An Automatic Detection Method and System for the Non-agriculturalization of Cultivated Land

Abstract

The present invention discloses an automatic detection method and system for the non - agriculturalization of cultivated land. The method includes: obtaining a historical cultivated land vector map of a first area to be detected; obtaining real - time remote sensing image data of the first area to be detected; constructing a multi - level image segmentation model; inputting the historical cultivated land vector map and the real - time remote sensing image data into the multi - level image segmentation model to obtain small cultivated land image patches; constructing a land use screening model; inputting the small cultivated land image patches output by the multi - level image segmentation model into the land use screening model for traversal detection to obtain multi - type construction land image patches; and generating a first occupancy detection result by merging the multi - type construction land image patches. It solves the technical problems in the prior art that the production efficiency of the manual visual interpretation method is low and it is not suitable for large - area detection, the operation speed of using multi - temporal remote sensing images for cultivated land change detection is slow, the detection accuracy is low, and the real - time requirements cannot be met.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing applications, and particularly to an automatic detection method and system for the non-agriculturalization of cultivated land. Background Art

[0002] Cultivated land is an important natural resource and the material basis for agricultural production in a region. Rational utilization and effective protection of cultivated land resources are important conditions for the social and economic development of a region. The non-agriculturalization of cultivated land mainly includes six behaviors. First, it is strictly prohibited to illegally occupy cultivated land for afforestation. Second, it is strictly prohibited to build green channels beyond the standard. Third, it is strictly prohibited to illegally occupy cultivated land to dig lakes and create landscapes. Fourth, it is strictly prohibited to occupy permanent basic farmland to expand nature reserves. Fifth, it is strictly prohibited to illegally occupy cultivated land for non-agricultural construction. Sixth, it is strictly prohibited to approve and use land illegally. Among these six behaviors, the phenomenon of illegally occupying cultivated land for non-agricultural construction is relatively common. For a long time, the detection of the non-agriculturalization of cultivated land mainly relies on the combination of manual visual interpretation and field investigation. In addition, there are also many studies on the detection of cultivated land changes using multi-temporal remote sensing images. The current detection methods have low timeliness and are difficult to meet the requirements of cultivated land protection in the new era for large-scale, high-frequency, and rapid dynamic monitoring of cultivated land changes. Therefore, further requirements are put forward for the detection of cultivated land changes.

[0003] However, in the process of implementing the technical solutions of the embodiments of the present application, it is found that the above technologies have at least the following technical problems:

[0004] In the prior art, there are technical problems that the production efficiency of the manual visual interpretation method is low, it is not suitable for large-scale detection, and the operation speed of detecting cultivated land changes using multi-temporal remote sensing images is slow and the detection accuracy is low, which cannot meet the real-time requirements. Summary of the Invention

[0005] Aiming at the defects in the prior art, the purpose of the embodiments of the present application is to provide an automatic detection method and system for the non-agriculturalization of cultivated land, so as to solve the technical problems in the prior art that the production efficiency of the manual visual interpretation method is low, it is not suitable for large-scale detection, and the operation speed of detecting cultivated land changes using multi-temporal remote sensing images is slow and the detection accuracy is low, which cannot meet the real-time requirements. The technical effect is achieved of improving the accuracy and calculation speed of cultivated land change detection by constructing an image segmentation model and a land use screening model, being suitable for engineering applications, and having promotability.

[0006] On the one hand, an embodiment of the present application provides an automatic detection method for non-agricultural conversion of cultivated land. The method includes: obtaining a historical cultivated land vector map of a first area to be detected; obtaining real-time remote sensing image data of the first area to be detected; constructing a multi-level image segmentation model; inputting the historical cultivated land vector map and the real-time remote sensing image data into the multi-level image segmentation model, and obtaining first output information according to the multi-level image segmentation model, where the first output information is small cultivated land image patches after image segmentation; constructing a land use screening model, where the land use screening model includes multiple construction land determination rules; inputting the small cultivated land image patches output by the multi-level image segmentation model into the land use screening model for traversal detection to obtain multiple types of construction land image patches; and generating a first occupancy detection result by performing multi-class merging on the multiple types of construction land image patches.

[0007] On the other hand, the present application also provides an automatic detection system for non-agricultural conversion of cultivated land. The system includes: a first obtaining unit configured to obtain a historical cultivated land vector map of a first area to be detected; a second obtaining unit configured to obtain real-time remote sensing image data of the first area to be detected; a first constructing unit configured to construct a multi-level image segmentation model; a first input unit configured to input the historical cultivated land vector map and the real-time remote sensing image data into the multi-level image segmentation model and obtain first output information according to the multi-level image segmentation model, where the first output information is small cultivated land image patches after image segmentation; a second constructing unit configured to construct a land use screening model, where the land use screening model includes multiple construction land determination rules; a third obtaining unit configured to input the small cultivated land image patches output by the multi-level image segmentation model into the land use screening model for traversal detection to obtain multiple types of construction land image patches; and a first generating unit configured to generate a first occupancy detection result by performing multi-class merging on the multiple types of construction land image patches.

[0008] In a third aspect, an embodiment of the present application provides an automatic detection system for non-agricultural conversion of cultivated land, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the method according to any one of the first aspect are implemented.

[0009] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0010] By adopting the method of obtaining the historical cultivated land vector map and real-time remote sensing image data of the first area to be detected, further, the obtained historical cultivated land vector map and the real-time remote sensing image data are input into a multi-level image segmentation model for image segmentation to obtain small cultivated land patches output by the model, and then a land use screening model constructed based on multiple construction land determination rules is used to detect and classify the corresponding categories of the small cultivated land patches, so as to generate multi-category construction land patches, and then further merge the multi-category construction land patches to obtain the first occupancy detection result of the final output. It achieves the technical effect of improving the accuracy and calculation speed of cultivated land change detection by constructing an image segmentation model and a land use screening model, adapting to engineering applications, and having the property of being popularizable.

[0011] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the description. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following specifically illustrates the specific embodiments of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] By reading the detailed description of the non-restrictive embodiments with reference to the following drawings, other features, objects and advantages of the present invention will become more obvious:

[0013] Figure 1 It is a schematic flow chart of an automatic detection method for cultivated land non-agriculturalization according to an embodiment of the present application;

[0014] Figure 2 It is a schematic flow chart of the first-level image segmentation of an automatic detection method for cultivated land non-agriculturalization according to an embodiment of the present application;

[0015] Figure 3 It is a schematic flow chart of the second-level image segmentation of an automatic detection method for cultivated land non-agriculturalization according to an embodiment of the present application;

[0016] Figure 4 It is a schematic flow chart of constructing a land use screening model of an automatic detection method for cultivated land non-agriculturalization according to an embodiment of the present application;

[0017] Figure 5 It is a schematic structural diagram of an automatic detection system for cultivated land non-agriculturalization according to an embodiment of the present application;

[0018] Figure 6 It is a schematic structural diagram of an exemplary electronic device according to an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] Embodiments of the present application provide an automatic detection method and system for the non - agriculturalization of cultivated land, which solve the technical problems in the prior art that the manual visual interpretation method has low production efficiency and is not suitable for large - area detection, and the operation speed of using multi - temporal remote sensing images for cultivated land change detection is slow and the detection accuracy is low, unable to meet the real - time requirements. It achieves the technical effect of improving the accuracy and calculation speed of cultivated land change detection by constructing an image segmentation model and a land use screening model, being suitable for engineering applications, and having generalizability.

[0020] Next, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the exemplary embodiments described herein.

[0021] Overview of the Application

[0022] The non - agriculturalization of cultivated land mainly includes six behaviors. First, it is strictly prohibited to illegally occupy cultivated land for afforestation. Second, it is strictly prohibited to build green channels beyond the standard. Third, it is strictly prohibited to illegally occupy cultivated land to dig lakes and create landscapes. Fourth, it is strictly prohibited to expand nature reserves by occupying permanent basic farmland. Fifth, it is strictly prohibited to illegally occupy cultivated land for non - agricultural construction. Sixth, it is strictly prohibited to approve land use and use land illegally and irregularly. Among these six behaviors, the phenomenon of illegally occupying cultivated land for non - agricultural construction is relatively common. For a long time, the detection of the non - agriculturalization of cultivated land mainly relies on the combination of manual visual interpretation and field investigation. In addition, there are also many studies on using multi - temporal remote sensing images for cultivated land change detection. The current detection methods have low timeliness and are difficult to meet the requirements of cultivated land protection in the new era for large - scale, high - frequency, and rapid dynamic monitoring of cultivated land changes. Therefore, further requirements are put forward for the detection of cultivated land changes.

[0023] In response to the above - mentioned technical problems, the general idea of the technical solution provided by the present application is as follows:

[0024] Embodiments of the present application provide an automatic detection method for the non - agriculturalization of cultivated land. The method includes: obtaining a historical cultivated land vector map of a first area to be detected; obtaining real - time remote sensing image data of the first area to be detected; constructing a multi - level image segmentation model; inputting the historical cultivated land vector map and the real - time remote sensing image data into the multi - level image segmentation model, and obtaining first output information according to the multi - level image segmentation model, where the first output information is small image patches of cultivated land after image segmentation; constructing a land use screening model, where the land use screening model includes multiple construction land determination rules; inputting the small image patches of cultivated land output by the multi - level image segmentation model into the land use screening model for traversal detection to obtain multiple types of construction land image patches; and generating a first occupancy detection result by performing multi - type merging on the multiple types of construction land image patches.

[0025] To better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings of the specification and specific embodiments.

[0026] Example 1

[0027] As Figure 1 shown, an embodiment of the present application provides an automatic detection method for non-agricultural conversion of cultivated land, and the method includes:

[0028] Step S100: Obtain the historical cultivated land vector map of the first area to be detected;

[0029] Step S200: Obtain the real-time remote sensing image data of the first area to be detected;

[0030] Specifically, the first area to be detected is the area where cultivated land change detection is required, and the historical cultivated land vector map is the historical cultivated land vector map of the first area to be detected. The graphic elements in the vector file are called patches. The historical cultivated land vector map can collect information based on big data for historical land surveys to obtain the historical cultivated land vector map.

[0031] Furthermore, by receiving remote sensing image data and implementing the data preprocessing process, and then using the real-time remote sensing image data as the input information of the model. For example, the historical cultivated land vector map is the cultivated land status in the previous period, and the real-time remote sensing image data is the data in the new period. For example, the historical cultivated land vector map is the cultivated land vector data obtained through land change surveys or the third national land survey in 2019. The remote sensing image data in the new period can be the remote sensing image data in 2020, or in 2021 or more recent years. The remote sensing image data in the new period must have a time interval from the historical cultivated land vector map. The historical cultivated land vector map and the real-time remote sensing image data are implemented through a data acquisition module, which is convenient for systematic and process-based data management.

[0032] Step S300: Construct a multi-level image segmentation model;

[0033] Step S400: Input the historical cultivated land vector map and the real-time remote sensing image data into the multi-level image segmentation model, and obtain the first output information according to the multi-level image segmentation model, where the first output information is the small cultivated land patches after image segmentation;

[0034] Specifically, the multi-level image segmentation model is a pre-constructed image segmentation model process, which is a mathematical model for image processing and is stored in the image segmentation module. Further, the multi-level image segmentation model includes a first-level image segmentation and a second-level image segmentation. Specifically, in the image segmentation module, the historical cultivated land vector map and real-time remote sensing image data are sent into a preset multi-level image segmentation model. The first-level image segmentation is an image segmentation that combines vector data. Using the boundary information of the historical cultivated land vector patches as prior knowledge, the image data of the new period is segmented according to the vector patch boundaries to obtain image patches consistent with the vector patch boundaries, which are used as the image object layer. The second-level image segmentation is to perform a second-level image segmentation on the cultivated land image patches on the image object layer. Thus, the operation of the multi-level image segmentation model is realized through the first-level image segmentation and the second-level image segmentation, and then the small cultivated land image patches after image segmentation output by the multi-level image segmentation model are obtained. Furthermore, the logical segmentation output of the multi-level image segmentation model is realized, thereby improving the detection accuracy.

[0035] Step S500: Construct a land use screening model, where the land use screening model includes multiple construction land determination rules;

[0036] Step S600: Input the small cultivated land image patches output by the multi-level image segmentation model into the land use screening model for traversal detection to obtain multiple types of construction land image patches;

[0037] Specifically, the land use screening model is obtained by constructing multiple construction land determination rules. The land use screening model is used to detect construction land image patches in small cultivated land image patches. Specifically, the small cultivated land image patches are sent into the land use screening model, and the land use screening model is used to detect multiple types of construction land in the small cultivated land image patches. And each type of construction land has a first color identifier. For example, preferably, the multiple construction land determination rules include 4 determination rules, that is, the preset land use screening model includes: a preset blue construction land determination rule, a preset red construction land determination rule, a preset white construction land determination rule, and a preset gray construction land determination rule. Thus, after the small cultivated land image patches output by the multi-level image segmentation model are input, the corresponding blue construction land image patches, red construction land image patches, white construction land image patches, and gray construction land image patches are obtained through traversal, and then multiple types of construction land image patches after screening detection are obtained, achieving the technical effect of accurately outputting multiple types of construction land occupying cultivated land and realizing efficient and accurate detection.

[0038] Step S700: Generate a first occupancy detection result by performing multi-class merging on the multiple types of construction land image patches.

[0039] Specifically, the multi-category construction land image patches are merged in multiple categories to output the occupancy detection results of the construction land categories. Through the method of multi-category merging, the detection results of actual cultivated land occupancy can be systematically output. It achieves the technical effects of improving the accuracy and calculation speed of cultivated land change detection by constructing an image segmentation model and a land use screening model, being suitable for engineering applications, and having generalizability.

[0040] Further, as Figure 2 shown, step S400 of the embodiment of the present application further includes:

[0041] Step S410: The multi-level image segmentation model includes a first-level image segmentation model and a second-level image segmentation model;

[0042] Step S420: Obtain the vector patch boundary information according to the historical cultivated land vector map;

[0043] Step S430: The first-level image segmentation model performs first-level image segmentation on the real-time remote sensing image data according to the vector patch boundary information to obtain the first-level image segmentation result.

[0044] Specifically, the vector patch boundary information is the boundary information of the historical cultivated land vector map, so that the first-level segmentation of the first-level image segmentation model can be realized according to the historical patch boundary information. Specifically, using the boundary information of the historical cultivated land vector map as prior knowledge, the real-time remote sensing image data is segmented according to the vector patch boundary to obtain image patches consistent with the vector patch boundary as the image object layer.

[0045] Further, as Figure 3 shown, the steps of the embodiment of the present application further include:

[0046] Step S440: Obtain the patch attribute category information of the historical cultivated land vector map based on the attribute table;

[0047] Step S450: Classify the patches of the first-level image segmentation result according to the patch attribute category information to obtain a first classification result and a second classification result, where the first classification result is the cultivated land patch and the second classification result is the non-cultivated land patch;

[0048] Step S460: Use the first classification result as the input information of the second-level image segmentation model for second-level image segmentation to obtain the cultivated land small patches output by the first output information.

[0049] Specifically, according to the field information in the patch attribute table of the historical cultivated land vector, the attribute category information of the historical cultivated land vector patches is obtained, and the attribute category information of the historical cultivated land vector patches is used as prior knowledge to classify the first-level image segmentation results, generating cultivated land image patches, and classifying the remaining unclassified image patches on the image object layer as non-cultivated land image patches. On the image object layer, a second-level image segmentation is performed on the cultivated land image patches, and all the cultivated land image patches are re-segmented using the multi-resolution segmentation method. The multi-resolution segmentation uses the region merging algorithm with the minimum heterogeneity and performs image segmentation under the set segmentation parameters to obtain small cultivated land patches; the multi-resolution segmentation formula is:

[0050] F = W1 × H color + W2 × H smoothness + W3 × H compactness ;

[0051] where F is the heterogeneity of the image object; W1, W2, and W3 are the weights of color, smoothness, and compactness respectively, and W1 + W2 + W3 = 1.

[0052] Furthermore, as Figure 4 shown, step S500 of the embodiment of the present application further includes:

[0053] Step S510: Obtain the first land use determination feature, the second land use determination feature, and the third land use determination feature, where the first land use determination feature is the hue feature, the second land use determination feature is the brightness feature, and the third land use determination feature is the saturation feature;

[0054] Step S520: Obtain the multiple construction land determination rules by performing parameter configuration on the first land use determination feature, the second land use determination feature, and the third land use determination feature;

[0055] Step S530: Construct the land use screening model according to the multiple construction land determination rules.

[0056] Specifically, the small cultivated land patches are sent into the land use screening model, and multiple types of construction land in the small cultivated land patches are detected using the land use screening model. The preset land use screening model includes: the preset blue construction land determination rule, the preset red construction land determination rule, the preset white construction land determination rule, and the preset gray construction land determination rule. By determining the number of determination rules and the parameter configuration of the rules, the multiple construction land determination rules are obtained, and then an accurate screening model is constructed. For example, when Hue is the hue feature, Intensity is the brightness feature, and Saturation is the saturation feature, the formula for the preset blue construction land determination rule is:

[0057] f (蓝色建设用地)={0.55 < Hue < 0.65 and Intensity > 0.5 and Saturation > {0.25, 0.35}};

[0058] The preset determination rule for red construction land has the formula:

[0059] f (红色建设用地) ={0.7 < Hue and Intensity > {0.4, 0.6} and Saturation > {0.16, 0.3}}

[0060] The preset determination rule for white construction land has the formula:

[0061] f (白色建设用地) ={Intensity > 0.7 and Saturation < 0.1}

[0062] The preset determination rule for gray construction land has the formula:

[0063] f (灰色建设用地) ={0.6 < Hue < 0.65 and 0.59 < Intensity < 0.85 and 0.2 < Saturation < 0.3}

[0064] Furthermore, by traversing each small cultivated land image patch, the small cultivated land image patches that meet the determination rule of blue construction land in the preset land use screening model are classified as blue construction land image patches; the small cultivated land image patches that meet the determination rule of red construction land in the preset land use screening model are classified as red construction land image patches; the small cultivated land image patches that meet the determination rule of white construction land in the preset land use screening model are classified as white construction land image patches; the small cultivated land image patches that meet the determination rule of gray construction land in the preset land use screening model are classified as gray construction land image patches. Thus, it has high usability and adaptability.

[0065] Furthermore, by performing multi-class merging on the multi-class construction land image patches to generate the first occupancy detection result, step S700 of the embodiment of the present application further includes:

[0066] Step S710: Obtain a first merging result by performing multi-class merging on the multi-class construction land image patches;

[0067] Step S720: Obtain a first preset area for mapping;

[0068] Step S730: Exclude the construction land image patches smaller than the first preset area for mapping from the first merging result to generate the first occupancy detection result.

[0069] Specifically, the first merging result is the result of merging multi-category construction land image patches. The first preset mapping area is the critical value of the minimum mapping area. By calculating the areas of the image patches in the first merging result, and then removing the small construction land image patches with areas smaller than the preset minimum mapping area, the automatically detected construction land result is finally output. Finally, the automatic detection of non-agriculturalization of cultivated land is realized, and the areas that occupy cultivated land for non-agricultural construction can be detected in a large range, with high frequency, quickly and accurately, which is suitable for engineering applications and has the property of being popularizable.

[0070] Compared with the prior art, the present invention has the following beneficial effects:

[0071] 1. Since the historical cultivated land vector map of the first area to be detected and the real-time remote sensing image data are obtained, and further, the obtained historical cultivated land vector map and the real-time remote sensing image data are input into a multi-level image segmentation model for image segmentation to obtain the small cultivated land image patches output by the model, and then a land use screening model constructed based on multiple construction land determination rules is used to detect and classify the corresponding categories of the small cultivated land image patches, so as to generate multi-category construction land image patches, and then further merge the multi-category construction land image patches to obtain the finally output first occupancy detection result. The technical effect of improving the accuracy and calculation speed of cultivated land change detection, being suitable for engineering applications and having the property of being popularizable is achieved by constructing an image segmentation model and a land use screening model.

[0072] 2. Since the automatic detection of non-agriculturalization of cultivated land is realized by parametric configuration of the land use screening model, the areas that occupy cultivated land for non-agricultural construction can be detected in a large range, with high frequency, quickly and accurately, which is suitable for engineering applications and has the property of being popularizable.

[0073] Example 2

[0074] Based on the same inventive concept as an automatic detection method for non-agriculturalization of cultivated land in the foregoing embodiment, the present invention also provides an automatic detection system for non-agriculturalization of cultivated land, as Figure 5 shown, the system includes:

[0075] The first obtaining unit 11: The first obtaining unit 11 is used to obtain the historical cultivated land vector map of the first area to be detected;

[0076] The second obtaining unit 12: The second obtaining unit 12 is used to obtain the real-time remote sensing image data of the first area to be detected;

[0077] The first constructing unit 13: The first constructing unit 13 is used to construct a multi-level image segmentation model;

[0078] First input unit 14: The first input unit 14 is used to input the historical cultivated land vector map and the real-time remote sensing image data into the multi-level image segmentation model, and obtain first output information according to the multi-level image segmentation model, where the first output information is the small cultivated land image patches after image segmentation;

[0079] Second construction unit 15: The second construction unit 15 is used to construct a land use screening model, where the land use screening model includes multiple construction land determination rules;

[0080] Third acquisition unit 16: The third acquisition unit 16 is used to input the small cultivated land image patches output by the multi-level image segmentation model into the land use screening model for traversal detection to obtain various types of construction land image patches;

[0081] First generation unit 17: The first generation unit 17 is used to generate a first occupancy detection result by performing multi-class merging on the various types of construction land image patches.

[0082] Further, the system further includes:

[0083] First inclusion unit: The first inclusion unit is used to indicate that the multi-level image segmentation model includes a first-level image segmentation model and a second-level image segmentation model;

[0084] Fourth acquisition unit: The fourth acquisition unit is used to obtain vector patch boundary information according to the historical cultivated land vector map;

[0085] First output unit: The first output unit is used to indicate that the first-level image segmentation model performs first-level image segmentation on the real-time remote sensing image data according to the vector patch boundary information to obtain a first-level image segmentation result.

[0086] Further, the system further includes:

[0087] Fifth acquisition unit: The fifth acquisition unit is used to obtain the patch attribute category information of the historical cultivated land vector map based on the attribute table;

[0088] First classification unit: The first classification unit is used to classify the image patches of the first-level image segmentation result according to the patch attribute category information to obtain a first classification result and a second classification result, where the first classification result is cultivated land image patches and the second classification result is non-cultivated land image patches;

[0089] Second input unit: The second input unit is used to use the first classification result as the input information of the second-level image segmentation model for second-level image segmentation to obtain the small cultivated land image patches output by the first output information.

[0090] Further, the system further includes:

[0091] A sixth acquisition unit: The sixth acquisition unit is configured to acquire a first land use determination feature, a second land use determination feature, and a third land use determination feature, where the first land use determination feature is a hue feature, the second land use determination feature is a brightness feature, and the third land use determination feature is a saturation feature;

[0092] A first configuration unit: The first configuration unit is configured to obtain the multiple construction land determination rules by performing parameter configuration on the first land use determination feature, the second land use determination feature, and the third land use determination feature;

[0093] A third construction unit: The third construction unit is configured to construct the land use screening model according to the multiple construction land determination rules.

[0094] Further, the system further includes:

[0095] A seventh acquisition unit: The seventh acquisition unit is configured to obtain a first merging result by performing multi-class merging on the multi-class construction land image patches;

[0096] An eighth acquisition unit: The eighth acquisition unit is configured to acquire a first preset area for mapping;

[0097] A first elimination unit: Eliminate the construction land image patches smaller than the first preset area for mapping from the first merging result to generate the first occupancy detection result.

[0098] The foregoing Figure 1 All the various change modes and specific examples of the automatic detection method for cultivated land non-agriculturalization in Embodiment 1 are equally applicable to the automatic detection system for cultivated land non-agriculturalization in this embodiment. Through the foregoing detailed description of the automatic detection method for cultivated land non-agriculturalization, those skilled in the art can clearly know the implementation method of the automatic detection system for cultivated land non-agriculturalization in this embodiment. Therefore, for the sake of simplicity of the specification, it will not be elaborated herein.

[0099] Example 3

[0100] Next, reference is made to Figure 6 to describe the electronic device according to an embodiment of the present application.

[0101] Figure 6 The structural schematic diagram of the electronic device according to an embodiment of the present application is illustrated.

[0102] Based on the inventive concept of an automatic detection method for non - agricultural conversion of cultivated land in the foregoing example, the present invention also provides an automatic detection system for non - agricultural conversion of cultivated land, on which a computer program is stored. When the program is executed by a processor, it implements the steps of any of the methods of the foregoing automatic detection system for non - agricultural conversion of cultivated land.

[0103] Among them, in Figure 6 In Figure 6 , the bus architecture (represented by bus 300), bus 300 may include any number of interconnected buses and bridges. Bus 300 links together various circuits including one or more processors represented by processor 302 and a memory represented by memory 304. Bus 300 can also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well - known in the art. Therefore, they will not be further described herein. Bus interface 305 provides an interface between bus 300 and receiver 301 and transmitter 303. Receiver 301 and transmitter 303 can be the same element, i.e., a transceiver, providing a unit for communicating with various other systems over a transmission medium. Processor 302 is responsible for managing bus 300 and general processing, while memory 304 can be used to store data used by processor 302 when performing operations.

[0104] The embodiments of the present application provide an automatic detection method for non - agricultural conversion of cultivated land. The method includes: obtaining a historical cultivated land vector map of a first area to be detected; obtaining real - time remote sensing image data of the first area to be detected; constructing a multi - level image segmentation model; inputting the historical cultivated land vector map and the real - time remote sensing image data into the multi - level image segmentation model, and obtaining first output information according to the multi - level image segmentation model, where the first output information is small cultivated land image patches after image segmentation; constructing a land use screening model, where the land use screening model includes multiple construction land determination rules; inputting the small cultivated land image patches output by the multi - level image segmentation model into the land use screening model for traversal detection to obtain multiple types of construction land image patches; and generating a first occupancy detection result by performing multi - type merging on the multiple types of construction land image patches. This solves the technical problems in the prior art that the manual visual interpretation method has low production efficiency, is not suitable for large - area detection, and the operation speed of using multi - temporal remote sensing images for cultivated land change detection is slow and the detection accuracy is low, unable to meet real - time requirements. It achieves the technical effect of improving the accuracy and calculation speed of cultivated land change detection by constructing an image segmentation model and a land use screening model, being suitable for engineering applications, and having generalizability.

[0105] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) that contain computer-usable program code.

[0106] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate a system for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or blocks or combinations of blocks.

[0107] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction system that implements the functions specified in Figure 1 one or more of the flows Figure 1 or blocks or combinations of blocks.

[0108] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are performed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or blocks or combinations of blocks.

[0109] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concepts. Therefore, the appended claims are intended to be construed as including the preferred embodiments as well as all changes and modifications falling within the scope of the present invention.

[0110] Obviously, those skilled in the art can make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. An automatic detection method for the non-agriculturalization of cultivated land, characterized in that, The method includes: Obtaining a historical cultivated land vector map of a first area to be detected; Obtaining real-time remote sensing image data of the first area to be detected; Constructing a multi-level image segmentation model; Inputting the historical cultivated land vector map and the real-time remote sensing image data into the multi-level image segmentation model, and obtaining first output information according to the multi-level image segmentation model, where the first output information is small cultivated land image patches after image segmentation, including: The multi-level image segmentation model includes a first-level image segmentation model and a second-level image segmentation model; obtaining vector patch boundary information according to the historical cultivated land vector map; the first-level image segmentation model performing first-level image segmentation on the real-time remote sensing image data according to the vector patch boundary information to obtain a first-level image segmentation result; obtaining patch attribute category information of the historical cultivated land vector map based on an attribute table; classifying the patches in the first-level image segmentation result according to the patch attribute category information to obtain a first classification result and a second classification result, where the first classification result is cultivated land image patches and the second classification result is non-cultivated land image patches; using the first classification result as input information for the second-level image segmentation model to perform second-level image segmentation to obtain the small cultivated land image patches output as the first output information; Constructing a land use screening model, where the land use screening model includes multiple construction land determination rules; Inputting the small cultivated land image patches output by the multi-level image segmentation model into the land use screening model for traversal detection to obtain multiple types of construction land image patches; Generating a first occupancy detection result by performing multi-class merging on the multiple types of construction land image patches, including: obtaining a first merging result by performing multi-class merging on the multiple types of construction land image patches; obtaining a first preset mapping area; excluding construction land image patches smaller than the first preset mapping area from the first merging result to generate the first occupancy detection result.

2. The method according to claim 1, characterized in that, The second-level image segmentation uses a multi-resolution segmentation formula to perform second-level image segmentation on the first classification result, and the multi-resolution segmentation formula is: F = W1 × H color + W2 × H smoothness + W3 × H compactness where F is the heterogeneity of the image object; W1, W2, and W3 are the weights of color, smoothness, and compactness respectively, and W1 + W2 + W3 = 1.

3. The method according to claim 1, wherein Regarding the construction of the land use screening model, the method further includes: Obtaining a first land use determination feature, a second land use determination feature, and a third land use determination feature, where the first land use determination feature is a hue feature, the second land use determination feature is a brightness feature, and the third land use determination feature is a saturation feature; Obtaining the multiple construction land determination rules by performing parameter configuration on the first land use determination feature, the second land use determination feature, and the third land use determination feature; Constructing the land use screening model according to the multiple construction land determination rules.

4. An automatic detection system for the non-agriculturalization of cultivated land, characterized in that, The system includes: A first obtaining unit: The first obtaining unit is used to obtain a historical cultivated land vector map of a first area to be detected; A second obtaining unit: The second obtaining unit is used to obtain real-time remote sensing image data of the first area to be detected; A first constructing unit: The first constructing unit is used to construct a multi-level image segmentation model; First input unit: The first input unit is used to input the historical cultivated land vector map and the real-time remote sensing image data into the multi-level image segmentation model, and obtain first output information according to the multi-level image segmentation model. The first output information is the cultivated land small image patches after image segmentation, including: The multi-level image segmentation model includes a first-level image segmentation model and a second-level image segmentation model; According to the historical cultivated land vector map, vector map patch boundary information is obtained; The first-level image segmentation model performs first-level image segmentation on the real-time remote sensing image data according to the vector map patch boundary information, and obtains a first-level image segmentation result; Based on the attribute table, the patch attribute category information of the historical cultivated land vector map is obtained; The first-level image segmentation result is classified into image patches according to the patch attribute category information, and a first classification result and a second classification result are obtained, where the first classification result is cultivated land image patches, and the second classification result is non-cultivated land image patches; The first classification result is used as the input information of the second-level image segmentation model for second-level image segmentation, and the cultivated land small image patches output as the first output information are obtained. Second construction unit: The second construction unit is used to construct a land use screening model, where the land use screening model includes multiple construction land determination rules. Third acquisition unit: The third acquisition unit is used to input the cultivated land small image patches output by the multi-level image segmentation model into the land use screening model for traversal detection, and obtain multiple types of construction land image patches. First generation unit: The first generation unit is used to generate a first occupancy detection result by performing multi-class merging on the multiple types of construction land image patches, including: By performing multi-class merging on the multiple types of construction land image patches, a first merging result is obtained; A first preset mapping area is obtained; Construction land image patches smaller than the first preset mapping area are removed from the first merging result to generate the first occupancy detection result.

5. An automatic detection system for the non-agricultural conversion of cultivated land, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method according to any one of claims 1-3.

Images