Geographic information acquisition system and method based on remote sensing image

Abstract

The invention relates to the technical field of geographic information systems, in particular to a geographic information acquisition system and method based on remote sensing images, and the system comprises a remote sensing image analysis module, a feature data extraction module, a classification matching module and a geographic data analysis module. According to the method, spectral reflection analysis is carried out on all pixels, color values are extracted in different wave band ranges, feature analysis is carried out, texture feature values containing geological information are extracted, and different land cover types can be identified more accurately; effective matching between the image units and the geographic classification labels is realized by calculating the weighted Euclidean distance between the feature vector of each image unit and each label classification vector; by extracting geographic classification label matching results, various types of geographic information data are counted, geographic distribution characteristics can be known, and data support is provided for subsequent geographic planning and resource utilization.

Description

Technical Field

[0001] The present invention relates to the technical field of geographic information systems, and in particular to a system and method for collecting geographic information based on remote sensing images. Background Art

[0002] Geographic Information System is a comprehensive technical field involving the collection, storage, management, analysis and display of geographic spatial data. Remote sensing technology collects and records electromagnetic wave information (such as visible light, infrared, microwave, etc.) from target objects through sensors on remote sensing platforms such as satellites, aircraft, and drones. These sensors can detect various features of the earth's surface, such as terrain, landforms, vegetation, water bodies, soil, etc., and convert them into digital signals or images. Remote sensing data has the characteristics of large area, rapid acquisition, multi-phase, and multi-spectral, and can provide rich information about the earth's surface.

[0003] However, there are still many problems with existing geographic information system technology, especially in the recognition, segmentation and classification of high-resolution remote sensing images.

[0004] The limitations of traditional remote sensing image recognition methods based on single pixels can easily lead to classification errors. These methods rely on a single data source for feature extraction and do not integrate multiple features such as spectrum and texture for feature extraction, which limits the accuracy and effectiveness of remote sensing image classification. The determination of the segmentation scale is easily affected by human subjective judgment and complex surface morphology, making it difficult to construct a feature space that is both comprehensive and highly discriminative, reducing the accuracy of remote sensing image recognition and segmentation.

[0005] In view of the above problems, it is necessary to propose a geographic information acquisition system and method based on remote sensing images. Summary of the invention

[0006] The purpose of the present invention is to solve the problems existing in the background technology and to propose a geographic information acquisition system and method based on remote sensing images.

[0007] The purpose of the present invention can be achieved through the following technical solutions:

[0008] In a first aspect, the present invention provides a geographic information acquisition system based on remote sensing images, including a remote sensing image analysis module, a feature data extraction module, a classification matching module and a geographic data analysis module.

[0009] The remote sensing image analysis module obtains the remote sensing image of the information collection area, and divides the remote sensing image into squares with a preset size of A×A pixels to obtain a number of image units. The image units are numbered, and the number symbol is i, i=1, 2, 3, ..., n. Where n is the total number of image units obtained after segmentation.

[0010] For each image unit, all pixels therein are located, and the positioning coordinates are (w, h), where w is the pixel height, and h is the pixel height; and 0≤w≤A, 0≤h≤A.

[0011] As a preferred method of the present invention, spectral reflectance analysis is performed on all pixels to obtain their reflectance values ​​Q at different wavelengths λ, and a reflectance-wavelength graph is drawn. The areas in different bands are extracted from the reflectance-wavelength graph and recorded as the color values ​​of the pixels (w, h) in different bands, including:

[0012] Purple color value c1(w,h), band range is 390nm-455nm;

[0013] The color value of indigo is c2 (w, h), and the wavelength range is 455nm-492nm;

[0014] Green color value c3 (w, h), band range is 492nm-577nm;

[0015] Yellow color value c4 (w, h), band range is 577nm-597nm;

[0016] Orange color value c5 (w, h), band range 597nm-622nm;

[0017] Red color value c6 (w, h), band range is 622nm-770nm;

[0018] Infrared color value C7 (w, h), the band range is greater than 770nm.

[0019] As a preferred embodiment of the present invention, the grayscale value G(w, h) of each pixel is obtained, where 0≤G(w, h)≤255. For each image unit i, a grayscale companion matrix A×A is generated

[0020] The feature data extraction module performs statistical feature analysis based on color value and gray value data, specifically:

[0021] By formula The overall spectral characteristic parameter Ci of the image unit i is calculated, where Kp is a set of preset weight factors, p=1, 2, 3, 4, 5, 6, 7.

[0022] By formula Calculate the first-order grayscale moment e1i(w) of each image unit i in the wth row and the first-order grayscale moment e1i(h) of the hth column,

[0023] By formula Calculate the gray second moment e2i(w) of each image unit i in the wth row and the gray second moment e2i(h) in the hth column.

[0024] By formula The overall first-order moment E1i and the overall second-order moment E2i of each image unit i are calculated.

[0025] As a preferred method of the present invention, further numerical analysis is performed based on the elements in the grayscale associated matrix to extract texture feature values ​​containing geological information, specifically:

[0026] By formula Calculate the homogeneous eigenvalue E3i of image unit i.

[0027] By formula Calculate the contrast feature value E4i of image unit i.

[0028] By formula Calculate the grayscale entropy E5 i of image unit i.

[0029] By formula Calculate the relevant eigenvalue E6i of image unit i.

[0030] For the same image unit i, the indigo color value C2 (w, h), the green color value C3 (w, h), the red color value C6 (w, h) and the infrared color value C7 (w, h) of each pixel (w, h) are obtained.

[0031] By formula Calculate the vegetation index U1i, water index U2i, building index U3i and soil index U4i of each image unit i.

[0032] The classification matching module integrates and merges the spectral characteristic parameters Ci, the overall first-order moment E1i, the overall second-order moment E2i, the homogeneous eigenvalue E3i, the contrast eigenvalue E4i, the grayscale entropy E5i, the correlation eigenvalue E6i, the vegetation index U1i, the water index U2i, the building index U3i and the soil index U4i of each image unit i. Generate the feature vector (Ci, E1i, E2i, E3i, E4i, E5i, E6i, U1i, U2i, U3i, U4i) of each image unit i.

[0033] Retrieve the preset geographic classification labels, including water system, forest land, cultivated land, garden land, sandy land, water system, residential buildings, roads and industrial and mining land. Each geographic classification label has a label classification vector (C, E1, E2, E3, E4, E5, E6, U1, U2, U3, U4).

[0034] The weighted Euclidean distance d between the feature vector of each image unit i and each label classification vector is calculated. When d is less than the preset threshold dMin, the image unit i is considered to have successfully matched the classification label corresponding to the label classification vector and is marked.

[0035] The calculation formula of weighted Euclidean distance is: Among them, α1 is the preset weight factor; β t1 is a set of preset weight factors; γ t2 A set of preset weight factors.

[0036] The geographic data analysis module obtains the geographic classification label matching results obtained by the classification matching module, counts the geographic information data, and identifies and marks them, specifically:

[0037] The number of image units matched to water system, forest land, cultivated land, garden land, sand land, residential buildings, unused urban open space, roads and industrial and mining land is S1, S2, S3, S4, S5, S6, S7, S8 and S9. Calculate the water system evaluation index ε1 and green space evaluation index ε2.

[0038] When the water system evaluation index ε1 is less than the preset threshold ε1Min, a water system degradation warning signal is generated;

[0039] When the greening evaluation index ε2 is less than the preset threshold ε2Min, a green space degradation warning signal is generated;

[0040] Obtain the slope angle θi and spectral characteristic parameter Ci of each image unit i, and calculate the illumination evaluation index ε3i of each image unit i by the formula ε3i=λ1×(θi-θ0)+λ2×Ci, where λ1 and λ2 are preset weight factors. The image unit i whose illumination evaluation index ε3i is greater than the preset threshold ε3Max is highlighted in yellow. The image unit i whose illumination evaluation index ε3i is less than the preset threshold ε3Min is highlighted in blue.

[0041] Obtain the weighted Euclidean distance d between each image unit i and other adjacent image units, and when d is greater than a preset threshold dMax, mark the image unit i as a classification boundary unit. The process is repeated through the number symbols i of all image units to obtain all classification boundary units, and the classification boundary units are highlighted in red.

[0042] In a second aspect, the present invention provides a system and method for collecting geographic information based on remote sensing images, which specifically includes the following steps:

[0043] Step 1: Remote sensing image analysis;

[0044] A remote sensing image of the information collection area is obtained, and the remote sensing image is segmented into squares of a preset size A×A pixels to obtain a plurality of image units. The image units are numbered, and the number symbol is i, i=1, 2, 3, ..., n. Where n is the total number of image units obtained after segmentation.

[0045] For each image unit, all pixels therein are located, and the positioning coordinates are (w, h), where w is the pixel height, and h is the pixel height; and 0≤w≤A, 0≤h≤A.

[0046] Perform spectral reflectance analysis on all pixels, obtain their reflectance values ​​Q at different wavelengths λ, and draw a reflectance-wavelength graph. Extract the areas in different bands from the reflectance-wavelength graph and record them as the color values ​​of the pixels (w, h) in different bands, including:

[0047] Purple color value c1(w,h), band range is 390nm-455nm;

[0048] The color value of indigo is c2 (w, h), and the wavelength range is 455nm-492nm;

[0049] Green color value c3 (w, h), band range is 492nm-577nm;

[0050] Yellow color value c4 (w, h), band range is 577nm-597nm;

[0051] Orange color value c5 (w, h), band range 597nm-622nm;

[0052] Red color value c6 (w, h), band range is 622nm-770nm;

[0053] Infrared color value C7 (w, h), the band range is greater than 770nm.

[0054] By formula The overall spectral characteristic parameter Ci of the image unit i is calculated, where Kp is a set of preset weight factors, p=1, 2, 3, 4, 5, 6, 7.

[0055] Get the grayscale value G(w,h) of each pixel, where 0≤G(w,h)≤255.

[0056] As a preferred embodiment of the present invention, for each image unit i, a grayscale companion matrix A×A is generated:

[0057] Step 2: Remote sensing image feature extraction;

[0058] By formula Calculate the first-order grayscale moment e1i(w) of each image unit i in the wth row and the first-order grayscale moment e1i(h) of the hth column,

[0059] By formula Calculate the gray second moment e2i(w) of each image unit i in the wth row and the gray second moment e2i(h) in the hth column.

[0060] By formula The overall first-order moment E1i and the overall second-order moment E2i of each image unit i are calculated.

[0061] Further numerical analysis is performed based on the elements in the grayscale associated matrix to extract the texture feature values ​​containing geological information, specifically:

[0062] By formula Calculate the homogeneous eigenvalue E3i of image unit i.

[0063] By formula Calculate the contrast feature value E4i of image unit i.

[0064] By formula Calculate the grayscale entropy E5 i of image unit i.

[0065] By formula Calculate the relevant eigenvalue E6i of image unit i.

[0066] Step 3: Extract geographical features;

[0067] For the same image unit i, the indigo color value C2 (w, h), the green color value C3 (w, h), the red color value C6 (w, h) and the infrared color value C7 (w, h) of each pixel (w, h) are obtained.

[0068] By formula Calculate the vegetation index U1i, water index U2i, building index U3i and soil index U4i of each image unit i.

[0069] Step 4: Image unit matching;

[0070] The spectral characteristic parameters Ci, the overall first-order moment E1i, the overall second-order moment E2i, the homogeneous eigenvalue E3i, the contrast eigenvalue E4i, the gray entropy E5i, the correlation eigenvalue E6i, the vegetation index U1i, the water index U2i, the building index U3i and the soil index U4i of each image unit i are integrated and merged. The characteristic vector (Ci, E1i, E2i, E3i, E4i, E5i, E6i, U1i, U2i, U3i, U4i) of each image unit i is generated.

[0071] Retrieve the preset geographic classification labels, including water system, forest land, cultivated land, garden land, sandy land, water system, residential buildings, roads and industrial and mining land. Each geographic classification label has a label classification vector (C, E1, E2, E3, E4, E5, E6, U1, U2, U3, U4).

[0072] The weighted Euclidean distance d between the feature vector of each image unit i and each label classification vector is calculated. When d is less than the preset threshold dMin, the image unit i is considered to have successfully matched the classification label corresponding to the label classification vector and is marked.

[0073] The calculation formula of weighted Euclidean distance is: Among them, α1 is the preset weight factor; β t1 is a set of preset weight factors; γ t2 A set of preset weight factors.

[0074] Step 5: Geographic data analysis;

[0075] Extract the geographic classification label matching results obtained in step 4 and count the geographic information data, including:

[0076] The number of image units matched to water system, forest land, cultivated land, garden land, sand land, residential buildings, unused urban open space, roads and industrial and mining land is S1, S2, S3, S4, S5, S6, S7, S8 and S9. Calculate the water system evaluation index ε1 and green space evaluation index ε2.

[0077] When the water system evaluation index ε1 is less than the preset threshold ε1Min, a water system degradation warning signal is generated;

[0078] When the greening evaluation index ε2 is less than the preset threshold ε2Min, a green space degradation warning signal is generated;

[0079] Obtain the slope angle θi and spectral characteristic parameter Ci of each image unit i, and calculate the illumination evaluation index ε3i of each image unit i by the formula ε3i=λ1×(θi-θ0)+λ2×Ci, where λ1 and λ2 are preset weight factors. The image unit i whose illumination evaluation index ε3i is greater than the preset threshold ε3Max is highlighted in yellow. The image unit i whose illumination evaluation index ε3i is less than the preset threshold ε3Min is highlighted in blue.

[0080] Obtain the weighted Euclidean distance d between each image unit i and other adjacent image units, and when d is greater than a preset threshold dMax, mark the image unit i as a classification boundary unit. The process is repeated through the number symbols i of all image units to obtain all classification boundary units, and the classification boundary units are highlighted in red.

[0081] Compared with the prior art, the present invention has the following beneficial effects:

[0082] (1) The present invention can more accurately capture the spectral characteristics of surface objects by performing spectral reflectance analysis on all pixels and extracting color values ​​within different bands; generate an A×A grayscale associated matrix, and perform statistical feature analysis such as first-order moment and second-order moment to extract texture eigenvalues ​​containing geological information, such as homogeneous eigenvalues, contrast eigenvalues, grayscale entropy and correlation eigenvalues; calculate parameters such as vegetation index, water index, building index and soil index by combining color values ​​of multiple bands, and can more accurately identify different surface cover types; perform quantitative analysis on the spectral reflectance characteristics, microstructure and roughness of surface objects, which helps to improve classification accuracy;

[0083] (2) The present invention provides clear classification standards for various geographic features by calling preset geographic classification labels; by calculating the weighted Euclidean distance between the feature vector of each image unit and the classification vector of each label, effective matching between the image unit and the geographic classification label is achieved; it not only takes into account the various dimensions of the feature vector, but also emphasizes the importance of certain features through weight factors, thereby improving the accuracy and efficiency of matching;

[0084] (3) The present invention extracts geographic classification label matching results to count various types of geographic information data, such as the number of image units of water systems, woodlands, and cultivated land. This helps to understand the geographic distribution characteristics and provide data support for subsequent geographic planning and resource utilization. By calculating the water system evaluation index and the green space evaluation index, the health status of the water system and green space can be evaluated, providing a scientific basis for environmental protection and ecological restoration. The lighting conditions are evaluated through the lighting evaluation index. This helps to identify areas with insufficient or excessive lighting, providing important references for urban planning, agricultural layout, etc. The classification boundary units are identified by calculating the weighted Euclidean distance between each image unit and other intersecting image units. This helps to distinguish the boundary areas of different geographic types and provides a reference for geographic classification and map making. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings:

[0086] Figure 1 is a system block diagram of the present invention;

[0087] Figure 2 is a reflectivity-wavelength graph of the present invention;

[0088] Figure 3 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0089] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. 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.

[0090] See also Figure 1 As shown, the geographic information acquisition system based on remote sensing images includes a remote sensing image analysis module, a feature data extraction module, a classification matching module and a geographic data analysis module.

[0091] The remote sensing image analysis module obtains the remote sensing image of the information collection area, and divides the remote sensing image into squares with a preset size of A×A pixels to obtain a number of image units. The image units are numbered, and the number symbol is i, i=1, 2, 3, ..., n. Where n is the total number of image units obtained after segmentation.

[0092] For each image unit, all pixels therein are located, and the positioning coordinates are (w, h), where w is the pixel height, and h is the pixel height; and 0≤w≤A, 0≤h≤A.

[0093] See also Figure 2 As shown, spectral reflectance analysis is performed on all pixels to obtain their reflectance values ​​Q at different wavelengths λ, and a reflectance-wavelength graph is drawn. In the reflectance-wavelength graph, the areas in different bands are extracted and recorded as the color values ​​of the pixels (w, h) in different bands, including:

[0094] Purple color value c1(w,h), band range is 390nm-455nm;

[0095] The color value of indigo is c2 (w, h), and the wavelength range is 455nm-492nm;

[0096] Green color value c3 (w, h), band range is 492nm-577nm;

[0097] Yellow color value c4 (w, h), band range is 577nm-597nm;

[0098] Orange color value c5 (w, h), band range 597nm-622nm;

[0099] Red color value c6 (w, h), band range is 622nm-770nm;

[0100] Infrared color value C7 (w, h), the band range is greater than 770nm.

[0101] Furthermore, the grayscale value G(w, h) of each pixel is obtained, where 0≤G(w, h)≤255. For each image unit i, an A×A grayscale companion matrix is ​​generated:

[0102]

[0103] The feature data extraction module performs further statistical feature analysis based on the color value and gray value data, specifically:

[0104] By formula The overall spectral characteristic parameter Ci of the image unit i is calculated, where Kp is a set of preset weight factors, p=1, 2, 3, 4, 5, 6, 7.

[0105] By formula Calculate the first-order grayscale moment e1i(w) of each image unit i in the wth row and the first-order grayscale moment e1i(h) of the hth column,

[0106] By formula Calculate the gray second moment e2i(w) of each image unit i in the wth row and the gray second moment e2i(h) in the hth column.

[0107] It should be noted that the grayscale first-order moment is a measure of the average grayscale value, and the grayscale second-order moment is a measure of the uniformity of the grayscale value. The larger the grayscale first-order moment, the higher the average grayscale value of the row or column. The larger the grayscale second-order moment, the more uneven the grayscale value distribution of the row or column.

[0108] By formula The overall first-order moment E1i and the overall second-order moment E2i of each image unit i are calculated.

[0109] Furthermore, further numerical analysis is performed based on the elements in the grayscale associated matrix to extract the texture feature values ​​containing geological information, specifically:

[0110] By formula Calculate the homogeneous eigenvalue E3i of image unit i.

[0111] It should be noted that the homogeneity eigenvalue is a measure of texture uniformity. The larger the homogeneity eigenvalue, the more uniform the texture distribution of the image unit i.

[0112] By formula Calculate the contrast feature value E4i of image unit i.

[0113] It should be noted that the contrast eigenvalue is a measure of the total amount of texture change. The larger the contrast eigenvalue, the more obvious the texture difference of image unit i is.

[0114] By formula Calculate the grayscale entropy E5i of image unit i.

[0115] It should be noted that the spectral value entropy is a measure of texture disorder and complexity. The larger the spectral value entropy, the more disordered the texture distribution of image unit i is.

[0116] By formula Calculate the relevant eigenvalue E6i of image unit i.

[0117] It should be noted that the correlation eigenvalue is a measure of the degree of correlation between spectral values ​​in different rows and columns. The larger the correlation eigenvalue is, the greater the degree of correlation between the spectral values ​​of image unit i in the row and column dimensions.

[0118] For the same image unit i, the indigo color value C2 (w, h), the green color value C3 (w, h), the red color value C6 (w, h) and the infrared color value C7 (w, h) of each pixel (w, h) are obtained.

[0119] By formula Calculate the vegetation index U1i, water index U2i, building index U3i and soil index U4i of each image unit i.

[0120] The classification matching module integrates and merges the spectral characteristic parameters Ci, the overall first-order moment E1i, the overall second-order moment E2i, the homogeneous eigenvalue E3i, the contrast eigenvalue E4i, the grayscale entropy E5i, the correlation eigenvalue E6i, the vegetation index U1i, the water index U2i, the building index U3i and the soil index U4i of each image unit i. Generate the feature vector (Ci, E1i, E2i, E3i, E4i, E5i, E6i, U1i, U2i, U3i, U4i) of each image unit i.

[0121] Retrieve the preset geographic classification labels, including water system, forest land, cultivated land, garden land, sandy land, water system, residential buildings, roads and industrial and mining land. Each geographic classification label has a label classification vector (C, E1, E2, E3, E4, E5, E6, U1, U2, U3, U4).

[0122] The weighted Euclidean distance d between the feature vector of each image unit i and each label classification vector is calculated. When d is less than the preset threshold dMin, the image unit i is considered to have successfully matched the classification label corresponding to the label classification vector and is marked.

[0123] The calculation formula of weighted Euclidean distance is: Among them, α1 is the preset weight factor; β t1 is a set of preset weight factors; γ t2 A set of preset weight factors.

[0124] The geographic data analysis module obtains the geographic classification label matching results obtained by the classification matching module, counts the geographic information data, and identifies and marks them, specifically:

[0125] The number of image units matched to water system, forest land, cultivated land, garden land, sand land, residential buildings, unused urban open space, roads and industrial and mining land is S1, S2, S3, S4, S5, S6, S7, S8 and S9. Calculate the water system evaluation index ε1 and green space evaluation index ε2.

[0126] When the water system evaluation index ε1 is less than the preset threshold ε1Min, a water system degradation warning signal is generated;

[0127] When the greening evaluation index ε2 is less than the preset threshold ε2Min, a green space degradation warning signal is generated;

[0128] Obtain the slope angle θi and spectral characteristic parameter Ci of each image unit i, and calculate the illumination evaluation index ε3i of each image unit i by the formula ε3i=λ1×(θi-θ0)+λ2×Ci, where λ1 and λ2 are preset weight factors. The image unit i whose illumination evaluation index ε3i is greater than the preset threshold ε3Max is highlighted in yellow. The image unit i whose illumination evaluation index ε3i is less than the preset threshold ε3Min is highlighted in blue.

[0129] Obtain the weighted Euclidean distance d between each image unit i and other adjacent image units, and when d is greater than a preset threshold dMax, mark the image unit i as a classification boundary unit. The process is repeated through the number symbols i of all image units to obtain all classification boundary units, and the classification boundary units are highlighted in red.

[0130] See also Figure 3 As shown, the geographic information acquisition system and method based on remote sensing images specifically include the following steps:

[0131] Step 1: Remote sensing image analysis;

[0132] A remote sensing image of the information collection area is obtained, and the remote sensing image is segmented into squares of a preset size A×A pixels to obtain a plurality of image units. The image units are numbered, and the number symbol is i, i=1, 2, 3, ..., n. Where n is the total number of image units obtained after segmentation.

[0133] For each image unit, all pixels therein are located, and the positioning coordinates are (w, h), where w is the pixel height, and h is the pixel height; and 0≤w≤A, 0≤h≤A.

[0134] Perform spectral reflectance analysis on all pixels, obtain their reflectance values ​​Q at different wavelengths λ, and draw a reflectance-wavelength graph. Extract the areas in different bands from the reflectance-wavelength graph and record them as the color values ​​of the pixels (w, h) in different bands, including:

[0135] Purple color value c1(w,h), band range is 390nm-455nm;

[0136] The color value of indigo is c2 (w, h), and the wavelength range is 455nm-492nm;

[0137] Green color value c3 (w, h), band range is 492nm-577nm;

[0138] Yellow color value c4 (w, h), band range is 577nm-597nm;

[0139] Orange color value c5 (w, h), band range 597nm-622nm;

[0140] Red color value c6 (w, h), band range is 622nm-770nm;

[0141] Infrared color value C7 (w, h), the band range is greater than 770nm.

[0142] By formula The overall spectral characteristic parameter Ci of the image unit i is calculated, where Kp is a set of preset weight factors, p=1, 2, 3, 4, 5, 6, 7.

[0143] Get the grayscale value G(w,h) of each pixel, where 0≤G(w,h)≤255.

[0144] Furthermore, for each image unit i, an A×A grayscale companion matrix is ​​generated:

[0145] Step 2: Remote sensing image feature extraction;

[0146] By formula Calculate the first-order grayscale moment e1i(w) of each image unit i in the wth row and the first-order grayscale moment e1i(h) of the hth column,

[0147] By formula Calculate the gray second moment e2i(w) of each image unit i in the wth row and the gray second moment e2i(h) in the hth column.

[0148] By formula The overall first-order moment E1i and the overall second-order moment E2i of each image unit i are calculated.

[0149] Further numerical analysis is performed based on the elements in the grayscale associated matrix to extract the texture feature values ​​containing geological information, specifically:

[0150] By formula Calculate the homogeneous eigenvalue E3i of image unit i.

[0151] By formula Calculate the contrast feature value E4i of image unit i.

[0152] By formula Calculate the grayscale entropy E5 i of image unit i.

[0153] By formula Calculate the relevant eigenvalue E6i of image unit i.

[0154] Step 3: Extract geographical features;

[0155] For the same image unit i, the indigo color value C2 (w, h), the green color value C3 (w, h), the red color value C6 (w, h) and the infrared color value C7 (w, h) of each pixel (w, h) are obtained.

[0156] By formula Calculate the vegetation index U1i, water index U2i, building index U3i and soil index U4i of each image unit i.

[0157] Step 4: Image unit matching;

[0158] The spectral characteristic parameters Ci, the overall first-order moment E1i, the overall second-order moment E2i, the homogeneous eigenvalue E3i, the contrast eigenvalue E4i, the gray entropy E5i, the correlation eigenvalue E6i, the vegetation index U1i, the water index U2i, the building index U3i and the soil index U4i of each image unit i are integrated and merged. The characteristic vector (Ci, E1i, E2i, E3i, E4i, E5i, E6i, U1i, U2i, U3i, U4i) of each image unit i is generated.

[0159] Retrieve the preset geographic classification labels, including water system, forest land, cultivated land, garden land, sandy land, water system, residential buildings, roads and industrial and mining land. Each geographic classification label has a label classification vector (C, E1, E2, E3, E4, E5, E6, U1, U2, U3, U4).

[0160] The weighted Euclidean distance d between the feature vector of each image unit i and each label classification vector is calculated. When d is less than the preset threshold dMin, the image unit i is considered to have successfully matched the classification label corresponding to the label classification vector and is marked.

[0161] The calculation formula of weighted Euclidean distance is: Among them, α1 is the preset weight factor; β t1 is a set of preset weight factors; γ t2 A set of preset weight factors.

[0162] Step 5: Geographic data analysis;

[0163] Extract the geographic classification label matching results obtained in step 4 and count the geographic information data, including:

[0164] The number of image units matched to water system, forest land, cultivated land, garden land, sand land, residential buildings, unused urban open space, roads and industrial and mining land is S1, S2, S3, S4, S5, S6, S7, S8 and S9. Calculate the water system evaluation index ε1 and green space evaluation index ε2.

[0165] When the water system evaluation index ε1 is less than the preset threshold ε1Min, a water system degradation warning signal is generated;

[0166] When the greening evaluation index ε2 is less than the preset threshold ε2Min, a green space degradation warning signal is generated;

[0167] Obtain the slope angle θi and spectral characteristic parameter Ci of each image unit i, and calculate the illumination evaluation index ε3i of each image unit i by the formula ε3i=λ1×(θi-θ0)+λ2×Ci, where λ1 and λ2 are preset weight factors. The image unit i whose illumination evaluation index ε3i is greater than the preset threshold ε3Max is highlighted in yellow. The image unit i whose illumination evaluation index ε3i is less than the preset threshold ε3Min is highlighted in blue.

[0168] Obtain the weighted Euclidean distance d between each image unit i and other adjacent image units, and when d is greater than a preset threshold dMax, mark the image unit i as a classification boundary unit. The process is repeated through the number symbols i of all image units to obtain all classification boundary units, and the classification boundary units are highlighted in red.

[0169] It should be understood that the terms “include” and “comprising” used in the specification and claims of the present disclosure indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0170] It should also be understood that the terms used in this disclosure are only for the purpose of describing specific embodiments and are not intended to limit the disclosure. As used in this disclosure and the claims, the singular forms "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should also be further understood that the term "and / or" used in this disclosure and the claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations;

[0171] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A geographic information acquisition system based on remote sensing images, including a feature data extraction module, characterized in that: The feature data extraction module performs statistical feature analysis based on color value and gray value data. Specifically, the formula Calculate the overall spectral characteristic parameter Ci of image unit i, where Kp is a set of preset weight factors, p = 1, 2, 3, 4, 5, 6, 7; by formula Calculate the first-order moment of grayscale e1i(w) in the wth row and e1i(h) in the hth column of each image unit i, using the formula Calculate the gray second moment e2i(w) of each image unit i in the wth row and the gray second moment e2i(h) of the hth column; by the formula Calculate the overall first-order moment E1i and the overall second-order moment E2i of each image unit i; Further numerical analysis is performed based on the elements in the grayscale associated matrix to calculate the texture feature values ​​containing geological information; the spectral analysis results are retrieved to calculate the vegetation index, water index, building index and soil index of each image unit.

2. The geographic information acquisition system based on remote sensing images according to claim 1 is characterized in that: It also includes remote sensing image analysis module, classification matching module and geographic data analysis module; The remote sensing image analysis module obtains the remote sensing image of the information collection area, and divides the remote sensing image into squares with a preset size of A×A pixels to obtain a number of image units; the image units are numbered, and the number symbol is i, i=1, 2, 3, ..., n; wherein n is the total number of image units obtained after segmentation; For each image unit, all pixels in it are located, and the location coordinates are (w, h), where w is the pixel height, h is the pixel height; where 0≤w≤A, 0≤h≤A; Perform spectral reflectance analysis on all pixels, obtain their reflectance values ​​Q at different wavelengths λ, and draw a reflectance-wavelength graph; extract the areas in different characteristic bands from the reflectance-wavelength graph, and record them as the color values ​​of pixels (w, h) in different characteristic bands, including: Get the gray value G(w, h) of each pixel, where 0≤G(w, h)≤255; for each image unit i, generate an A×A grayscale companion matrix The classification matching module integrates and merges the spectral characteristic parameters Ci, the overall first-order moment E1i, the overall second-order moment E2i, the homogeneous eigenvalue E3i, the contrast eigenvalue E4i, the grayscale entropy E5i, the correlation eigenvalue E6i, the vegetation index U1i, the water index U2i, the building index U3i and the soil index U4i of each image unit i; generates the characteristic vector (Ci, E1i, E2i, E3i, E4i, E5i, E6i, U1i, U2i, U3i, U5i) of each image unit i; calls the preset geographic classification label, each geographic classification label has a label classification vector (C, E1, E2, E3, E4, E5, E6, U1, U2, U3, U5); Calculate the weighted Euclidean distance d between the feature vector of each image unit i and each label classification vector. When d is less than the preset threshold dMin, it is considered that the image unit i and the classification label corresponding to the label classification vector match successfully and are marked; The geographic data analysis module obtains the geographic classification label matching results obtained by the classification matching module, counts the geographic information data, and performs identification and marking; The number of image units matched to each geographic classification label is obtained and the water system evaluation index ε1 and the green space evaluation index ε2 are obtained through numerical calculation; when the water system evaluation index ε1 is less than the preset threshold ε1Min, a water system degradation warning signal is generated; when the green space evaluation index ε2 is less than the preset threshold ε2Min, a green space degradation warning signal is generated; Obtain the slope angle θi and spectral characteristic parameter Ci of each image unit i, and calculate the illumination evaluation index ε3i of each image unit i by the formula ε3i=λ1×(θi-θ0)+λ2×Ci, where λ1 and λ2 are preset weight factors; highlight the image unit i whose illumination evaluation index ε3i is greater than the preset threshold ε3Max in yellow; highlight the image unit i whose illumination evaluation index ε3i is less than the preset threshold ε3Min in blue; Obtain the weighted Euclidean distance d between each image unit i and other intersecting image units. When d is greater than a preset threshold dMax, mark the image unit i as a classification boundary unit. Perform the process on the number symbols i of all image units to obtain all classification boundary units, and highlight the classification boundary units in red.

3. The geographic information acquisition system based on remote sensing images according to claim 1 is characterized in that: The specific process of numerically analyzing the elements in the grayscale associated matrix and calculating the texture feature values ​​containing geological information is as follows: By formula Calculate the homogeneous eigenvalue E3i of image unit i; By formula Calculate the contrast eigenvalue E4i of image unit i; By formula Calculate the grayscale entropy E5i of image unit i; By formula Calculate the relevant eigenvalue E6i of image unit i.

4. The geographic information acquisition system based on remote sensing images according to claim 1 is characterized in that: The specific process of retrieving the spectral analysis results and calculating the vegetation index, water index, building index and soil index of each image unit is as follows: For the same image unit i, obtain the indigo color value C2(w, h), the green color value C3(w, h), the red color value C6(w, h) and the infrared color value C7(w, h) of each pixel (w, h); By formula Calculate the vegetation index U1i, water index U2i, building index U3i and soil index U4i of each image unit i.

5. The geographic information acquisition system based on remote sensing images according to claim 2 is characterized in that: The characteristic bands include: Purple color value c1(w,h), band range is 390nm-455nm; The color value of indigo is c2 (w, h), and the wavelength range is 455nm-492nm; Green color value c3 (w, h), band range is 492nm-577nm; Yellow color value c4 (w, h), band range is 577nm-597nm; Orange color value c5 (w, h), band range 597nm-622nm; Red color value c6 (w, h), band range is 622nm-770nm; Infrared color value C7 (w, h), the band range is greater than 770nm.

6. The geographic information acquisition system based on remote sensing images according to claim 2 is characterized in that: The geographical classification labels include: water system, forest land, cultivated land, garden land, sandy land, water system, residential buildings, roads and industrial and mining land.

7. The geographic information acquisition system based on remote sensing images according to claim 2 is characterized in that: The calculation formula of the weighted Euclidean distance is specifically: Where α1 is the preset weight factor; β t1 is a set of preset weight factors; γ t2 A set of preset weight factors.

8. The geographic information acquisition system based on remote sensing images according to claim 2 is characterized in that: The specific process of obtaining the water system evaluation index and green space evaluation index through numerical calculation is as follows: Get the preset geographic classification labels, i.e., the number of image units S1, S2, S3, S4, S5, S6, S7, S8 and S9 matched to water system, woodland, cultivated land, garden, sandy land, residential buildings, unused urban open space, roads and industrial and mining land; through the formula Calculate the water system evaluation index ε1 and green space evaluation index ε2.

9. A method for collecting geographic information based on remote sensing images, characterized in that The geographical information acquisition system based on remote sensing images applied to any one of claims 1 to 8 comprises the following steps: Step 1: Remote sensing image analysis; Acquire a remote sensing image of the information collection area, segment the remote sensing image into squares of a preset size of A×A pixels to obtain a plurality of image units; number the image units, where the number symbol is i, i=1, 2, 3, ..., n; wherein n is the total number of image units obtained after segmentation; For each image unit, all pixels in it are located, and the positioning coordinates are (w, h), where w is the pixel height and h is the pixel height; Perform spectral reflectance analysis on all pixels, obtain their reflectance values ​​Q at different wavelengths λ, and draw a reflectance-wavelength graph; extract the areas in different bands from the reflectance-wavelength graph and record them as the color values ​​of the pixels (w, h) in different bands; By formula Calculate the overall spectral characteristic parameter Ci of the image unit i, where Kp is a set of preset weight factors, p = 1, 2, 3, 4, 5, 6, 7; Get the grayscale value G(w,h) of each pixel, where 0≤G(w,h)≤255; For each image unit i, generate an A×A grayscale companion matrix Step 2: Remote sensing image feature extraction; By formula Calculate the first-order grayscale moment e1i(w) of each image unit i in the wth row and the first-order grayscale moment e1i(h) of the hth column, By formula Calculate the gray second moment e2i(w) of each image unit i in the wth row and the gray second moment e2i(h) of the hth column; By formula Calculate the overall first-order moment E1i and the overall second-order moment E2i of each image unit i; Further numerical analysis is performed based on the elements in the grayscale associated matrix to extract the texture feature values ​​containing geological information, specifically: By formula Calculate the homogeneous eigenvalue E3i of image unit i; By formula Calculate the contrast eigenvalue E4i of image unit i; By formula Calculate the grayscale entropy E5i of image unit i; By formula Calculate the relevant eigenvalue E6i of image unit i; Step 3: Extract geographical features; For the same image unit i, obtain the indigo color value C2(w, h), the green color value C3(w, h), the red color value C6(w, h) and the infrared color value C7(w, h) of each pixel (w, h); By formula Calculate the vegetation index U1i, water index U2i, building index U3i and soil index U4i of each image unit i; Step 4: Image unit matching; Integrate and merge the spectral characteristic parameters Ci, the overall first-order moment E1i, the overall second-order moment E2i, the homogeneous eigenvalue E3i, the contrast eigenvalue E4i, the grayscale entropy E5i, the correlation eigenvalue E6i, the vegetation index U1i, the water index U2i, the building index U3i and the soil index U4i of each image unit i; generate the characteristic vector (Ci, E1i, E2i, E3i, E4i, E5i, E6i, U1i, U2i, U3i, U4i) of each image unit i; Retrieve preset geographic classification labels, including water system, forest land, cultivated land, garden land, sandy land, water system, residential buildings, roads and industrial and mining land; each geographic classification label has a label classification vector (C, E1, E2, E3, E4, E5, E6, U1, U2, U3, U4); Calculate the weighted Euclidean distance d between the feature vector of each image unit i and each label classification vector. When d is less than the preset threshold dMin, it is considered that the image unit i and the classification label corresponding to the label classification vector match successfully and are marked; The calculation formula of weighted Euclidean distance is: Among them, α1 is the preset weight factor; β t1 is a set of preset weight factors; γ t2 is a set of preset weight factors; Step 5: Geographic data analysis; Extract the geographic classification label matching results obtained in step 4 and count the geographic information data, including: The number of image units matched to water system, forest land, cultivated land, garden land, sand land, residential buildings, unused urban open space, roads and industrial and mining land is S1, S2, S3, S4, S5, S6, S7, S8 and S9; through the formula Calculate the water system evaluation index ε1 and the green space evaluation index ε2; When the water system evaluation index ε1 is less than the preset threshold ε1Min, a water system degradation warning signal is generated; When the greening evaluation index ε2 is less than the preset threshold ε2Min, a green space degradation warning signal is generated; Obtain the slope angle θi and spectral characteristic parameter Ci of each image unit i, and calculate the illumination evaluation index ε3i of each image unit i by the formula ε3i=λ1×(θi-θ0)+λ2×Ci, where λ1 and λ2 are preset weight factors; highlight the image unit i whose illumination evaluation index ε3i is greater than the preset threshold ε3Max in yellow; highlight the image unit i whose illumination evaluation index ε3i is less than the preset threshold ε3Min in blue; Obtain the weighted Euclidean distance d between each image unit i and other intersecting image units. When d is greater than a preset threshold dMax, mark the image unit i as a classification boundary unit. Perform the process on the number symbols i of all image units to obtain all classification boundary units, and highlight the classification boundary units in red.

Images