform boundary of the study area according to the embodiment of the present invention, wherein (a) is a result diagram of the element type classification, and (b) is a spatial distribution diagram of the uncertainty of the estimated position of the landform boundary; Fig. 9 This is a classification result map of the landform type of the study area in an embodiment of the present invention, wherein (a) is the classification result map, and (b) is the correction result map. DETAILED DESCRIPTION

[0016] The following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] See also Figure 1-Figure 9 , a landform boundary estimation and classification method based on spatial composition structure information, comprising the following steps: Step 1: Extract geomorphic elements. For the identification of geomorphic elements, the present invention uses the Geomorphons method to exhaustively enumerate the geometric morphology of the study area and establish a mapping relationship with the corresponding geomorphic elements. On this basis, a multi-scale comprehensive identification of the element identification results at different analysis scales is adopted to determine the final element type based on the principle of the highest element occurrence frequency. The specific process is as follows: S11, setting a number of analysis scale windows of different scales, and then identifying terrain elements through each analysis scale window; S12, constructing a set of element classification results for each position according to the recognition results of the terrain elements in S11; S13. Use the comprehensive evaluation method to determine the final element type at each position based on the principle of the element with the highest frequency of occurrence. The expression is as follows: (1) In the formula, represents the analysis scale, Indicates at location On a single scale of analysis The element type below, Indicates at location The final element type of the .

[0018] Step 2: Obtain spatial composition structure information to form a two-dimensional sample space; the specific process is as follows: S21. Based on the expert's prior knowledge, an approximate dividing line is outlined at the location where the boundary may exist, and a buffer zone of the dividing line is established to limit the calculation area. The width of the buffer zone should be set to at least cover half of the internal area of ​​the morphological type or ensure that the covered area can reflect the composition structure information of the element type within the type, such as Figure 2 The dark solid line is the fuzzy boundary set according to the expert prior knowledge, and the range encircled by the light solid line is the fuzzy boundary calculation area set according to the fuzzy boundary and width parameters d1 and d2; S22, obtaining element type and elevation information of each position in the fuzzy boundary buffer; S23, storing the element types at the positions with the same elevation as a sub-information set respectively; S24, arrange the sub-information sets at each elevation in the order of elevation change to form a two-dimensional sample space, such as Figure 3 shown.

[0019] Step 3: Based on the two-dimensional sample space constructed in step 2, estimate the landform boundary to obtain the corresponding boundary line and segmentation area; the specific process is as follows: S31, respectively calculate the information entropy of the sub-information sets of elements located at the same elevation, and transform the two-dimensional information in the two-dimensional sample space into one-dimensional information, such as Figure 4 As shown; the expressions for calculating the information entropy of the sub-information sets of elements at the same elevation are as follows: (2) In the formula, are the element types located at the same elevation, is the total number of element types at the same elevation, For element type The probability statistics of It represents the information entropy of the elements at the same elevation. Indicates the total number of elements of type M.

[0020] S32, different elevations are set as demarcation thresholds, and the second-order information entropy is calculated for the information entropy values ​​on both sides of the demarcation threshold respectively; the specific calculation expression is as follows: (3) In the formula, is the total number of elements on one side of the dividing threshold that make up the information entropy, The information entropy of the elements The probability statistics of The information entropy of the second-order elements on one side of the dividing threshold is represented by It represents the total number of elements whose information entropy is E(M).

[0021] S33, calculate the sum of the second-order information entropy on both sides of the demarcation threshold; the specific calculation expression is as follows: (4) In the formula, Indicates that it is less than the demarcation threshold The entropy value of Indicates that it is greater than the demarcation threshold The entropy value of The demarcation threshold is In addition, the calculation result of formula (4) can also give the uncertainty spatial distribution of the boundary position estimation results under different threshold conditions; S34, the one-dimensional information generated in S31 is traversed and calculated in the order of elevation, and the elevation line with the maximum sum of the second-order information entropy values ​​on both sides is taken as the corresponding boundary line, such as Figure 5 As shown in the figure, (a) is the statistical distribution of the sum of the second-order information entropy on both sides of the dividing threshold, and (b) is the spatial distribution of the estimated uncertainty of the dividing position. The color changes from red to blue, indicating that the uncertainty changes from low to high. Among them, the elevation threshold with the smallest uncertainty is the dividing position between types.

[0022] Step 4: Classify and map the areas divided by the boundary lines according to the topological relationship of the spatial composition structure information of the landform; wherein the topological relationship of the spatial composition structure information of the landform is specifically spatial adjacency and spatial inclusion. The present invention generalizes the topological relationship of the spatial structure composition of the composite landform based on the description of the composite landform structure and spatial hierarchy and combines expert knowledge, such as Figure 6 As shown in the figure, 1 represents the reef slope, 2 represents the reef flat, and 3 represents the lagoon; (1) represents the outer reef flat, (2) represents the reef bulge, (3) represents the inner reef flat, (4) represents the lagoon slope, (5) represents the lagoon bottom, and (6) represents the point reef; I represents the sparse coral zone, II represents the coral cluster zone, and III represents the reef pit development zone. The spatial distribution of the three types of landforms, the reef slope, the reef flat, and the lagoon, all have the characteristics of seaward order and a landform type classification system of at least two levels, and the topological relationship between the landform types is mainly spatial adjacency and inclusion. According to the topological relationship of the composite landform morphological type, the landforms separated by the estimated boundaries can be divided into certain landform morphological types.

[0023] Example Taking the application of coral atoll landform classification as an example, the landform classification method proposed in the present invention is applied for analysis and accuracy evaluation.

[0024] The research area selected in this embodiment is a typical coral atoll located in the southern part of the Xisha Islands in the South China Sea, which includes three primary landform types, namely reef flats, lagoons and point reefs. Among them, the reef flat only covers the secondary landform types of coral cluster belts and reef pit development belts close to the lagoon slope. The lagoon contains two secondary landform types, the lagoon slope and the lagoon bottom. The morphological characteristics and terrain complexity of different landform types are very different, and the four landform types of coral cluster belts, lagoon slopes, lagoon bottoms and point reefs are relatively typical in terms of morphological characteristics and component compositions and differ greatly from each other. Therefore, this embodiment selected some typical areas of the above four landform types as test areas, such as Figure 7As shown in the figure, it represents a DEM data. The elevation in the figure includes the coral cluster belt, lagoon slope, lagoon bottom and point reef from high to low. The specific landform classification system and definition description of the study area are shown in Table 1: Table 1 Geomorphic classification system within the study area ; Results and Analysis: 1. Multi-scale comprehensive identification results of geomorphic elements The results of multi-scale comprehensive identification of geomorphological elements in the study area are as follows: Figure 8 As shown in (a). The coral biomass belt with shallow water depth is covered with coral reefs, so the surface morphology is more complex and the elemental composition is more fragmented. The lagoon bottom with deeper water depth also has a more fragmented elemental composition, but due to the influence of the distribution of point reefs, the valley-type element distribution area is relatively large. In addition, due to the uneven surface of the reef pit development zone, the elemental composition is also relatively fragmented. Compared with the reef pit development zone, the lagoon slope is mainly a slope of sandy debris accumulation, so the elemental composition is relatively simple, mainly dominated by the slope. The point reef is hilly as a whole, so the element type is also relatively simple, mainly slope. From the above results, it can be seen that the spatial distribution of elemental type composition has typical spatial differentiation characteristics. This spatial differentiation is consistent with the spatial distribution of atoll geomorphological types.

[0025] 2. Estimated uncertainty distribution results of each demarcation threshold Based on the element recognition results and combined with the expert prior knowledge, the maximum second-order entropy estimation model was used to estimate the fuzzy area at the boundary position, and the estimated uncertainty distribution results of each demarcation threshold were obtained, such as Figure 8 As shown in (b). In the location area where the element type composition has a clear turning point, the uncertainty of the boundary estimation is relatively low. In the location area where the element type composition does not change significantly, the uncertainty of the boundary estimation is relatively high, especially inside the sand slope where the bottom composition and morphology are relatively homogeneous. In the location area with weak spatial heterogeneity of element composition, the high and low distribution of uncertainty in boundary estimation is relatively concentrated. In the location area with strong spatial heterogeneity of element composition, the high and low distribution of uncertainty in boundary estimation is relatively scattered. The uncertainty results of the boundary estimation between the interior of the lagoon and the point reef are distributed in a discrete ring shape, while the uncertainty results of the boundary estimation from the coral cluster zone to the bottom of the lagoon show an intermittent high and low distribution. It can also be seen from the above results that the spatial distribution law of the uncertainty of the boundary position estimation conforms to the characteristics of the atoll geomorphological type and the spatial topological relationship.

[0026] 3. Landform classification results The present invention uses the line with the least uncertainty as the dividing line between types based on the boundary estimation results. Then the range contained in the dividing line is converted into surface data, and the surface data is assigned a morphological type based on the prior knowledge of the spatial topological relationship of the atoll landform type, and finally a landform morphological type classification map is obtained, as shown in Fig. 9 As shown in (a) of the classification diagram, it can be seen from the coverage of each type in the classification diagram that the overall classification result is consistent with the actual situation. After the classification is completed, this example corrects and adjusts the type boundary position based on the sampling data obtained from the field survey to obtain the final classification result, as shown in Fig. 9 as shown in (b).

[0027] 4. Classification accuracy evaluation In this example, a total of 27,233 sample points were selected based on the field survey data and the boundary buffer range. The confusion matrix and Kappa coefficient were used to evaluate the accuracy of the classification results, as shown in Table 2: Table 2 Comparison of classification results and sample data ; As can be seen from Table 2, the mapping accuracy of point reefs, coral clusters, lagoon slopes and lagoon bottoms is over 93%, and the user accuracy is between 91% and 100%. The user accuracy of the reef pit development zone is 100%, and the mapping accuracy is 86%. The overall accuracy of the landform morphology classification result is 94%, and the Kappa coefficient is 0.91, and the classification result is good. The above results show that the method of the present invention performs well in the classification application of coral reef composite landform morphology, and can obtain ideal landform morphology classification results, which can meet the needs of landform mapping.

[0028] In summary, in the application of composite landform morphological type classification, the landform morphological boundary estimation and classification method based on spatial composition information proposed in the present invention can maintain a relatively ideal accuracy in the landform morphological classification results and can meet the needs of composite landform mapping.

[0029] Therefore, the present invention adopts the above-mentioned method of estimating and classifying geomorphic boundaries based on spatial composition structure information, which makes full use of the differences in spatial composition structure information between geomorphic morphological types and expert prior knowledge such as the spatial distribution and topological relationship of geomorphic morphological types. There is no need to consider issues such as the selection of geomorphic morphological factors and the optimal analysis scale. The overall process is simple, effective and highly operational.

[0030] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for estimating and classifying landform boundaries based on spatial composition structure information, characterized in that: The following steps are involved: Step 1: Extract geomorphic elements; Step 2: Obtain spatial composition structure information to form a two-dimensional sample space; Step 3: Based on the two-dimensional sample space formed in step 2, estimate the landform boundary to obtain the corresponding boundary line and segmentation area; Step 4: Classify and map the areas divided by the boundary lines according to the topological relationship of the spatial composition structure information of the landform.

2. A method for estimating and classifying landform boundaries based on spatial composition structure information according to claim 1, characterized in that: The process of extracting geomorphic elements in step 1 is as follows: S11, setting a number of analysis scale windows of different scales, and then identifying terrain elements through each analysis scale window; S12, constructing a set of element classification results for each position according to the recognition results of the terrain elements in S11; S13. Use the comprehensive evaluation method to determine the final element type at each position based on the principle of the element with the highest frequency of occurrence. The expression is as follows: （1） In the formula, represents the analysis scale, Indicates at location On a single scale of analysis The element type below, Indicates at location The final element type of the .

3. A method for estimating and classifying landform boundaries based on spatial composition structure information according to claim 2, characterized in that: In step 2, the process of obtaining the spatial composition structure information and forming a two-dimensional sample space is as follows: S21. Outline an approximate dividing line at a location where a boundary may exist based on the expert's prior knowledge, establish a buffer zone for the dividing line, and limit the calculation area; S22, obtaining element type and elevation information of each position in the fuzzy boundary buffer; S23, storing the element types at the positions with the same elevation as a sub-information set respectively; S24. Arrange the sub-information sets at each elevation in the order of elevation change to form a two-dimensional sample space.

4. The method for estimating and classifying landform boundaries based on spatial composition structure information according to claim 3, characterized in that: In step 3, based on the two-dimensional sample space constructed in step 2, the landform boundary is estimated to obtain the corresponding boundary line and segmentation area as follows: S31, respectively calculating information entropy for sub-information sets of elements located at the same elevation, so as to transform the two-dimensional information in the two-dimensional sample space into one-dimensional information; S32, setting different elevations as demarcation thresholds, and calculating the second-order information entropy for the information entropy values ​​on both sides of the demarcation threshold; S33, calculating the sum of the second-order information entropy on both sides of the demarcation threshold; S34, traversing and calculating the one-dimensional information generated in S31 in the order of elevation, and taking the elevation line with the maximum sum of the second-order information entropy values ​​on both sides as the corresponding boundary line.

5. A method for estimating and classifying landform boundaries based on spatial composition structure information according to claim 4, characterized in that: The expressions for calculating the information entropy of the sub-information sets of elements at the same elevation in S31 are as follows: （2） In the formula, are the element types located at the same elevation, is the total number of element types at the same elevation, For element type The probability statistics of It represents the information entropy of the elements at the same elevation. Indicates the total number of elements of type M.

6. A method for estimating and classifying landform boundaries based on spatial composition structure information according to claim 5, characterized in that: In S32, the calculation expressions for the second-order information entropy of the information entropy values ​​on both sides of the demarcation threshold are as follows: （3） In the formula, is the total number of elements on one side of the dividing threshold that make up the information entropy, The information entropy of the elements The probability statistics of The information entropy of the second-order elements on one side of the dividing threshold is represented by The information entropy of element composition is The total quantity.

7. A method for estimating and classifying landform boundaries based on spatial composition structure information according to claim 6, characterized in that: The calculation expression of the sum of the second-order information entropy on both sides of the demarcation threshold in S33 is as follows: （4） In the formula, Indicates that it is less than the demarcation threshold The entropy value of Indicates that it is greater than the demarcation threshold The entropy value of The demarcation threshold is The sum of the entropy values ​​at time .

8. The method for estimating and classifying landform boundaries based on spatial composition structure information according to claim 7, characterized in that: The topological relationship of the spatial composition structure information of landforms is specifically spatial adjacency and spatial inclusion.

Images