Land use dynamic monitoring and optimizing method based on AI

Abstract

The invention belongs to the technical field of land monitoring, and particularly relates to an AI-based land use dynamic monitoring and optimization method. The technical scheme adopted by the invention is as follows: the invention provides an AI-based land use dynamic monitoring and optimization method. The method comprises the following steps: data acquisition; processing data; classifying the images; performing change analysis; the invention provides a land use dynamic monitoring and optimizing system based on AI, and the system comprises an image recognition module, an image marking module, a data analysis module, a historical data module, a vector module, a prediction module, and a planning module. The urban land utilization change is analyzed and researched based on the satellite remote sensing technology, the effects of urban planning and environment monitoring can be improved, and very valuable results can be generated by obtaining high-quality and high-resolution remote sensing data and analyzing the urban land utilization change.

Description

Technical Field

[0001] The present invention belongs to the field of land monitoring technology, and specifically refers to an AI-based land use dynamic monitoring and optimization method. Background Art

[0002] As the global population continues to increase, land resources are becoming more scarce and the human living environment is seriously threatened. Understanding the current status and changing patterns of land use in order to make correct decisions is the goal currently pursued by people.

[0003] Large-scale human activities and natural resource development have caused rapid changes in the ecological environment around the world. As one of the main causes of global change, land use change has outstanding practical significance for dynamic remote sensing monitoring. However, on the one hand, the current dynamic remote sensing monitoring of land use mainly uses human-computer interactive interpretation to detect changes, or uses spectral classification to extract change information. However, the visual interpretation method has high requirements for the interpretation personnel, and the workload of interpretation is very huge. The classification method also has the problem of low accuracy. These deficiencies seriously restrict the promotion of dynamic remote sensing monitoring technology of land use. On the other hand, with the improvement of satellite resolution, spatial information in remote sensing images is very prominent. Therefore, how to use the rich texture information of remote sensing images to improve the efficiency and accuracy of dynamic remote sensing monitoring of land use and reduce the requirements for operators is an important research direction at present. Summary of the invention

[0004] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides an AI-based land use dynamic monitoring and optimization method.

[0005] The technical solution adopted by the present invention is as follows: On the one hand, the present invention provides an AI-based land use dynamic monitoring and optimization method, comprising the following steps:

[0006] Data collection, based on satellite remote sensing images, using image processing to collect and monitor land use and generate data sets;

[0007] Data processing, including correction, mosaicking, fusion, and radiation correction of remote sensing effects, to form remote sensing orthophotos with high spatial resolution and high spectral resolution;

[0008] Image classification, which divides pixels into different categories, analyzes urban land use changes, including single-phase image classification and multi-phase image classification. Single-phase image classification divides different pixels in the same remote sensing image into different categories to obtain land use types, and multi-phase image classification compares remote sensing images of different phases to obtain land use change information;

[0009] Change analysis: compare the single-phase image classification data with the multi-phase image classification data to obtain urban land use change information, and update the land use status map by overlaying the remote sensing image with the land use status map;

[0010] The change analysis can be performed by methods such as cross-tabulation, comparison chart, change region extraction and change vector analysis.

[0011] Furthermore, the data processing includes image correction, including control point selection, correction model selection, and geometric correction accuracy inspection.

[0012] Furthermore, in image correction, the control points are evenly distributed, and the number of control points is 15; the relationship between the number of control points and the polynomial order (n) is that the number of control points is twice (n+1)(n+2) / 2.

[0013] Furthermore, the radiation correction is to obtain the radiation calibration coefficient of the sensor, and the calibration coefficient is used to complete the calculation of the radiation brightness:

[0014] L=DN*Gain+Bias,

[0015] Among them, L represents the radiation brightness, DN represents the grayscale of the pixel, and Gain and Bias are calibration coefficients.

[0016] Furthermore, the geometric correction accuracy check includes a rough check and a precise check. The rough check superimposes the detected land status image and the corrected remote sensing image, and uses a rolling shutter method to check the accuracy of the remote sensing image; the precise check includes taking the land status map as the basis, and then randomly extracting check points other than the correction control points for correction.

[0017] Furthermore, the image fusion method includes HIS transformation method, principal component transformation method, addition transformation method, and weighted multiplication transformation method;

[0018] Furthermore, after the image fusion, the data needs to be post-processed to check for ghosting and other phenomena, check the image texture details and colors, and judge the correctness of the pre-fusion processing; if the image brightness is low or the grayscale is narrow after fusion, linear stretching, brightness contrast, color balance, chroma, saturation and brightness adjustment methods are used to adjust the tone.

[0019] Furthermore, in image classification, two-phase remote sensing orthophoto image data are used to highlight the changed parts using computer automatic processing methods such as image difference and spectral feature variation, and the changed spots are extracted and numbered one by one based on AI; after obtaining the changed spots, GIS is used to establish the topological relationship of the land change spots and produce a land change information table.

[0020] On the other hand, this solution provides an AI-based land use dynamic monitoring and optimization system, including:

[0021] Image recognition module, which identifies different types of land use based on images acquired by remote sensing satellites;

[0022] The image marking module marks the type of land in the image based on the acquired remote sensing image and generates marking data results;

[0023] The data analysis module extracts land use indicators based on the labeling results, performs time series analysis, and generates land use assessment results;

[0024] The historical data module obtains historical land use information and adds spatiotemporal dimension tags as training reference objects;

[0025] The vector module processes the historical land use information of historical resource data into vector form to form a vector model;

[0026] The prediction module detects the trend of land use change based on the assessment results, combines historical data to predict future land use and development trends, makes risk assessments, and generates prediction results;

[0027] The planning module combines the prediction results with the requirements of land use planning to optimize land use, simulates the impact of different planning schemes based on the optimization results, and conducts integrated analysis to generate land planning adjustment suggestions.

[0028] The beneficial effects achieved by the present invention using the above scheme are as follows: This scheme provides an AI-based land use dynamic monitoring and optimization method, which analyzes and studies urban land use changes based on satellite remote sensing technology, can improve the effect of urban planning and environmental monitoring, and obtains urban land use change information through data processing, image classification and change analysis methods. The analysis of urban land use changes can produce very valuable results. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is the monitoring system architecture diagram of this solution;

[0030] Figure 2 The following is a flow chart of the monitoring method for this program. DETAILED DESCRIPTION

[0031] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0032] The technical solution adopted by the present invention is as follows: In one aspect, the present invention provides a method for dynamic monitoring and optimization of land use based on AI, comprising the following steps:

[0033] Data collection, based on satellite remote sensing images, using image processing to collect and monitor land use and generate data sets;

[0034] In the embodiment, the satellites used include spot, landsat, and tm, and the resolution and band are rotated according to the requirements to collect the corresponding image data;

[0035] The contents of the collection and monitoring are urban land, wasteland, construction land, water area, wetland, cultivated land, land for animal husbandry production, forestry land, pond water surface, and highway construction land;

[0036] Data processing, including correction, mosaicking, fusion, and radiation correction of remote sensing effects, to form remote sensing orthophotos with high spatial resolution and high spectral resolution;

[0037] Image classification, which divides pixels into different categories, analyzes urban land use changes, including single-phase image classification and multi-phase image classification. Single-phase image classification divides different pixels in the same remote sensing image into different categories to obtain land use types, and multi-phase image classification compares remote sensing images of different phases to obtain land use change information;

[0038] Change analysis: compare the single-phase image classification data with the multi-phase image classification data to obtain urban land use change information, and update the land use status map by overlaying the remote sensing image with the land use status map;

[0039] The change analysis can be performed by methods such as cross-tabulation, comparison chart, change region extraction and change vector analysis.

[0040] Furthermore, the data processing includes image correction, including control point selection, correction model selection, and geometric correction accuracy inspection.

[0041] In the embodiment, correction is performed using SPOT5 whole-number color images as units, the image resampling interval is 2.5m, the resampling method is bilinear interpolation, and the image correction control points are selected mainly with reference to the land use status map. In some areas, since there is no land use status map, a topographic map is used instead. At the same time, the mountainous area also refers to DEM for point selection correction.

[0042] Furthermore, in image correction, the control points are evenly distributed, and the number of control points is 15; the relationship between the number of control points and the polynomial order (n) is that the number of control points is twice (n+1)(n+2) / 2.

[0043] In the embodiment, a finite element model is used for partition correction to control errors, model points are selected evenly, control points with large residuals are checked and processed, and the control points are maintained to facilitate retrieval of the control points when correcting again.

[0044] Furthermore, the radiation correction is to obtain the radiation calibration coefficient of the sensor, and the calibration coefficient is used to complete the calculation of the radiation brightness:

[0045] L=DN*Gain+Bias,

[0046] Among them, L represents the radiation brightness, DN represents the grayscale of the pixel, and Gain and Bias are calibration coefficients.

[0047] Furthermore, the geometric correction accuracy check includes a rough check and a precise check. The rough check superimposes the detected land status image and the corrected remote sensing image, and uses a rolling shutter method to check the accuracy of the remote sensing image; the precise check includes taking the land status map as the basis, and then extracting check points other than the correction control points for correction; in an embodiment, the correction control points are 30, the correction error is within 5 meters for plains, and within 8 meters for hills.

[0048] Furthermore, the image fusion method includes HIS transformation method, principal component transformation method, addition transformation method, and weighted multiplication transformation method;

[0049] In the embodiment, tone adjustment is performed before image fusion to improve the brightness of high-resolution data, enhance local contrast, highlight texture details, and reduce noise; color enhancement is performed on multispectral data to increase the color contrast between different land types and highlight their spectral color information.

[0050] After fusion, the data needs to be post-processed to check for ghosting and other phenomena, check the image texture details and colors, and judge the correctness of the pre-fusion processing. If the image brightness is low or the grayscale is narrow after fusion, the color tone can be adjusted by linear stretching, brightness contrast, color balance, chroma, saturation and brightness adjustment.

[0051] Furthermore, in image classification, two-phase remote sensing orthophoto image data are used to highlight the changed parts using computer automatic processing methods such as image difference and spectral feature variation, and the changed spots are extracted and numbered one by one based on AI; after obtaining the changed spots, GIS is used to establish the topological relationship of the land change spots and produce a land change information table.

[0052] In an embodiment, different marks and numbers are used for positive change patterns and uncertain change patterns.

[0053] On the other hand, this solution provides an AI-based land use dynamic monitoring and optimization system, including:

[0054] Image recognition module, which identifies different types of land use based on images acquired by remote sensing satellites;

[0055] The image marking module marks the type of land in the image based on the acquired remote sensing image and generates marking data results;

[0056] The data analysis module extracts land use indicators based on the labeling results, performs time series analysis, and generates land use assessment results;

[0057] The historical data module obtains historical land use information and adds spatiotemporal dimension tags as training reference objects;

[0058] The vector module processes the historical land use information of historical resource data into vector form to form a vector model;

[0059] The AI ​​prediction module detects the trend of land use change based on the assessment results, combines historical data to predict future land use and development trends, makes risk assessments, and generates prediction results;

[0060] The planning module combines the prediction results with the requirements of land use planning to optimize land use, simulates the impact of different planning schemes based on the optimization results, and conducts integrated analysis to generate land planning adjustment suggestions.

[0061] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0062] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

[0063] The present invention and its implementation methods are described above, and such description is not restrictive. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design structures and embodiments similar to the technical solution without creativity, which should all fall within the protection scope of the present invention.

Claims

1. AI-based land use dynamic monitoring and optimization method, characterized in that: The steps include: Data collection, based on satellite remote sensing images, using image processing to collect and monitor land use and generate data sets; Data processing, including correction, mosaicking, fusion, and radiation correction of remote sensing effects, to form remote sensing orthophotos with high spatial resolution and high spectral resolution; Image classification, which divides pixels into different categories, analyzes urban land use changes, including single-phase image classification and multi-phase image classification. Single-phase image classification divides different pixels in the same remote sensing image into different categories to obtain land use types, and multi-phase image classification compares remote sensing images of different phases to obtain land use change information; Change analysis: compare the single-phase image classification data with the multi-phase image classification data to obtain urban land use change information, and update the land use status map by overlaying the remote sensing image with the land use status map; The change analysis can be performed by methods such as cross-tabulation, comparison chart, change region extraction and change vector analysis.

2. The AI-based dynamic land use monitoring and optimization method according to claim 1, characterized in that: The data processing includes image correction, including control point selection, correction model selection, and geometric correction accuracy inspection.

3. The AI-based dynamic land use monitoring and optimization method according to claim 1, characterized in that: In image correction, the control points are evenly distributed, and the number of control points is 15. The relationship between the number of control points and the polynomial order (n) is that the number of control points is twice (n+1)(n+2) / 2.

4. The method for dynamic monitoring and optimization of land use based on AI according to claim 1, characterized in that: Radiation correction is to obtain the radiation calibration coefficient of the sensor, and use the calibration coefficient to complete the calculation of radiation brightness: L=DN*Gain+Bias, Among them, L represents the radiation brightness, DN represents the grayscale of the pixel, and Gain and Bias are calibration coefficients.

5. The method for dynamic monitoring and optimization of land use based on AI according to claim 1, characterized in that: The geometric correction accuracy check includes a rough check and a precise check. The rough check superimposes the detected land status image and the corrected remote sensing image, and uses a rolling shutter method to check the accuracy of the remote sensing image; the precise check includes taking the land status map as the basis, and then extracting check points other than the correction control points for correction.

6. The AI-based dynamic land use monitoring and optimization method according to claim 1, characterized in that: The image fusion method includes HIS transformation method, principal component transformation method, addition transformation method and weighted multiplication transformation method.

7. The AI-based dynamic land use monitoring and optimization method according to claim 1, characterized in that: After the image fusion, the data needs to be post-processed to check for ghosting and other phenomena, check the image texture details and color, and judge the correctness of the pre-fusion processing; after fusion, if the image brightness is low or the grayscale is narrow, use linear stretching, brightness contrast, color balance, chroma, saturation and brightness adjustment methods to adjust the tone.

8. The AI-based dynamic land use monitoring and optimization method according to claim 1, characterized in that: In image classification, two-phase remote sensing orthophoto image data are used to highlight the changed parts using computer automatic processing methods such as image difference and spectral feature variation, and the changed spots are extracted and numbered one by one based on AI. After obtaining the changed spots, GIS is used to establish the topological relationship of the land change spots and produce a land change information table.

9. AI-based land use dynamic monitoring and optimization system, characterized by: include: Image recognition module, which identifies different types of land use based on images acquired by remote sensing satellites; The image marking module marks the type of land in the image based on the acquired remote sensing image and generates marking data results; The data analysis module extracts land use indicators based on the labeling results, performs time series analysis, and generates land use assessment results; The historical data module obtains historical land use information and adds spatiotemporal dimension tags as training reference objects; The vector module processes the historical land use information of historical resource data into vector form to form a vector model; The prediction module detects the trend of land use change based on the assessment results, combines historical data to predict future land use and development trends, makes risk assessments, and generates prediction results; The planning module combines the prediction results with the requirements of land use planning to optimize land use, simulates the impact of different planning schemes based on the optimization results, and conducts integrated analysis to generate land planning adjustment suggestions.

Images