Inventory method of natural resource assets owned by all people based on real-life 3D model

Abstract

The embodiments of the present application relate to the field of natural resource asset inventory, and specifically to a method for inventorying natural resource assets owned by the whole people based on a real-life three-dimensional model. The specific implementation of the method includes: generating a local resource area mask image through a resource area segmentation model according to resource area division image feature information and resource area mask division image; generating resource area segmentation information; matching natural resource asset area information from a natural resource asset database according to resource area segmentation information and resource area geographic reference information; performing real-life three-dimensional image acquisition on the resource area corresponding to the resource area segmentation information; inputting the resource area real-life three-dimensional model, resource area segmentation information and natural resource asset area information into a natural resource asset inventory identification model to obtain resource area asset inventory identification information. This implementation can quickly and accurately identify resource area assets of each individual resource area within the natural resource asset inventory area.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of natural resource asset inventory, and specifically to a method for inventorying natural resource assets owned by the whole people based on a real-life three-dimensional model. Background Art

[0002] The inventory of natural resource assets owned by the whole people is an important way to find out the assets. It is based on the results of various special surveys (inventories) on resource ownership, quantity, quality, use, distribution, etc., and through a unified benchmark time point (period) to supplement the investigation of prices, use rights, income, etc., estimate the economic value of assets, and basically grasp the situation of six types of natural resource assets owned by the whole people, such as land, minerals, oceans, forests, grasslands, and wetlands, to provide important support for the effective performance of the responsibilities of the owner of natural resource assets owned by the whole people. In the existing natural resource asset inventory plan, the method usually adopted is to conduct an inventory of natural resource assets through manual statistics. However, the above method often has the following technical problems: manual statistics are time-consuming and have statistical errors, making it difficult to effectively manage and protect natural resources. Summary of the invention

[0003] The content of this application is used to introduce concepts in a brief form, which will be described in detail in the detailed implementation section below. The content of this application is not intended to identify the key features or essential features of the technical solution claimed for protection, nor is it intended to limit the scope of the technical solution claimed for protection.

[0004] Some embodiments of the present application propose a method for inventorying natural resource assets owned by the whole people based on a real-life three-dimensional model, a computer device, and a computer-readable storage medium to solve one or more of the technical problems mentioned in the above background technology section.

[0005] In a first aspect, some embodiments of the present application provide a method for inventorying natural resource assets owned by the whole people based on a real-life three-dimensional model, the method comprising: performing resource area segmentation on a regional remote sensing image of a natural resource asset inventory area to be inspected, and obtaining a resource area segmentation image set, wherein the natural resource asset inventory area has corresponding resource area geographic reference information; for each resource area segmentation image in the resource area segmentation image set, performing the following identification steps: generating resource area segmentation information corresponding to the resource area segmentation image; matching the corresponding natural resource asset area information from a pre-established natural resource asset database based on the resource area segmentation information and the resource area geographic reference information; performing real-life three-dimensional image acquisition on the resource area corresponding to the resource area segmentation information, and obtaining a corresponding resource area real-life three-dimensional image set; and constructing a resource area real-life three-dimensional model based on the resource area real-life three-dimensional image set; inputting the resource area real-life three-dimensional model, the resource area segmentation information, and the natural resource asset area information into a pre-trained natural resource asset inventory identification model to obtain corresponding resource area asset inventory identification information; and storing each resource area asset inventory identification information in the corresponding natural resource asset inventory database owned by the whole people.

[0006] In a second aspect, the present application also provides a computer device, comprising a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, the method described in any implementation of the first aspect is implemented.

[0007] In a third aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, wherein when the computer program is executed by a processor, the method described in any implementation of the first aspect is implemented.

[0008] The above-mentioned various embodiments of the present application have the following beneficial effects: through the method for inventorying natural resource assets owned by the whole people based on the real-life three-dimensional model of some embodiments of the present application, the resource regional assets of each individual resource area in the natural resource asset inventory area can be quickly and accurately identified, thereby facilitating the management and protection of resource regional assets according to different resource areas. First, obtain the regional remote sensing image corresponding to the area to be inventoried for natural resource assets and the resource regional geographic reference information corresponding to the above-mentioned area to be inventoried for natural resource assets. Thus, data support is provided for identifying resource regional assets in different resource areas. Secondly, the above-mentioned regional remote sensing image is segmented into resource regions to obtain a resource region segmentation image set. Thus, separate resource regions can be segmented to facilitate asset differentiation. Afterwards, for each resource region segmentation image in the resource region segmentation image set, the following identification steps are performed: generating resource region segmentation information corresponding to the resource region segmentation image; matching the corresponding natural resource asset region information from a pre-established natural resource asset database according to the resource region segmentation information and the resource region geographic reference information; collecting real-scene three-dimensional images of the resource region corresponding to the resource region segmentation information to obtain the corresponding resource region real-scene three-dimensional image set, and constructing a resource region real-scene three-dimensional model according to the resource region real-scene three-dimensional image set; inputting the resource region real-scene three-dimensional model, the resource region segmentation information and the natural resource asset region information into the pre-trained natural resource asset inventory identification model to obtain the corresponding resource region asset inventory identification information. Thus, accurate image segmentation for resource region segmentation images is achieved through a multi-level segmentation method. Thus, other non-resource regions in the roughly segmented resource region segmentation image can be segmented out. Furthermore, resource region asset inventory is performed on the resource region through the pre-trained natural resource asset inventory identification model. The resource region assets of each individual resource region in the natural resource asset inventory area can be quickly and accurately identified. Finally, the asset inventory identification information of each resource area is stored in the corresponding national natural resource asset inventory database. Thus, the resource area assets of each individual resource area in the natural resource asset inventory area can be quickly and accurately identified, thereby facilitating the management and protection of resource area assets according to different resource areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The above and other features, advantages and aspects of the embodiments of the present application will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that components and elements are not necessarily drawn to scale.

[0010] Figure 1 It is a flowchart of some embodiments of the method for checking the natural resource assets owned by all people based on the real-life three-dimensional model according to the present application; Figure 2 is a schematic diagram of a computer device suitable for implementing some embodiments of the present application; Figure 3 It is a schematic diagram of the grid resolution of the resource area division image in the method for checking the natural resource assets owned by all people based on the real-life three-dimensional model of this application; Figure 4 It is a schematic diagram of pixel supplementation of resource area segmentation image in the method for inventorying natural resources assets owned by the whole people based on the real-life three-dimensional model of the present application. DETAILED DESCRIPTION

[0011] Embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be construed as being limited to the embodiments set forth herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present application. It should be understood that the drawings and embodiments of the present application are only for exemplary purposes and are not intended to limit the scope of protection of the present application.

[0012] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other.

[0013] It should be noted that the concepts such as "first" and "second" mentioned in this application are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0014] It should be noted that the modifications of "one" and "plurality" mentioned in the present application are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0015] The names of the messages or information exchanged between multiple devices in the embodiments of the present application are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0016] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments.

[0017] Figure 1 The process 100 of some embodiments of the method for checking the natural resource assets owned by the whole people based on the real scene 3D model according to the present application is shown. The method for checking the natural resource assets owned by the whole people based on the real scene 3D model comprises the following steps: Step 101, performing resource region segmentation on the regional remote sensing image of the natural resource asset inventory area to obtain a resource region segmentation image set.

[0018] In some embodiments, the execution subject (e.g., computing device) of the method for checking all natural resource assets owned by the whole people based on the real-life three-dimensional model can perform resource area segmentation on the regional remote sensing image of the area to be checked for natural resource assets, and obtain a resource area segmentation image set, wherein the area to be checked for natural resource assets has corresponding resource area geographic reference information.

[0019] For example, the execution subject may obtain the regional remote sensing image corresponding to the area to be checked for natural resource assets and the resource area geographic reference information corresponding to the area to be checked for natural resource assets from the pre-built database. The area to be checked for natural resource assets may be an area to be checked for natural resource assets. For example, the area to be checked for natural resource assets may be an area to be checked for various resources including land resources, mineral resources, water resources, etc. of each user. The regional remote sensing image may refer to a remote sensing image of a natural resource area taken by a remote sensing satellite technology shooting device. The resource area geographic reference information may be a real geographic reference data set pre-stored for each different resource area in the area to be checked for natural resource assets. For example, the resource area in each resource area may refer to a land resource area, a forest resource area, or a water resource area owned by a certain user. For another example, the resource area in each resource area may refer to a mineral area or a lake area. For example, the resource area geographic reference information may be a plurality of resource area key-value pairs composed of each resource area in the area to be checked for natural resource assets. The key in the resource region key-value pair is the resource region identifier, and the pair in the resource region key-value pair is the geographic reference information corresponding to the resource region. The pre-built database pre-stores the regional remote sensing images corresponding to the area to be inventoried for natural resource assets and the resource region geographic reference information corresponding to the area to be inventoried for natural resource assets. The regional remote sensing images can be scaled, and the resolution can be shot at the specified resolution to ensure image clarity.

[0020] For example, the above-mentioned execution entity can send the regional remote sensing image to the associated image segmentation end, so that the technician can operate the image segmentation end to segment or edge detect each individual resource area in the regional remote sensing image to obtain a resource area segmentation image of the individual resource area.

[0021] In an actual application scenario, the execution subject can perform resource region segmentation on the above-mentioned regional remote sensing image through the following steps to obtain a resource region segmentation image set: The first step is to perform correction processing on the above regional remote sensing image to obtain a corrected regional remote sensing image. For example, the correction processing may include: radiation correction, atmospheric correction, and geometric correction.

[0022] In the second step, the above-mentioned correction area remote sensing image is input into a pre-trained resource area segmentation model to obtain a resource area segmentation image set. The resource area segmentation model may refer to a neural network model for segmenting a single independent resource area based on image recognition technology. For example, the resource area segmentation model may be a YOLO model.

[0023] Furthermore, the resource region segmentation model can be trained through the following steps: The first step is to obtain remote sensing images of the sample area.

[0024] The second step is to determine the network structure of the initial resource region segmentation model. The initial resource region segmentation model includes: a convolutional network, a pooling network, and an upsampling layer. For example, the convolutional network may include multiple convolutional layers. The pooling network may include an average pooling layer.

[0025] The third step is to input the remote sensing images of the sample area into the initial resource area segmentation model to obtain an initial resource area segmentation image set.

[0026] The fourth step is to determine the difference between the initial resource region segmentation image set and the corresponding sample segmentation label. For example, the difference between the initial resource region segmentation image set and the corresponding sample segmentation label can be determined by a preset loss function. The loss function can be a hinge loss function or a cross entropy loss function.

[0027] In a fifth step, in response to determining that the difference value is less than or equal to a preset threshold, the initial resource region segmentation model is determined as a trained resource region segmentation model.

[0028] Step 102: for each resource region segmentation image in the resource region segmentation image set, perform the following identification steps: Step 1021, generating resource region segmentation information corresponding to the resource region segmentation image.

[0029] In some embodiments, the execution entity may generate resource region segmentation information corresponding to the resource region segmentation image.

[0030] In an actual application scenario, the execution subject generates resource region segmentation information corresponding to the resource region segmentation image through the following steps: The first step is to generate a resource region division image feature information set and global resource region division image feature information corresponding to the above resource region division image. The resource region division image feature information can represent the image semantic content corresponding to the resource region division image. The resource region division image feature information can be the semantic content corresponding to a part of the image region in the resource region division image. The resource region division image feature information can be feature information in vector form. One resource region division image feature information corresponds to one resource region division image. The global resource region segmentation image feature information can be a vector form representing the overall resource region segmentation image feature semantic content corresponding to the resource region segmentation image.

[0031] It should be clear that the resource region segmentation image is only a remote sensing image of a separate resource region that has been segmented. The resource region segmentation image may contain images of other non-resource regions, and therefore needs to be further divided / segmented.

[0032] For example, the execution subject may generate the resource region partition image feature information set and the global resource region segmentation image feature information corresponding to the resource region segmentation image through the following steps: First, the resolution of the mask image output by the resource region segmentation model is determined as the model resolution, wherein the input and output corresponding to the resource region segmentation model and the model structure are predetermined, and the corresponding model resolution is also predetermined.

[0033] Second, according to the resolution corresponding to the resource region segmentation image and the model resolution, the resource region segmentation image is divided to obtain a resource region segmentation image set. The resolution corresponding to each resource region segmentation image is the same as the model resolution. For example, the execution subject can determine whether the pixel points of the resolution corresponding to the resource region segmentation image in the horizontal and vertical directions of the screen are integer multiples of the pixel points of the model resolution in the horizontal and vertical directions of the screen. Afterwards, if it is an integer multiple, the image corresponding to the resolution is evenly divided with the area corresponding to the model resolution as a division unit to obtain a resource region segmentation image set. Next, if it is not an integer multiple, the resource region segmentation image under the above resolution is filled in the corresponding direction so that the resolution corresponding to the filled resource region segmentation image is an integer multiple of the model resolution in each direction. Then, the area corresponding to the model resolution is used as a division unit, and the filled resource region segmentation image is evenly divided to obtain a resource region segmentation image set.

[0034] Third, each resource region division image in the resource region division image set is input into a vector model to obtain a resource region division image feature information set. The vector model may be a neural network model that converts an image into a vector form and represents the semantic content of image features. For example, the vector model may be an embedding model.

[0035] Fourth, the resource region segmentation image is input into the vector model to obtain corresponding global resource region segmentation image feature information.

[0036] like Figure 3 As shown, the corresponding resolution of the resource area segmentation image is 512*512 and the model resolution can be 256*256. The resource area segmentation image can be grid-divided into 4 blocks, and the resolution of the corresponding grid of each block is 256*256.

[0037] like Figure 4 As shown in the figure, the corresponding resolution of the resource region segmentation image is 1920*1080 and the model resolution can be 256*256. 1920 and 1080 cannot divide 256 evenly, so the resource region segmentation image is padded with padding pixels to obtain a padded image of 2048*1280. The padded image of 2048*1280 is then evenly divided to obtain a resource region segmentation image set in which each resource region segmentation image is 256*256.

[0038] In the second step, based on the above-mentioned global resource region segmentation image feature information, a pre-trained resource region segmentation model is used to generate a global resource region mask image.

[0039] For example, the execution subject may generate a global resource region mask image based on the feature information of the global resource region segmentation image using a pre-trained resource region segmentation model. The global resource region mask image may be an image representing the mask position information of the resource region in the resource region segmentation image. The image region where the resource region is located in the global resource region mask image is set to black, and the image region where the non-resource region is located is set to yellow. The resource region segmentation model may be a neural network model for segmenting the resource region in the image. For example, the resource region segmentation model may be an encoding and decoding model. For example, the resource region segmentation model may be a U-net model. The resolution corresponding to the global resource region mask image is the same as the resolution corresponding to the resource region partition image. For example, in response to determining that the resource region segmentation model is a non-interactive segmentation model, the execution subject may input the feature information of the global resource region segmentation image into the resource region segmentation model to obtain a global resource region mask image.

[0040] For another example, the resource region segmentation model is an interactive segmentation model. The interactive segmentation model may be a neural network model that performs image segmentation based on interactive information. For example, the interactive segmentation model may be a SAM (SegmentAnything Model) model. The above-mentioned overall image feature information and resource region image marking information are input into the above-mentioned interactive segmentation model to obtain a global resource region mask image. The interactive segmentation model includes: an encoding model (Encoder) and a decoding model (Dencoder). The resource region image marking information may be a positive example set and a negative example set of the corresponding image content in the resource region segmentation image. The positive example set and the negative example set may be set on the resource region segmentation image, and the corresponding number may be calibrated in turn by observing the segmentation effect. For example, a positive example may be that the corresponding point (mask area) is on the resource region in the resource region segmentation image. A negative example may be that the corresponding point is not on the resource region in the resource region segmentation image.

[0041] The third step is to perform image division on the global resource region mask image according to the division method corresponding to the resource region division image set to obtain the resource region mask division image set. The division method may be a division method in which the resource region segmentation image is divided into a resource region division image set. One resource region mask division image corresponds to one resource region division image. The resource region mask division image may be an image representing the mask position information of the resource region in the resource region division image. As an example, the division method is as follows: the above-mentioned execution subject may determine whether the pixel points of the resolution corresponding to the global resource region mask image in the horizontal and vertical directions of the screen are integer multiples of the pixel points of the model resolution in the horizontal and vertical directions of the screen. Afterwards, if it is an integer multiple, the image corresponding to the resolution is uniformly divided with the region corresponding to the model resolution as a division unit to obtain the resource region mask division image set. Then, if it is not an integer multiple, the global resource region mask image at the above resolution is filled in the corresponding direction so that the resolution corresponding to the filled global resource region mask image in each direction is an integer multiple of the model resolution in each direction. Then, the area corresponding to the model resolution is taken as a division unit, and the filled global resource region mask image is evenly divided to obtain a resource region mask division image set.

[0042] In the fourth step, for each resource region division image feature information in the resource region division image feature information set, according to the resource region division image feature information and the corresponding resource region mask division image, a local resource region mask image corresponding to the resource region division image feature information is generated through the resource region segmentation model. Wherein, if there is a resource region division image with a corresponding position relationship for the resource region division image feature information, there is also a resource region mask division image with a corresponding position relationship. Wherein, the local resource region mask image may be an image of the mask position information of the resource region in the resource region division image corresponding to the resource region division image feature information. The resolution corresponding to the local resource region mask image is equal to the resolution of the corresponding resource region division image. The segmentation of the resource region based on the overall image content may be a global segmentation of the image content corresponding to the resource region. The regional content in the resource region and the resource region may be segmented out. For example, the resource region division image feature information and the corresponding resource region mask division image may be input into the resource region segmentation model to obtain the corresponding local resource region mask image.

[0043] In an actual application scenario, the execution subject may generate a local resource region mask image corresponding to the resource region division image feature information through the following steps: First, generate image marking information corresponding to the resource region mask partition image.

[0044] Second, the resource region mask partition image is input into a vector model to obtain feature information of the resource region mask image.

[0045] Third, the resource region mask image feature information, the resource region partition image feature information and the image tag information are input into the resource region segmentation model to obtain a corresponding local resource region mask image.

[0046] The resource region mask image feature information, the resource region partition image feature information and the image tag information are input into the resource region segmentation model to obtain the corresponding local resource region mask image, including: First, the resource region division image corresponding to the resource region division image feature information is determined as the target resource region division image.

[0047] Second, at least one resource area division image that is adjacent to the target resource area division image is selected from the resource area division image set as an adjacent resource area division image group. The adjacent resource area division image may be an image content in the target resource area division image that is adjacent to the target resource area division image.

[0048] Third, the adjacent resource region division image group and the target resource region division image are combined to obtain a combined resource region division image.

[0049] Fourth, the image marking information and the resource region division image feature information subset corresponding to the above-mentioned combined resource region division image are generated, and the image marking information and the resource region division image feature information subset are used as candidate image marking information and target resource region division image subset respectively.

[0050] Fifth, the target resource region partition image subset and the candidate image labeling information are input into the resource region segmentation model to generate a combined resource region mask image.

[0051] Sixth, a sub-resource region mask image corresponding to the target resource region division image is selected from the combined resource region mask image. For example, first, the corresponding image splitting method can be determined according to the image combination method corresponding to the adjacent resource region division image group and the target resource region division image. Then, according to the image splitting method, a sub-resource region mask image corresponding to the target resource region division image is selected from the combined resource region mask image.

[0052] Seventh, the image feature information corresponding to the sub-resource region mask image, the resource region mask image feature information, the resource region partition image feature information and the image tag information are input into the resource region segmentation model to obtain the corresponding local resource region mask image. For example, the resource region mask image feature information represents the mask content corresponding to the resource region partition image feature information after the resource region segmentation image is subjected to resource region segmentation.

[0053] Therefore, inputting the regional mask features centered on the target resource region partition image into the resource region segmentation model can allow the model to learn mask content with higher quality relative to the mask image feature information, making subsequent local mask images more accurate.

[0054] The fifth step is to generate resource region segmentation information corresponding to the resource region segmentation image based on each local resource region mask image. The resource region segmentation information may be the image position information of the resource region in the resource region segmentation image, or may be the image after semantic segmentation of the resource region. For example, first, each local resource region mask image may be combined according to the division method to generate a combined resource region mask image. Then, based on the combined resource region mask image, the resource region segmentation information corresponding to the resource region is generated.

[0055] Step 1022, according to the above resource area segmentation information and resource area geographic reference information, the corresponding natural resource asset area information is matched from a pre-established natural resource asset database.

[0056] In some embodiments, the above-mentioned execution entity can match the corresponding natural resource asset area information from a pre-established natural resource asset database based on the above-mentioned resource area segmentation information and the resource area geographic reference information. The natural resource asset database can be a pre-established database corresponding to each resource area in the area to be natural resource asset inventory. The natural resource asset area information can represent the asset area information of the resource area corresponding to the above-mentioned resource area segmentation information and the resource area geographic reference information. For example, the natural resource asset area information may include: the name of the natural resource area, the type of natural resources, the output of natural resources, and the unit value sequence of natural resources (unit price). The unit value sequence of natural resources can represent the value change sequence of natural resources. For example, on January 1, the value is A; on January 2, the value is B...

[0057] Step 1023, collect real-scene 3D images of the resource area corresponding to the resource area segmentation information to obtain a corresponding real-scene 3D image set of the resource area, and construct a real-scene 3D model of the resource area based on the real-scene 3D image set of the resource area.

[0058] In some embodiments, the execution subject may collect real-scene 3D images of the resource area corresponding to the resource area segmentation information to obtain a corresponding real-scene 3D image set of the resource area, and construct a real-scene 3D model of the resource area based on the real-scene 3D image set of the resource area. For example, a drone carrying a panoramic 3D camera / 3D camera may be used to collect real-scene 3D images of the resource area corresponding to the resource area segmentation information to obtain a corresponding real-scene 3D image set of the resource area. The real-scene 3D image set of the resource area may comprehensively represent a set of 3D images of the corresponding resource area. The real-scene 3D images of each resource area in the real-scene 3D image set of the resource area have a sequence.

[0059] In an actual application scenario, the above execution entity can build a real-life 3D model of the resource area through the following steps: The first step is to generate an initial resource area real-scene three-dimensional model corresponding to the resource area segmentation image according to the resource area segmentation image feature information set and the global resource area segmentation image feature information. For example, a three-dimensional modeling tool can be used to construct an initial resource area real-scene three-dimensional model corresponding to the resource area segmentation image according to the resource area segmentation image feature information set and the global resource area segmentation image feature information. For example, the three-dimensional modeling tool can be: Maya, 3Dmax, Zbrush. It should be noted that these image feature information can be used to represent the edge features and three-dimensional features required for the initial resource area real-scene three-dimensional model.

[0060] The second step is to fill the initial resource area real-scene 3D model with the resource area real-scene 3D image according to the above-mentioned resource area real-scene 3D image set, and obtain the filled initial resource area real-scene 3D model as the resource area real-scene 3D model. For example, the Substance Painter tool can be used to fill the above-mentioned resource area real-scene 3D image set into the initial resource area real-scene 3D model in sequence; then, the edge area of ​​each resource area real-scene 3D image in the filled initial resource area real-scene 3D model is optimized, that is, the same 3D image edge pixels are removed. Thus, the resource area real-scene 3D model is obtained. It should be noted that the resource area real-scene 3D model can refer to the constructed 3D image model of the corresponding resource area.

[0061] Step 1024, input the above-mentioned resource area real-life three-dimensional model, the above-mentioned resource area segmentation information and the above-mentioned natural resource asset area information into a pre-trained natural resource asset inventory identification model to obtain corresponding resource area asset inventory identification information.

[0062] In some embodiments, the execution entity may input the above-mentioned resource area real-life three-dimensional model, the above-mentioned resource area segmentation information, and the above-mentioned natural resource asset area information into a pre-trained natural resource asset inventory identification model to obtain the corresponding resource area asset inventory identification information. The natural resource asset inventory identification model may be a neural network model for estimating and identifying the natural resource asset results of the corresponding resource area. For example, the natural resource asset inventory identification model may be a Transformer deep learning model, or a resource-capability-value model (RCV model), a CA-Markov model, a FLUS model, or an InVEST model. The resource area asset inventory identification information may represent the estimated asset value of the resource area. For example, if resource area A is a mining area, the resource area asset inventory identification information may represent a value of XX yuan. For another example, if resource area A is reservoir No. 001, the resource area asset inventory identification information may represent a value of YY yuan.

[0063] Furthermore, the natural resource asset inventory identification model can be trained through the following steps: The first step is to obtain sample data of natural resource assets.

[0064] The second step is to determine the causal coefficients of various factors affecting the value of natural resource assets based on the sample data of natural resource assets and the causal model in the initial natural resource asset inventory identification model. The above-mentioned sample data of natural resource assets include characteristic data sequences under various data dimensions, as well as sample labels of sample natural resource asset values ​​at different times and different data dimensions. The causal model may include a linear causal model (Linear DML (Double Machine Learning)), a tree series causal model (Causal Forest DML), etc. For example, the execution entity may determine the causal coefficients of various factors affecting the value of natural resource assets based on the sample data of natural resource assets and the causal model in the initial natural resource asset inventory identification model.

[0065] The second step may include: 1. Determine the appropriate causal model based on the data characteristics of different feature data sequences in the natural resource asset sample data. For example, you can select an appropriate causal model for different data features. For example, if it is linear data, you can choose a linear causal model; and if it is hierarchical data, you can choose a tree series causal model.

[0066] 2. Based on the above-mentioned natural resource asset sample data and the determined causal model, generate the causal effect of each factor affecting the value of natural resource assets. For example, the execution entity can input the natural resource asset sample data into the determined causal model, analyze the sample data through the causal model, and thus obtain the causal effect of each factor affecting the value of natural resource assets. For example, the execution entity can transform the causal effect of each factor to obtain the causal coefficient of each factor.

[0067] 3. Transform the causal effect of each factor to obtain the causal coefficient of each factor. The execution subject can perform exponential transformation on the causal effect of each factor to obtain the transformed causal coefficient.

[0068] Among them, the causal effect of each factor is transformed to obtain the causal coefficient of each factor, including: 1. Perform exponential transformation on the causal effect of each factor to obtain the transformed causal coefficient.

[0069] 2. Evaluate the transformed causal coefficients, and determine the evaluated causal coefficients as the causal coefficients of each factor.

[0070] The third step is to select the causal coefficient of the target factor from the determined causal coefficients. The target factor is a factor corresponding to the target time and target dimension. The target time and target dimension here are usually the time and dimension of the resource area value to be predicted. In the model training process, the target time and target dimension are usually the time and data dimensions in the sample data of natural resource assets. The execution entity can select the factor coefficient that matches the forecast date and granularity.

[0071] The fourth step is to obtain the basic natural resource asset identification results based on the above-mentioned natural resource asset sample data and the basic natural resource asset identification model in the above-mentioned initial natural resource asset inventory identification model. Among them, the above-mentioned basic natural resource asset identification model is used to identify the value of natural resource assets at the above-mentioned target time. The basic natural resource asset identification model can adopt the natural resource asset value estimation model commonly used in related technologies, such as the Transformer deep learning model. Here, the execution entity can input the natural resource asset sample data into the basic natural resource asset identification model to obtain the preliminary estimation results of the natural resource asset value under the above-mentioned target time and target dimension.

[0072] The fifth step is to generate the initial natural resource asset identification value output by the initial natural resource asset inventory identification model based on the causal coefficient of the target factor and the basic natural resource asset identification result. For example, the causal coefficient of the target factor can be multiplied by the basic natural resource asset identification result, and the result can be used as the initial natural resource asset identification value output by the initial natural resource asset inventory identification model.

[0073] In the sixth step, the initial natural resource asset inventory identification model is adjusted according to the initial natural resource asset identification value and the corresponding sample label to continue training the adjusted initial natural resource asset inventory identification model. For example, the loss value between the initial natural resource asset identification value and the corresponding sample label can be determined by a preset loss function. Finally, the initial natural resource asset inventory identification model can be adjusted by gradient descent.

[0074] In the seventh step, in response to determining that the training of the initial natural resource asset inventory identification model is completed, the trained initial natural resource asset inventory identification model is determined as the natural resource asset inventory identification model.

[0075] Step 103, storing the asset inventory identification information of each resource area into the corresponding public natural resource asset inventory database.

[0076] In some embodiments, the execution entity may store the asset inventory identification information of each resource region in the corresponding public natural resource asset inventory database. The public natural resource asset inventory database may be a pre-established database for storing the asset inventory identification information of resource regions that have been completed. Figure 2 A schematic block diagram of the structure of a computer device provided in an embodiment of the present disclosure. The computer device may be a terminal.

[0077] like Figure 2 As shown, the computer device includes a processor, a memory, and a network interface connected via a system bus, wherein the memory may include a non-volatile storage medium and an internal memory.

[0078] The non-volatile storage medium can store an operating system and a computer program. The computer program includes program instructions, and when the program instructions are executed, the processor can execute any method for checking the natural resource assets owned by the whole people based on the real-life three-dimensional model.

[0079] The processor is used to provide computing and control capabilities and support the operation of the entire computer equipment.

[0080] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor, the processor can execute any method for inventorying natural resource assets owned by the whole people based on the real-life three-dimensional model.

[0081] The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art will appreciate that Figure 2 The structure shown in the figure is only a block diagram of a part of the structure related to the scheme of the present disclosure, and does not constitute a limitation on the computer device to which the scheme of the present disclosure is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0082] It should be understood that the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0083] In one embodiment, the processor is used to run a computer program stored in a memory to implement the following steps: performing resource region segmentation on a regional remote sensing image of a natural resource asset inventory area to obtain a resource region segmentation image set, wherein the natural resource asset inventory area has corresponding resource region geographic reference information; for each resource region segmentation image in the resource region segmentation image set, performing the following identification steps: generating resource region segmentation information corresponding to the resource region segmentation image; matching corresponding natural resource asset region information from a pre-established natural resource asset database based on the resource region segmentation information and the resource region geographic reference information; performing real-scene three-dimensional image acquisition on the resource region corresponding to the resource region segmentation information to obtain a corresponding resource region real-scene three-dimensional image set, and constructing a resource region real-scene three-dimensional model based on the resource region real-scene three-dimensional image set; inputting the resource region real-scene three-dimensional model, the resource region segmentation information and the natural resource asset region information into a pre-trained natural resource asset inventory recognition model to obtain corresponding resource region asset inventory recognition information; and storing each resource region asset inventory recognition information in the corresponding natural resource asset inventory database owned by the whole people.

[0084] An embodiment of the present disclosure also provides a computer-readable storage medium, on which a computer program is stored. The computer program includes program instructions. The method implemented when the program instructions are executed can refer to the various embodiments of the method for inventorying natural resource assets owned by the whole people based on the real-life three-dimensional model of the present disclosure.

[0085] The computer-readable storage medium may be an internal storage unit of the computer device in the aforementioned embodiment, such as a hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a smart memory card (SmartMediaCard, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc., provided on the computer device.

[0086] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system 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 system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0087] The serial numbers of the embodiments of the present disclosure are only for description and do not represent the advantages and disadvantages of the embodiments. The above description is only a specific implementation mode of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present disclosure, and these modifications or replacements should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure shall be based on the protection scope of the claims.

Claims

1. A method for checking all natural resources assets owned by the whole people based on a real-life three-dimensional model, characterized in that: include: Perform resource region segmentation on the regional remote sensing image of the area to be investigated for natural resource assets, and obtain a resource region segmentation image set, wherein the area to be investigated for natural resource assets has corresponding resource region geographic reference information; For each resource region segmentation image in the resource region segmentation image set, the following identification steps are performed: Generating resource region segmentation information corresponding to the resource region segmentation image; According to the resource area segmentation information and the resource area geographic reference information, matching corresponding natural resource asset area information from a pre-established natural resource asset database; Performing real-scene 3D image acquisition on the resource area corresponding to the resource area segmentation information to obtain a corresponding real-scene 3D image set of the resource area, and constructing a real-scene 3D model of the resource area according to the real-scene 3D image set of the resource area; Inputting the resource area real-scene three-dimensional model, the resource area segmentation information and the natural resource asset area information into a pre-trained natural resource asset inventory identification model to obtain corresponding resource area asset inventory identification information; The asset inventory identification information of each resource area is stored in the corresponding public natural resource asset inventory database.

2. The method for checking the natural resources assets owned by all people based on the real-life three-dimensional model according to claim 1 is characterized in that: The generating of resource region segmentation information corresponding to the resource region segmentation image includes: Generate a resource region partition image feature information set and global resource region segmentation image feature information corresponding to the resource region segmentation image; Generate a global resource region mask image based on the global resource region segmentation image feature information and using a pre-trained resource region segmentation model; Performing image division on the global resource region mask image according to the division mode corresponding to the resource region division image set to obtain the resource region mask division image set; For each resource region partition image feature information in the resource region partition image feature information set, generating a local resource region mask image corresponding to the resource region partition image feature information through the resource region segmentation model according to the resource region partition image feature information and the corresponding resource region mask partition image; The resource region segmentation information corresponding to the resource region segmentation image is generated according to each local resource region mask image.

3. The method for checking the natural resource assets owned by all people based on the real-life three-dimensional model according to claim 2 is characterized in that: The method of constructing a real-scene three-dimensional model of the resource area according to the real-scene three-dimensional image set of the resource area includes: Generate an initial resource region real-scene three-dimensional model corresponding to the resource region segmentation image according to the resource region segmentation image feature information set and the global resource region segmentation image feature information; According to the resource area real-scene 3D image set, the initial resource area real-scene 3D model is filled with resource area real-scene 3D images to obtain a filled initial resource area real-scene 3D model as the resource area real-scene 3D model.

4. The method for checking the natural resource assets owned by all people based on the real-life three-dimensional model according to claim 2 is characterized in that: The method of generating a local resource region mask image corresponding to the resource region partition image feature information according to the resource region partition image feature information and the corresponding resource region mask partition image through the resource region segmentation model includes: Generate image marking information corresponding to the resource region mask partition image; Inputting the resource region mask partition image into a vector model to obtain resource region mask image feature information; The resource region mask image feature information, the resource region partition image feature information and the image tag information are input into the resource region segmentation model to obtain a corresponding local resource region mask image.

5. The method for checking the natural resources assets owned by all people based on the real-life three-dimensional model according to claim 2 is characterized in that: The generating of the resource region partition image feature information set and the global resource region partition image feature information corresponding to the resource region segmentation image comprises: Determining the resolution of the mask image output by the resource region segmentation model as the model resolution; According to the resolution corresponding to the resource region segmentation image and the model resolution, the resource region segmentation image is divided to obtain a resource region segmentation image set, wherein the resolution corresponding to each resource region segmentation image is the same as the model resolution; Inputting each resource area division image in the resource area division image set into a vector model to obtain a resource area division image feature information set; The resource region segmentation image is input into the vector model to obtain corresponding global resource region segmentation image feature information.

6. The method for checking the natural resources assets owned by all people based on the real-life three-dimensional model according to claim 2 is characterized in that: The resource region segmentation is performed on the regional remote sensing image of the area to be natural resource asset inventory to obtain a resource region segmentation image set, including: Performing correction processing on the regional remote sensing image to obtain a corrected regional remote sensing image; The correction area remote sensing image is input into a pre-trained resource area segmentation model to obtain a resource area segmentation image set.

7. The method for checking the natural resources assets owned by all people based on the real-life three-dimensional model according to claim 6 is characterized in that: Before inputting the correction area remote sensing image into a pre-trained resource area segmentation model to obtain a resource area segmentation image set, the method further includes: Acquire remote sensing images of the sample area; Determine a network structure of an initial resource region segmentation model, wherein the initial resource region segmentation model includes: a convolutional network, a pooling network, and an upsampling layer; Inputting the sample area remote sensing image into the initial resource area segmentation model to obtain an initial resource area segmentation image set; Determine a difference value between the initial resource region segmentation image set and the corresponding sample segmentation label; In response to determining that the difference value is less than or equal to a preset threshold, the initial resource region segmentation model is determined as a trained resource region segmentation model.

8. A computer device, characterized in that: The computer device comprises a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images