High-precision building information extraction and analysis system based on drone

Abstract

The present invention relates to the technical field of building data extraction, and specifically discloses a high-precision building information extraction and analysis system based on an unmanned aerial vehicle. The system sets a model generation accuracy determination module, a category recognition module, an information extraction accuracy module and an extraction state feedback module. First, the model generation accuracy of a three-dimensional aerial survey model of a target building is determined, the category of the target building is identified, and an information extraction method of the corresponding category is obtained. Information is extracted from the target building, and measured data of the target building is collected at the same time. The information extraction accuracy of the target building is comprehensively determined, and the information is verified with a predefined information extraction accuracy compliance value. The collected data can be grasped and processed in near real time, and the data processing efficiency is improved. Feedback is provided to the high-precision building information extraction and analysis of the unmanned aerial vehicle, and the monitoring efficiency and accuracy of the unmanned aerial vehicle on the building are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of building data extraction, and in particular to a high-precision building information extraction and analysis system based on unmanned aerial vehicles. Background Art

[0002] With the remarkable progress made in UAV technology in recent years, a solid foundation has been laid for the application of UAV in the field of building information extraction. The current technical solutions for building information extraction are aimed at the traditional building type survey and information acquisition in earthquake damage pre-assessment work, which has the disadvantages of large workload, slow progress, and large manpower consumption. At present, the height of buildings is mainly extracted by indirectly inferring the actual height of buildings through shadow information, which will lead to calculation errors. In addition, the spatial resolution of civil remote sensing images is low, which cannot meet the requirements of fine extraction of building information, and it is difficult to make an initial judgment on the type of buildings.

[0003] For example, the invention patent with announcement number CN104484668B provides a method for extracting building contours from multiple overlapping remote sensing images of unmanned aerial vehicles, including generating a three-dimensional point cloud using an aerial triangulation combined with a dense matching method, filtering the point cloud, and detecting buildings therefrom; after deleting the walls of the detected buildings, extracting the rough contour of the building from the top surface information of the building; superimposing the rough contour of the building on the spliced ​​image as a buffer zone, using the rough contour of the building as shape prior information, evolving it within the buffer zone using a level set algorithm, and finally obtaining the precise contour of the building; the above application makes full use of the three-dimensional information of the point cloud generated by multiple overlapping images, and combines it with the high-precision geometric information of high-resolution remote sensing images, which not only significantly improves the accuracy of building contour extraction, but also reduces the complexity of the method.

[0004] Combining the above technical solutions, it is found that with the application of drones in the field of building information extraction and analysis, it is necessary to organize more accurate, efficient and convenient building information extraction technical solutions. However, most building information extraction technical solutions only consider a single factor affecting the extraction of building information, which will reduce the accuracy of building information extraction and affect the subsequent analysis of building information. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a high-precision building information extraction and analysis system based on drones, which can effectively solve the problems involved in the above-mentioned background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a high-precision building information extraction and analysis system based on an unmanned aerial vehicle, including a model generation accuracy module, which is used to perform aerial survey on a target building through an unmanned aerial vehicle to obtain a three-dimensional aerial survey model of the target building, obtain the characteristic data of the three-dimensional aerial survey model of the target building, determine the model generation accuracy of the three-dimensional aerial survey model of the target building, and compare it with a predefined model generation accuracy threshold. If the model generation accuracy of the three-dimensional aerial survey model of the target building is greater than or equal to the model generation accuracy threshold, then enter the category recognition module; the category recognition module is used to perform category recognition on the three-dimensional aerial survey model of the target building through the model recognition mechanism of the unmanned aerial vehicle, and identify the category of the target building, so as to match the information extraction method corresponding to the category of the target building; the information extraction accuracy module is used to extract information from the target building according to the information extraction method corresponding to the category of the target building, obtain the extraction data of the target building, and collect the measured data of the target building at the same time, and comprehensively determine the information extraction accuracy of the target building; the extraction state feedback module is used to verify the information extraction accuracy of the target building with the predefined information extraction accuracy compliance value, so as to provide feedback on the high-precision building information extraction and analysis of the unmanned aerial vehicle.

[0007] As a further solution, the model generation accuracy of the three-dimensional aerial survey model of the target building is specifically analyzed in the following process: the total edge length of the three-dimensional aerial survey model of the target building is processed with the predefined total edge reference length to obtain the contour integrity of the three-dimensional aerial survey model of the target building; the spatial resolution of each pixel of the three-dimensional aerial survey model of the target building is processed with the standard deviation to obtain the spatial resolution discreteness of the three-dimensional aerial survey model of the target building; the time resolution of the drone at each data acquisition time point is processed with the standard deviation to obtain the time resolution discreteness of the drone within the data acquisition cycle; the total edge reference length, spatial resolution reference discreteness and time resolution reference discreteness are extracted from the building information database; the contour integrity of the three-dimensional aerial survey model of the target building, the spatial resolution discreteness of the three-dimensional aerial survey model of the target building and the time resolution discreteness of the drone within the data acquisition cycle are comprehensively analyzed to obtain the model generation accuracy of the three-dimensional aerial survey model of the target building, and the specific expression is:

[0008] ;

[0009] In the formula The accuracy of the model generation of the 3D aerial survey model of the target building, is the outline completeness of the 3D aerial survey model of the target building, is the spatial resolution discreteness of the 3D aerial survey model of the target building, For spatial resolution, refer to discreteness. is the time resolution discreteness of the UAV during the data acquisition cycle, For time resolution, refer to discreteness. is the impact factor corresponding to the unit value of contour integrity predefined in the building information database, is the correction factor corresponding to the spatial resolution discreteness predefined in the building information database, is the correction factor corresponding to the time resolution discreteness predefined in the building information database, is a natural constant.

[0010] As a further solution, the comparison with a predefined model generation accuracy threshold is specifically analyzed as follows: the model generation accuracy of the three-dimensional aerial survey model of the target building is compared with the predefined model generation accuracy threshold; if the model generation accuracy of the three-dimensional aerial survey model of the target building is greater than or equal to the model generation accuracy threshold, the three-dimensional aerial survey model of the target building is classified; if the model generation accuracy of the three-dimensional aerial survey model of the target building is less than the model generation accuracy threshold, the target building is aerially surveyed again until the model generation accuracy of the drone is greater than or equal to the model generation accuracy threshold.

[0011] As a further solution, the information extraction method corresponding to the category to which the target building belongs is matched. The specific matching process is: according to the category to which the target building belongs, the information extraction method corresponding to each predefined category is matched, so as to match the information extraction method corresponding to the category to which the target building belongs.

[0012] As a further solution, the information extraction accuracy of the target building is determined by the following specific process: the roof height of the target building, the number of floors of the target building and the roof angle of the target building are respectively comprehensively analyzed with the measured roof height of the target building, the measured number of floors of the target building and the measured roof angle of the target building to obtain the information extraction accuracy of the target building.

[0013] As a further solution, the information extraction accuracy of the target building is determined by:

[0014] ;

[0015] In the formula is the accuracy of information extraction of the target building, is the roof height of the target building, is the measured roof height of the target building, is the number of floors of the target building, is the measured number of floors of the target building, is the roof angle of the target building, is the measured roof angle of the target building, is the weight factor corresponding to the roof height predefined in the building information database, is the weight factor corresponding to the number of floors of the building predefined in the building information database, It is the weight factor corresponding to the roof angle predefined in the building information library.

[0016] As a further solution, the high-precision building information extraction and analysis of the drone is fed back, and the specific feedback process is: the information extraction accuracy of the target building is verified with the predefined information extraction accuracy and compliance value to generate an information extraction label. If the information extraction accuracy of the target building is greater than or equal to the information extraction accuracy and compliance value, the information extraction label is defined as an information extraction qualified label. If the information extraction accuracy of the target building is less than the information extraction accuracy and compliance value, the information extraction label is defined as an information extraction abnormal label, so as to provide feedback on the high-precision building information extraction and analysis of the drone.

[0017] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0018] (1) The present invention provides a high-precision building information extraction and analysis system based on an unmanned aerial vehicle. First, the model generation accuracy of the three-dimensional aerial survey model of the target building is determined. If the model generation accuracy is greater than or equal to the model generation accuracy threshold, it means that the quality of the three-dimensional aerial survey model of the target building generated by the unmanned aerial vehicle is reliable. Then, the category of the target building is identified, and the information extraction method of the corresponding category is obtained. Information of the target building is extracted, and the measured data of the target building is collected at the same time. The information extraction accuracy of the target building is comprehensively determined and verified with the predefined information extraction accuracy compliance value. The collected data can be grasped and processed in near real time, thereby improving data processing efficiency. Feedback is provided to the high-precision building information extraction and analysis of the unmanned aerial vehicle, thereby improving the efficiency and accuracy of unmanned aerial vehicle monitoring of buildings.

[0019] (2) The present invention uses a three-dimensional aerial survey model of the target building and a model recognition mechanism to identify the category of the target building, thereby matching the information extraction method corresponding to the category of the target building. By determining the category of the building, the drone can extract and analyze the relevant information of the building in a more targeted manner. Accurate building category determination can reduce misjudgments and missed judgments in the information extraction process and improve the accuracy of drone information extraction.

[0020] (3) The present invention obtains the characteristic data of the three-dimensional aerial survey model of the target building and determines the model generation accuracy of the three-dimensional aerial survey model of the target building. It can discover and correct the errors in the generated three-dimensional aerial survey model, thereby obtaining a more accurate three-dimensional aerial survey model of the target building, improving the accuracy of the model generation of the drone, and providing a guarantee for the subsequent information extraction of the building. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0022] Figure 1 It is a schematic diagram of system module connection of the present invention.

[0023] Figure 2 This is a simplified schematic diagram of the building categories involved in the present invention.

[0024] Figure numerals: 1. Flat roof building; 2. Sloped roof building; 3. Triangular roof building; 4. Roof height; 5. Roof angle. DETAILED DESCRIPTION

[0025] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0026] See also Figure 1 As shown, an embodiment of the present invention provides a technical solution: a high-precision building information extraction and analysis system based on a drone, including a model generation accuracy module, a category recognition module, an information extraction accuracy module and an extraction status feedback module.

[0027] An embodiment of the present invention provides a technical solution: a high-precision building information extraction and analysis system based on an unmanned aerial vehicle, which also includes a building information database, which is used to store the total length of edge reference, spatial resolution reference discreteness, time resolution reference discreteness and initial preset values ​​of various factors.

[0028] The model generation accuracy module is connected to the category recognition module, the category recognition module is connected to the information extraction accuracy module, the information extraction accuracy module is connected to the extraction state feedback module, and the model generation accuracy module, the category recognition module, the information extraction accuracy module and the extraction state feedback module are all connected to the building information library.

[0029] The model generation accuracy module is used to perform aerial survey of the target building through a drone, obtain a three-dimensional aerial survey model of the target building, obtain characteristic data of the three-dimensional aerial survey model of the target building, determine the model generation accuracy of the three-dimensional aerial survey model of the target building, and compare it with a predefined model generation accuracy threshold. If the model generation accuracy of the three-dimensional aerial survey model of the target building is greater than or equal to the model generation accuracy threshold, enter the category recognition module.

[0030] Specifically, the characteristic data of the three-dimensional aerial survey model of the target building include the total edge length of the three-dimensional aerial survey model of the target building, the spatial resolution of each pixel of the three-dimensional aerial survey model of the target building, and the temporal resolution of the drone at each data acquisition time point.

[0031] The above-mentioned total edge length, spatial resolution and temporal resolution can all be extracted from the operation report generated by the UAV aerial survey terminal.

[0032] Furthermore, the model generation accuracy of the three-dimensional aerial survey model of the target building is analyzed in the following specific process:

[0033] The total edge length of the three-dimensional aerial survey model of the target building is processed with the predefined total edge reference length to obtain the contour integrity of the three-dimensional aerial survey model of the target building.

[0034] The total edge length of the above-mentioned target building three-dimensional aerial survey model refers to the sum of all edge lengths of the generated target building three-dimensional aerial survey model, and the total edge reference length refers to the standard value of the total edge length of the target building three-dimensional aerial survey model; the specific expression of data processing is: ,in is the outline completeness of the 3D aerial survey model of the target building, is the total edge length of the 3D aerial survey model of the target building, is the total length of the edge reference, It is the correction factor corresponding to the outline completeness predefined in the building information library.

[0035] In a specific embodiment, the correction factor corresponding to the contour completeness can be directly obtained from the building information library during use, and the corresponding relationship can be a pre-set mapping relationship. For example, the contour completeness and the correction factor corresponding to the contour completeness preset in the building information library form a mapping set, and the real-time contour completeness is brought into the mapping set to obtain the correction factor corresponding to the contour completeness. The mapping relationship can be a one-to-one correspondence or a many-to-one relationship. In this embodiment, its value range is [0, 1].

[0036] In this embodiment, the correction factor corresponding to the contour completeness is set to 0.33.

[0037] In this embodiment, the contour integrity of the three-dimensional aerial survey model of the target building is shown in Table 1:

[0038] Table 1 Numerical changes between contour integrity and its parameters

[0039] ;

[0040] In this embodiment, according to the table of contour completeness of the above-mentioned three-dimensional aerial survey model of the target building, when the total edge length is equal to or close to the total edge reference length, the contour completeness is the largest. At this time, the model generation accuracy of the three-dimensional aerial survey model of the target building will also be higher. Therefore, if the integrity of the contour is to be guaranteed, necessary measures need to be taken when collecting the total edge length to ensure that the collected total edge length is closer to the actual length.

[0041] The spatial resolution of each pixel of the three-dimensional aerial survey model of the target building is processed by standard deviation to obtain the spatial resolution dispersion of the three-dimensional aerial survey model of the target building.

[0042] The above-mentioned pixels refer to the basic units that constitute the surface of the three-dimensional model. In three-dimensional imaging and remote sensing technology, pixels are the smallest identifiable units in the model, and they together constitute the details and structure of the entire model.

[0043] The specific process of standard deviation is as follows: the spatial resolution of each pixel of the 3D aerial survey model of the target building is recorded as , where i is the number of each pixel, , m is the total number of pixels.

[0044] Calculate the average: ,in It is the average value of the spatial resolution of the 3D aerial survey model of the target building.

[0045] Calculate the variance: ,in is the variance of the spatial resolution of the 3D aerial survey model of the target building.

[0046] Calculate the standard deviation: ,in is the standard deviation of the spatial resolution of the three-dimensional aerial survey model of the target building, and is marked as the spatial resolution dispersion of the three-dimensional aerial survey model of the target building.

[0047] The time resolution of the UAV at each data acquisition time point is processed by standard deviation to obtain the time resolution discreteness of the UAV within the data acquisition cycle.

[0048] The standard deviation processing process of the temporal resolution is as shown in the standard deviation processing process of the spatial resolution, and will not be described in detail in this embodiment.

[0049] The total edge reference length, spatial resolution reference discreteness and temporal resolution reference discreteness are extracted from the building information database.

[0050] The data stored in the above-mentioned building information database is fitted based on the historical extraction information of multiple buildings. For example, the total edge reference length is obtained by obtaining historical monitoring data of multiple buildings. The information extraction personnel determine the total edge length that best matches the actual total edge length under the monitoring data of different total edge lengths, and use it as the total edge reference length. In this way, the value in the building information database is obtained, and the value may change due to different extraction environments, actual needs, and extraction equipment status. This embodiment does not make any special limitations on this.

[0051] In this embodiment, the contour integrity of the three-dimensional aerial survey model of the target building, the spatial resolution discreteness of the three-dimensional aerial survey model of the target building, and the time resolution discreteness of the drone during the data acquisition cycle are comprehensively analyzed to determine the model generation accuracy of the three-dimensional aerial survey model of the target building. In this embodiment, a more accurate calculation method is used to obtain the model, and the specific expression is:

[0052] ;

[0053] In the formula The model generation accuracy of the three-dimensional aerial survey model of the target building refers to the degree of consistency between the model obtained by shooting with the high-precision camera carried by the drone, obtaining the image data of the target building, and generating a three-dimensional model with the actual building in terms of geometric shape, size, positional relationship, etc.

[0054] The outline completeness of the three-dimensional aerial survey model of the target building refers to the degree of closeness and completeness of the outline of the target building model constructed by UAV three-dimensional aerial survey technology with the actual outline of the original building.

[0055] The spatial resolution discreteness of the three-dimensional aerial survey model of the target building refers to the degree of resolution difference of the target building at different spatial positions in the three-dimensional model constructed by UAV aerial survey technology. This discreteness can reflect the uniformity and consistency of the model in terms of spatial detail expression.

[0056] The spatial resolution reference discreteness refers to the standard value of the set spatial resolution discreteness.

[0057] It is the time resolution discreteness of the UAV within the data acquisition cycle, which refers to the degree of discreteness between each time point when the UAV obtains model data at different time points within the data acquisition cycle.

[0058] The time resolution reference discreteness refers to the standard value of the set time resolution discreteness.

[0059] It is the impact factor corresponding to the unit value of outline completeness predefined in the building information database.

[0060] In a specific embodiment, the influence factor corresponding to the unit value of contour completeness can be directly obtained from the building information library during use, and the corresponding relationship can be a preset mapping relationship. For example, the contour completeness and the influence factor corresponding to the unit value of contour completeness preset in the building information library form a mapping set, and the real-time contour completeness is brought into the mapping set to obtain the influence factor corresponding to the unit value of contour completeness. The mapping relationship can be a one-to-one correspondence or a many-to-one relationship. In this embodiment, its value range is [0, 1].

[0061] It is the correction factor corresponding to the spatial resolution discreteness predefined in the building information database.

[0062] In a specific embodiment, the correction factor corresponding to the spatial resolution discreteness can be directly obtained from the building information library when in use, and the corresponding relationship can be a preset mapping relationship. For example, the spatial resolution discreteness and the correction factor corresponding to the spatial resolution discreteness preset in the building information library form a mapping set, and the real-time spatial resolution discreteness is brought into the mapping set to obtain the correction factor corresponding to the spatial resolution discreteness. The mapping relationship can be a one-to-one correspondence or a many-to-one relationship. In this embodiment, its value range is [0, 1].

[0063] is the correction factor corresponding to the time resolution discreteness predefined in the building information database, and e is a natural constant.

[0064] In a specific embodiment, the correction factor corresponding to the time resolution discreteness can be directly obtained from the building information library during use, and the corresponding relationship can be a preset mapping relationship. For example, the time resolution discreteness and the correction factor corresponding to the time resolution discreteness preset in the building information library form a mapping set, and the real-time time resolution discreteness is brought into the mapping set to obtain the correction factor corresponding to the time resolution discreteness. The mapping relationship can be a one-to-one correspondence or a many-to-one relationship. In this embodiment, its value range is [0, 1].

[0065] In this embodiment, a high-completeness model can accurately reflect the external shape and structural features of the building. The higher the contour completeness, the greater the model generation accuracy. A model with high contour completeness often requires a smaller spatial resolution discreteness that is closer to the spatial resolution reference discreteness to support it, so as to more accurately restore the contour and details of the building. This is because a higher spatial resolution discreteness that is closer to the spatial resolution reference discreteness can not only restore the detailed performance of the three-dimensional aerial survey model of the target building in different areas or directions, but also reflect the more refined geometric shape of the three-dimensional aerial survey model of the target building. If the spatial resolution reference discreteness is too high, the model may have blurred or missing contours due to discrete data points, thereby reducing the model generation accuracy of the model; if the time resolution discreteness is large, it may mean that the data has changed significantly during the construction of the three-dimensional aerial survey model of the building at different time points, resulting in reduced model generation accuracy.

[0066] Specifically, the comparison is made with the predefined model to generate an accurate threshold value, and the specific analysis process is as follows:

[0067] The model generation accuracy of the three-dimensional aerial survey model of the target building is compared with the model generation accuracy threshold predefined in the building information database. If the model generation accuracy of the three-dimensional aerial survey model of the target building is greater than or equal to the model generation accuracy threshold, it means that the generated three-dimensional aerial survey model can more accurately reflect the real form, size, structure, texture and other characteristics of the target building, and the generation quality is high and the accuracy is as expected. Then, the three-dimensional aerial survey model of the target building is classified. If the model generation accuracy of the three-dimensional aerial survey model of the target building is less than the model generation accuracy threshold, it means that the shape, size, proportion, etc. of the generated three-dimensional aerial survey model may be significantly different from the real building, and the model cannot truly reflect the required characteristics of the building. There may be problems such as missing details and inconsistent accuracy. Then, the target building is re-surveyed until the model generation accuracy of the drone is greater than or equal to the model generation accuracy threshold.

[0068] It should be explained that if the model generation accuracy of the three-dimensional aerial survey model of the target building is always less than the model generation accuracy threshold, high-precision drone acquisition equipment can be selected to ensure that the accuracy of the acquisition equipment meets the requirements and high-quality data can be obtained; at the same time, select an environment with good lighting conditions and no obstructions for data collection to ensure the integrity and accuracy of the data, so as to improve the model generation accuracy of the three-dimensional aerial survey model of the target building.

[0069] The category recognition module is used to perform category recognition on the three-dimensional aerial survey model of the target building through the model recognition mechanism of the drone, and identify the category to which the target building belongs, so as to match the information extraction method corresponding to the category to which the target building belongs.

[0070] The model recognition mechanism of the above-mentioned drone is a mechanism that automatically detects, identifies and classifies target buildings through a series of technical means and methods. This process mainly relies on advanced technologies such as computer vision, image processing, and deep learning. Its core lies in extracting distinctive features from drone images or videos, and performing efficient and accurate recognition based on these features.

[0071] Furthermore, the information extraction method corresponding to the category of the target building is matched, and the specific matching process is:

[0072] According to the category of the target building, it is matched with the information extraction methods corresponding to the predefined categories in the building information database. The specific matching process is: construct a mapping set between the category of the target building and its corresponding information extraction method, input the real-time target building type, match the specific category of the type, and obtain the information extraction method corresponding to the target building category through the mapping set, wherein the mapping relationship is a one-to-one relationship, so as to obtain the information extraction method corresponding to the category of the target building.

[0073] It should be explained that the categories of the above-mentioned target buildings include but are not limited to flat roof buildings, sloping roof buildings, and triangular roof buildings. In the embodiments of the present invention, only three categories of flat roof buildings, sloping roof buildings and triangular roof buildings are discussed. By limiting the scope of discussion to these three typical building types, in the face of a variety of building forms, the problem can be simplified and the implementation complexity of the embodiments of the present invention can be greatly reduced. Although only three types are discussed, these three building types are widely representative in practical applications, which improves the versatility and extensibility of the embodiments of the present invention.

[0074] The simplified schematic diagrams of the above-mentioned flat roof building, sloping roof building and triangular roof building are as follows: Figure 2 As shown, in the simplified schematic diagram of building categories, 1 is a flat roof building, 2 is a sloping roof building, 3 is a triangular roof building, 4 is the roof height, and 5 is the roof angle.

[0075] according to Figure 2 From the simplified schematic diagram of building categories, we can see that a flat roof building has no roof structure, that is, the roof height of a flat roof building is zero and the roof angle is zero. If this happens to the building to be judged, the building can be judged to be a flat roof building; the roof angle of a sloping roof building directly affects its roof height. When the roof angle increases within a reasonable range, its roof height also increases accordingly; the roof height of a triangular roof building is lower and the roof angle is smaller.

[0076] The information extraction methods corresponding to the categories of the above target buildings are:

[0077] The information extraction method corresponding to the flat roof building is to obtain parameters, including the contour perimeter, area and surface coefficient of the target building, so as to evaluate the accuracy of the information extraction method corresponding to the target building in the information extraction process.

[0078] The information extraction method corresponding to the sloping roof building is to obtain parameters, including the roof angle of the target building, the building height and the reflectivity of the sloping roof, so as to evaluate the information extraction method corresponding to the target building and whether it is accurate in the information extraction process.

[0079] The information extraction method corresponding to the triangular roof building is to obtain parameters, including the roof height of the target building, the number of floors of the target building and the roof angle of the target building, so as to evaluate the accuracy of the information extraction method corresponding to the target building during the information extraction process. In the embodiment of the present invention, the category of the target building is set as a triangular roof building.

[0080] The information extraction accuracy module is used to extract information from the target building according to the information extraction method corresponding to the category of the target building, obtain the extracted data of the target building, and collect the measured data of the target building to comprehensively determine the information extraction accuracy of the target building.

[0081] Specifically, the extracted data of the target building includes the roof height of the target building, the number of floors of the target building, and the roof angle of the target building.

[0082] The above-mentioned roof height, number of floors and roof angle can all be extracted from the operating terminal of the drone.

[0083] Furthermore, the measured data of the target building specifically includes the measured roof height of the target building, the measured number of floors of the target building, and the measured roof angle of the target building.

[0084] The above-mentioned measured roof height, measured number of floors and measured roof angle can all be obtained by on-site measurements by information extraction personnel.

[0085] Specifically, the information extraction accuracy of the target building is determined by the following process:

[0086] In this embodiment, the roof height of the target building, the number of floors of the target building, and the roof angle of the target building are comprehensively analyzed with the measured roof height of the target building, the measured number of floors of the target building, and the measured roof angle of the target building, so as to determine the accuracy of information extraction of the target building. In this embodiment, a more accurate calculation method is used to obtain the information, and the specific expression is:

[0087] ;

[0088] In the formula The accuracy of information extraction of the target building is used to quantify the accuracy of the information extraction method used by the drone for the target building in this embodiment.

[0089] The roof height of the target building refers to the vertical distance from the bottom of the building's roof structure to the top of the building's roof structure that can be detected by the drone.

[0090] The measured roof height of the target building refers to the vertical distance from the bottom of the building's roof structure to the top of the building's roof structure obtained by on-site measurement by information extraction personnel.

[0091] The number of floors of the target building refers to the natural number of floors of the building that can be detected by the drone. It is the number of floors above the outdoor floor of the building divided by floor slabs and floor structures, and does not include the underground floors of the building.

[0092] The measured number of floors of the target building refers to the natural number of floors of the building obtained by on-site measurement by information extraction personnel.

[0093] The roof angle of the target building refers to the angle formed between the roof structure surface and the horizontal plane of the target building that can be detected by the drone.

[0094] The measured roof angle of the target building refers to the angle formed by the roof structure surface and the horizontal plane of the target building obtained by on-site measurement by the information extraction personnel.

[0095] The weight factor corresponding to the roof height predefined in the building information library.

[0096] In a specific embodiment, the weight factor corresponding to the roof height can be directly obtained from the building information database during use, and the corresponding relationship can be a pre-set mapping relationship. For example, the roof height and the weight factor corresponding to the roof height preset in the building information database form a mapping set, and the real-time roof height is brought into the mapping set to obtain the weight factor corresponding to the roof height. The mapping relationship can be a one-to-one correspondence or a many-to-one relationship. In this embodiment, its value range is [0, 1].

[0097] The weight factor corresponding to the number of floors of a building predefined in the building information library.

[0098] In a specific embodiment, the weight factor corresponding to the number of floors of the building can be directly obtained from the building information library during use, and the corresponding relationship can be a pre-set mapping relationship. For example, the number of floors of the building and the weight factors corresponding to the number of floors of the building preset in the building information library form a mapping set, and the real-time number of floors of the building is brought into the mapping set to obtain the weight factor corresponding to the number of floors of the building. The mapping relationship can be a one-to-one correspondence or a many-to-one relationship. In this embodiment, its value range is [0, 1].

[0099] It is the weight factor corresponding to the roof angle predefined in the building information library.

[0100] In a specific embodiment, the weight factor corresponding to the roof angle can be directly obtained from the building information library during use, and the corresponding relationship can be a preset mapping relationship. For example, the roof angle and the weight factor corresponding to the roof angle preset in the building information library form a mapping set, and the real-time roof angle is brought into the mapping set to obtain the weight factor corresponding to the roof angle. The mapping relationship can be a one-to-one correspondence or a many-to-one relationship. In this embodiment, its value range is [0, 1].

[0101] In this embodiment, according to the category to which the target building belongs, parameters are obtained, including the roof height of the target building, the number of floors of the target building and the roof angle of the target building, so as to evaluate the accuracy of the information extraction method corresponding to the target building. If the differences between the roof height of the target building, the number of floors of the target building and the roof angle of the target building and the actually measured roof height, the actually measured number of floors of the target building and the actually measured roof angle of the target building are large, it means that there are significant errors and uncertainties in the process of using the information extraction method, which results in the extracted building information being inaccurate.

[0102] The extraction status feedback module is used to verify the information extraction accuracy of the target building with the predefined information extraction accuracy compliance value, so as to provide feedback on the high-precision building information extraction and analysis of the drone.

[0103] Specifically, the feedback of the extraction and analysis of high-precision building information by the drone is performed as follows:

[0104] The information extraction accuracy of the target building is verified with the predefined information extraction accuracy compliance value to generate an information extraction label. If the information extraction accuracy of the target building is greater than or equal to the information extraction accuracy compliance value, it means that the information extraction method used can effectively identify and extract key information such as roof height, number of floors, roof angle, etc. when processing the target building data, and the accuracy of this information meets the predetermined compliance standard, then the information extraction label is defined as an information extraction qualified label; if the information extraction accuracy of the target building is less than the information extraction accuracy compliance value, it means that there are significant errors in the information extraction process, resulting in the inability to accurately identify or extract key information, then the information extraction label is defined as an information extraction abnormality label, so as to provide feedback on the high-precision building information extraction and analysis of the drone.

[0105] The above-mentioned generated information extraction label refers to a label obtained after verifying the information extraction accuracy of the target building with the predefined information extraction accuracy compliance value. This label can help information extraction personnel understand whether the information extraction accuracy of the target building has reached the preset standard, so as to determine whether the information extraction process needs to be adjusted.

[0106] The above feedback process is specifically as follows: if the above verification process generates an information extraction anomaly label, the information extraction personnel will be reminded of the information extraction anomaly through SMS, phone calls, emails, etc., and an anomaly identification report will be generated so that the information extraction personnel can take necessary measures to resolve the anomaly.

[0107] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the protection scope of the present invention.

Claims

1. A high-precision building information extraction and analysis system based on drones, characterized by: include: The model generation accuracy module is used to perform aerial survey of the target building by using a drone to obtain a three-dimensional aerial survey model of the target building, obtain the characteristic data of the three-dimensional aerial survey model of the target building, determine the model generation accuracy of the three-dimensional aerial survey model of the target building, and compare it with a predefined model generation accuracy threshold. If the model generation accuracy of the three-dimensional aerial survey model of the target building is greater than or equal to the model generation accuracy threshold, enter the category recognition module; The category recognition module is used to classify the three-dimensional aerial survey model of the target building through the model recognition mechanism of the drone, and identify the category of the target building, so as to match the information extraction method corresponding to the category of the target building; The information extraction accuracy module is used to extract information from the target building according to the information extraction method corresponding to the category of the target building, obtain the extracted data of the target building, and collect the measured data of the target building to comprehensively determine the information extraction accuracy of the target building; The extraction status feedback module is used to verify the accuracy of information extraction of the target building with the predefined information extraction accuracy compliance value, so as to provide feedback on the high-precision building information extraction and analysis of the drone.

2. The high-precision building information extraction and analysis system based on drone according to claim 1 is characterized by: The characteristic data of the three-dimensional aerial survey model of the target building specifically include the total edge length of the three-dimensional aerial survey model of the target building, the spatial resolution of each pixel of the three-dimensional aerial survey model of the target building, and the time resolution of the drone at each data acquisition time point.

3. The high-precision building information extraction and analysis system based on drone according to claim 2 is characterized by: The model generation accuracy of the three-dimensional aerial survey model of the target building is analyzed in the following specific process: The total edge length of the three-dimensional aerial survey model of the target building is processed with the predefined total edge reference length to obtain the contour integrity of the three-dimensional aerial survey model of the target building; The spatial resolution of each pixel of the three-dimensional aerial survey model of the target building is processed by standard deviation to obtain the spatial resolution dispersion of the three-dimensional aerial survey model of the target building; The time resolution of the UAV at each data acquisition time point is processed by standard deviation to obtain the time resolution dispersion of the UAV within the data acquisition cycle; Extracting the total edge reference length, spatial resolution reference dispersion and time resolution reference dispersion from the building information database; The model generation accuracy of the target building 3D aerial survey model is obtained by comprehensively analyzing the contour integrity of the target building 3D aerial survey model, the spatial resolution discreteness of the target building 3D aerial survey model, and the time resolution discreteness of the UAV during the data acquisition cycle. The specific expression is: ； In the formula The accuracy of the model generation of the 3D aerial survey model of the target building, is the outline completeness of the 3D aerial survey model of the target building, is the spatial resolution discreteness of the 3D aerial survey model of the target building, For spatial resolution, refer to discreteness. is the time resolution discreteness of the UAV during the data acquisition cycle, For time resolution, refer to discreteness. is the impact factor corresponding to the unit value of contour integrity predefined in the building information database, is the correction factor corresponding to the spatial resolution discreteness predefined in the building information database, is the correction factor corresponding to the time resolution discreteness predefined in the building information database, is a natural constant.

4. The high-precision building information extraction and analysis system based on drone according to claim 1 is characterized by: The above is compared with the predefined model to generate an accurate threshold, and the specific analysis process is as follows: The model generation accuracy of the three-dimensional aerial survey model of the target building is compared with the predefined model generation accuracy threshold. If the model generation accuracy of the three-dimensional aerial survey model of the target building is greater than or equal to the model generation accuracy threshold, the three-dimensional aerial survey model of the target building is classified; if the model generation accuracy of the three-dimensional aerial survey model of the target building is less than the model generation accuracy threshold, the target building is re-aerial surveyed until the model generation accuracy of the UAV is greater than or equal to the model generation accuracy threshold.

5. The high-precision building information extraction and analysis system based on drone according to claim 1 is characterized by: The information extraction method corresponding to the category of the target building is matched, and the specific matching process is as follows: According to the category to which the target building belongs, the information extraction method corresponding to each predefined category is matched, so as to match the information extraction method corresponding to the category to which the target building belongs.

6. The high-precision building information extraction and analysis system based on drone according to claim 1 is characterized by: The extracted data of the target building specifically includes the roof height of the target building, the number of floors of the target building, and the roof angle of the target building.

7. The high-precision building information extraction and analysis system based on drone according to claim 1 is characterized by: The measured data of the target building specifically includes the measured roof height of the target building, the measured number of floors of the target building, and the measured roof angle of the target building.

8. The high-precision building information extraction and analysis system based on drone according to claim 6 is characterized by: The specific determination process of the information extraction accuracy of the target building is as follows: The roof height of the target building, the number of floors of the target building and the roof angle of the target building are comprehensively analyzed with the measured roof height of the target building, the measured number of floors of the target building and the measured roof angle of the target building to obtain the information extraction accuracy of the target building.

9. The high-precision building information extraction and analysis system based on drone according to claim 8, characterized in that: The information extraction accuracy of the target building is determined by: ； In the formula is the accuracy of information extraction of the target building, is the roof height of the target building, is the measured roof height of the target building, is the number of floors of the target building, is the measured number of floors of the target building, is the roof angle of the target building, is the measured roof angle of the target building, is the weight factor corresponding to the roof height predefined in the building information database, is the weight factor corresponding to the number of floors of the building predefined in the building information database, It is the weight factor corresponding to the roof angle predefined in the building information library.

10. The high-precision building information extraction and analysis system based on drone according to claim 1, characterized in that: The feedback of the extraction and analysis of high-precision building information from drones is as follows: The information extraction accuracy of the target building is verified with the predefined information extraction accuracy compliance value to generate an information extraction label. If the information extraction accuracy of the target building is greater than or equal to the information extraction accuracy compliance value, the information extraction label is defined as an information extraction qualified label. If the information extraction accuracy of the target building is less than the information extraction accuracy compliance value, the information extraction label is defined as an information extraction abnormal label, so as to provide feedback on the high-precision building information extraction and analysis of the drone.

Images