Rural landscape local feature acquisition method and system based on radar and spectrum fusion

Abstract

The invention discloses a spectral image acquisition method and system for rural landscape local features, and belongs to the technical field of image acquisition and processing, and the method comprises the steps: S1, carrying a spectrum acquisition device and a distance measurement device to fly at a preset fixed height by means of an unmanned aerial vehicle based on a predefined route, acquiring a spectral data set A: # imgabs0 # and a height data set B: # imgabs1 # of a ground object; s2, performing data matching processing to obtain a spectral data set A'with elevation characteristics; s3, obtaining a spectral data set A ''with geographic position and elevation information from the spectral data set A'' with elevation characteristics in the step S2; and S4, generating a rural landscape observation model with accurate spatial resolution and three-dimensional characteristics, the rural landscape three-dimensional model with multi-dimensional information is formed by fusing ground object height data acquisition with the assistance of corresponding geographic position information, and compared with traditional two-dimensional image information, the rural landscape three-dimensional model is more visual and has higher analysis value.

Description

Technical Field

[0001] The present invention belongs to the technical field of image acquisition and processing, and particularly relates to a method for collecting local features of rural landscapes based on radar and spectrum fusion and a collection system using the above method. Background Art

[0002] Ecosystem services are the link between human well-being and the natural environment. As a spatial settlement at the landscape scale, the countryside is a rural organic body and its residential and industrial synergy formed on the basis of various rural ecosystems such as mountains, waters, forests, farmlands, lakes, grasslands, and sands. Rural landscape ecosystem services can play multiple roles in the process of urban-rural integration development, such as improving the well-being of rural residents, improving the ecological environment, stabilizing the social economy, and inheriting local culture.

[0003] At present, the research on local features of rural landscapes has become a hot topic. One direction is the analysis of the correlation characteristics between rural settlement landscapes and the spatial layout of natural elements, exploring planning methods for the protection, activation, and improvement of rural local landscapes under the background of new urbanization, and providing references for the construction and development of rural areas under the national territorial space planning system. At present, the spatial information of rural landscapes is mostly collected from existing map databases, and only planar image information can be obtained, which is difficult to form a three-dimensional spatial layout display. Especially for the study of a specific rural area, the above information has a large error and is not conducive to carrying out targeted research. The prior art also discloses the use of drones to collect the above planar images, but the problem of how to achieve a three-dimensional spatial layout display has not been solved yet. Therefore, it is necessary to improve the existing image collection method. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for collecting local features of rural landscapes based on radar and spectrum fusion to solve the above-mentioned defects caused by the prior art.

[0005] The described method for collecting local features of rural landscapes based on radar and spectrum fusion includes the following steps: S1. Based on a predefined flight path, a drone is used to carry a spectrum acquisition device and a ranging device to fly at a preset fixed height to obtain a spectrum data set A : and a height data set of ground objects B : , and store them in a data memory; S2. Match the spectrum data set A obtained in step S1 with the height data set B according to the shooting time point T to perform data matching processing to obtain a spectrum data set with elevation characteristics A' , and store it in a data memory; S3. Combine the positioning system built into the drone to attach the corresponding geographical location information to each collection point, so that the spectral data set already having elevation characteristics in step S2 A' Obtain a spectral data set with geographical location and elevation information A'' , and store it in the data storage; S4. Perform assignment and modeling on the spectral data set A'' to generate a rural landscape observation model with accurate spatial resolution and three-dimensional characteristics.

[0006] Furthermore, the spectral data set A : The matching method with the height data set B : is as follows: ; wherein, is the height correction coefficient of the drone, is the ranging correction coefficient caused by the time delay of the ranging device.

[0007] Furthermore, the marking strategy of the geographical location in step S3 is as follows: S31. Mark the position of the drone at a certain time point to the spectral data A'n of that time point by means of the marking model to complete the preliminary calibration of the geographical location. The marking method is as follows: L(A'n)= LnA'n ; wherein, L(A'n) is the spectral data A'n spectral data after completing the preliminary calibration, is the time correction coefficient, Ln is the position information; S32. Then perform smoothing processing on the data preliminarily calibrated in step S31 to obtain a spectral data set A'' with a certain smoothness.

[0008] Furthermore, in the preset flight path, the drone adopts a serpentine movement, and the distance between two adjacent parallel flight paths is not greater than 10 m.

[0009] Furthermore, the specific steps of performing assignment and modeling on the spectral data set A'' in step S4 are as follows: S41. With the help of the data processor, for the image with height data in the spectral data set A'' , perform equivalent stretching in the vertical direction according to its height data to form a three-dimensional model with elevation characteristics; S42. Attach the geographical location information to the above spectral data set with a three-dimensional model and integrate it into the three-dimensional map model library to form a three-dimensional model with both elevation characteristics and geographical location characteristics; A'' S43. Project the three-dimensional model with spectral image characteristics, elevation characteristics, and geographical location characteristics into the display cabinet by means of three-dimensional projection technology to form a holographic projection of the rural landscape three-dimensional model. S43. Project the three-dimensional model with spectral image characteristics, elevation characteristics, and geographical location characteristics into the display cabinet by means of three-dimensional projection technology to form a holographic projection of the rural landscape three-dimensional model.

[0010] Further, the flight altitude of the drone is set at 500 - 1000m.

[0011] Further, a temperature and humidity sensor can also be mounted on the drone to detect the temperature and humidity data at the location; The temperature and humidity data can be matched to the spectral data set A'' to form a rural landscape three-dimensional model with temperature and humidity characteristics.

[0012] Further, before generating the model in step S4, the following steps are also included: Classify the spectral data set in step S1 according to the element categories of buildings and natural landscapes by means of machine learning and image recognition technologies, and count the spatial layout frequencies of each element category A Classify the spectral data set in step S1 according to the element categories of buildings and natural landscapes by means of machine learning and image recognition technologies, and count the spatial layout frequencies of each element category , and then assign values and model the spectral data set using the spatial layout frequencies A'' ; Before classification, first divide the spectral data set A into n sampling areas by grid division, and then classify these n sampling areas; The calculation method of the spatial layout frequency is as follows: where is the spatial layout frequency of the l th element category, k is the total number of sampling areas, is the i th sampling area, and l is the proportion of the image area of the l th element category in the

[0013] The element categories of buildings and natural landscapes specifically include: buildings, mountains, waters, fields, forests, lakes, grasslands, and sands. A ranging device for collecting the elevation information of ground objects; A data memory for storing spectral data sets, elevation information, and geographical location information; A drone for carrying the above-mentioned spectral acquisition device, ranging device, and data memory.

[0014] A computer device includes a data processor that can assign the elevation information and geographical location information stored in the data memory to the spectral data set according to the method described above to form a three-dimensional rural landscape stereo model with three-dimensional features.

[0015] The present invention has the following advantages: 1. Based on traditional spectral data acquisition, the present invention proposes a new method, that is, integrating the acquisition of ground object height data and supplemented by corresponding geographical location information to form a three-dimensional rural landscape stereo model with multi-dimensional information. Compared with traditional two-dimensional image information, it is more intuitive and has greater analysis value.

[0016] 2. Through the corresponding matching method and marking model, the present invention realizes the organic combination of spectral information, elevation information, and geographical location information, providing a basis for forming a relatively accurate three-dimensional rural landscape stereo model. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the acquisition method of the present invention.

[0018] Figure 2 It is a schematic diagram of the structure of the acquisition system of the present invention.

[0019] Figure 3 It is a schematic diagram of the data processing working principle of the present invention Among them, 100 - drone, 200 - spectral acquisition device, 300 - ranging device. DETAILED DESCRIPTION OF THE INVENTION

[0020] The following is a further detailed description of the specific implementation of the present invention by describing the embodiments with reference to the accompanying drawings, so as to help those skilled in the art have a more complete, accurate, and in-depth understanding of the inventive concept and technical solution of the present invention.

[0021] As Figure 2 shown, the present invention provides an acquisition system using the above method, including: A spectral acquisition device 200 for acquiring spectral data on the ground; A ranging device 300 for acquiring elevation information of ground objects; A data memory for storing spectral data sets, elevation information, and geographical location information; A drone 100 for carrying the above-mentioned spectral acquisition device, ranging device, and data memory.

[0022] like Figure 1 As shown, in this embodiment, the present invention also provides a method for collecting local features of rural landscapes based on radar and spectrum fusion using the above-mentioned collection system, comprising the following steps: S1. Based on the predefined route, the drone 100 is used to carry the spectrum acquisition device 200 and the distance measurement device 300 to fly at a preset fixed altitude. In hilly areas, plains or areas without mountains nearby, the flight altitude of the drone 100 is set at 500-1000m. In mountainous areas, the flight altitude can be set at a height of 100m from the highest point. At the set altitude, the ground personnel control the drone 100 to fly and obtain the spectral data set. A : and ground object height dataset B : , stored in a data storage device, wherein the ranging device 300 may use a laser radar or a millimeter wave radar; S2, the spectral data set obtained in step S1 A With height dataset B Based on the time of shooting T Perform data matching processing to obtain a spectral data set with elevation characteristics A' , stored in the data storage device, because the spectrum acquisition device 200 and the distance measurement device 300 work simultaneously, both collect their own data at the same time point, and through matching processing, the elevation information of the ground object can be assigned to the spectrum data set A' In the process of creating a 3D model, it provides a basis for the later establishment of a 3D model. S3, combining the positioning system on the drone 100 to attach corresponding geographic location information to each collection point, so that the spectral data set with elevation features in step S2 A' Obtain spectral dataset with geographic location and elevation information A'' , and store it in a data storage device; The marking strategy is as follows: S31, marking the position of the drone 100 at a certain time point to the spectral data at the time point by means of a marking model A'n In the process of completing the preliminary calibration of the geographic location, the marking method is as follows: L(A'n)= LnA'n ; in, L(A'n) For spectral data A'n After completing the preliminary calibration of the spectral data, is the time correction factor, Ln For location information, the formed data set can be expressed as: L(A') : ; S32. Then, smooth the data preliminarily calibrated in step S31 to obtain a spectral data set with a certain degree of smoothness. A'' ; S4. Assign values and build a model for the spectral data set A'' to generate a rural landscape observation model with precise spatial resolution and three-dimensional features; The specific steps are as follows: S41. With the help of a data processor, for the image with height data in the spectral data set A'' , perform equivalent stretching vertically according to its height data to form a three-dimensional model with elevation features. During the stretching process, the color of the spectral image at the corresponding position is not adjusted. In this way, a relatively realistic display effect can be brought. S42. Attach geographical location information to the above spectral data set with a three-dimensional model A'' and integrate it into the three-dimensional map model library to form a three-dimensional model with both elevation features and geographical location features; S43. With the help of three-dimensional projection technology, project the three-dimensional model with spectral image features, elevation features, and geographical location features into the display cabinet to form a holographic projection of the rural landscape three-dimensional model.

[0023] In this embodiment, the spectral data set A : and the height data set B : The matching method is as follows: ; where is the height correction coefficient of the drone 100, and is the ranging correction coefficient caused by the time delay of the ranging device 300.

[0024] In the preset flight path, the drone 100 adopts a serpentine movement, and the distance between two adjacent parallel flight paths is not greater than 10 m. In this way, the accuracy of information collection can be ensured on the premise of shortening the collection time as much as possible.

[0025] As an improvement of the present invention, a temperature and humidity sensor can also be mounted on the drone 100 to detect the temperature and humidity data at the location; The temperature and humidity data can be matched to the spectral data set A'' to form a rural landscape three-dimensional model with temperature and humidity features. Combining with the elevation features and geographical location features obtained previously, a rural landscape three-dimensional model with multi-dimensional information is thus formed.

[0026] In this embodiment, before generating the model in step S4, the following steps are further included: With the aid of machine learning and image recognition technologies, the spectral data set in step S1 A is classified according to the element categories of buildings and natural landscapes, and the spatial layout frequencies of each element category are counted , and then the spatial layout frequencies are used to assign values and model the spectral data set A'' ; Before classification, the spectral data set A is divided into grids to form n sampling regions, and then these n sampling regions are classified; The calculation method of the spatial layout frequency is as follows: Wherein, is the spatial layout frequency of the l th element category, k is the total number of sampling regions, is the proportion of the image area of the i th sampling region in the l th element category, and the element categories of buildings and natural landscapes l specifically include: buildings, mountains, waters, fields, forests, lakes, grasslands and sands. In this way, a data set with spatial layout frequencies P can be obtained based on the spectral data. The spatial layout frequencies P are used to assign values and model the spectral data set A'' , so that a display with the spatial layout situation is formed at the top of the holographic projection of the three-dimensional model of the rural landscape. The presentation method can be a density distribution map (dot-shaped) in a plane state or a density distribution map (bubble-shaped) in a spatial state, and can be used to indirectly evaluate the rationality of the spatial layout.

[0027] The rationality of the above spatial layout is evaluated by the following calculation method: In the formula, n is the number of sampling regions in the acquisition area, is the cumulative percentage of the proportion of a certain element category in the sampling area in the entire sampling area ranked from large to small at the i th position. takes values from 0 to 1, the closer the

[0028] value is to 1, the more reasonable the spatial layout of the area is. Figure 3As shown, an embodiment of the present invention also discloses a computer device, including a data processor, which can assign the elevation information and geographical location information stored in the data memory to the spectral data set according to the described method to form a three-dimensional rural landscape stereo model.

[0029] The above are only the preferred embodiments of this application. It should be noted that for those of ordinary skill in the art of this technology, without departing from the principle of this application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of this application.

[0030] In the above embodiments of this application, the descriptions of each embodiment have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0031] The above are only the preferred embodiments of this application. It should be noted that for those of ordinary skill in the art of this technology, without departing from the principle of this application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of this application.

Claims

1. A method for collecting local features of rural landscapes based on radar and spectral fusion, characterized in that: The following steps are involved: S1. Based on the predefined route, the spectral data set is obtained by flying the drone with the spectral acquisition equipment and the ranging equipment at a preset fixed altitude. A : and ground object height dataset B : , stored in a data storage device; S2, the spectral data set obtained in step S1 A With height dataset B Based on the time of shooting T Perform data matching processing to obtain a spectral data set with elevation characteristics A' , stored in a data storage device; S3, combine the positioning system on the drone to attach the corresponding geographical location information to each collection point, so that the spectral data set with elevation features in step S2 A' Obtain spectral dataset with geographic location and elevation information A'' , and store it in a data storage device; S4. Generate rural landscape observation models with accurate spatial resolution and three-dimensional features.

2. A method for collecting local features of rural landscapes based on radar and spectrum fusion according to claim 1, characterized in that: The spectral dataset A : With height dataset B : The matching method is as follows: ； in, is the altitude correction factor of the UAV, It is the ranging correction factor caused by the time delay of the ranging equipment.

3. The method for collecting local features of rural landscape based on radar and spectrum fusion according to claim 1, characterized in that: The marking strategy of the geographic location in step S3 is as follows: S31, marking the position of the drone at a certain time point to the spectral data at the time point by means of a marking model A'n In the process of completing the preliminary calibration of the geographic location, the marking method is as follows: L(A'n)= L A'n ； in, L(A'n) For spectral data A'n After completing the preliminary calibration of the spectral data, is the time correction factor, L is location information; S32: Smoothing the data initially calibrated in step S31 to obtain a spectral data set with a certain degree of smoothness. A'' .

4. The method for collecting local features of rural landscape based on radar and spectrum fusion according to claim 1, characterized in that: In the preset route, the drone adopts a serpentine movement, and the distance between two adjacent parallel routes is no more than 10m.

5. The method for collecting local features of rural landscape based on radar and spectrum fusion according to claim 1, characterized in that: In step S4, the spectral data set is A'' The specific steps for assignment and modeling are as follows: S41, using a data processor to process the spectral data set A'' The image with height data is stretched vertically equivalently according to its height data to form a three-dimensional model with elevation features; S42, for the above spectral data set with three-dimensional model A'' Attach geographic location information and integrate it into the 3D map model library to form a 3D model with both elevation and geographic location features; S43. With the help of three-dimensional projection technology, a three-dimensional model with spectral image characteristics, elevation characteristics and geographical location characteristics is projected into a display cabinet to form a holographic projection of a three-dimensional model of the rural landscape.

6. The method for collecting local features of rural landscape based on radar and spectrum fusion according to claim 1, characterized in that: The flight altitude of the UAV is set at 500-1000m.

7. The method for collecting local features of rural landscape based on radar and spectrum fusion according to claim 1, characterized in that: The drone may also be equipped with a temperature and humidity sensor to detect the temperature and humidity data of the location; The temperature and humidity data can be matched to the spectral data set A'' A three-dimensional model of the rural landscape with temperature and humidity characteristics is formed.

8. The method for collecting local features of rural landscape based on radar and spectrum fusion according to claim 1, characterized in that: The step S4 further comprises the following steps before generating the model: The spectral dataset in step S1 is analyzed by using machine learning and image recognition technology. A Classify according to the element categories of buildings and natural landscapes, and count the spatial layout frequency of each element category , and then the spatial layout frequency For spectral datasets A'' Perform assignment and modeling; Before classification, the spectral dataset is A Perform grid division to form n sampling areas, and then classify these n sampling areas; The spatial layout frequency The calculation method is as follows: in, For the l The spatial distribution frequency of the class feature category, k is the total number of sampling areas, For the i In the sampling area l Image area ratio of feature categories, feature categories of buildings and natural landscapes l Specifically including: buildings, mountains, water, fields, forests, lakes, grass and sand.

9. A collection system using the method of claim 1, characterized in that: include: Spectral acquisition equipment, used to collect spectral data on the ground; Distance measuring equipment, used to collect elevation information of ground objects; A data storage device for storing spectral data sets, elevation information, and geographic location information; The unmanned aerial vehicle is used to carry the above-mentioned spectrum collection equipment, distance measurement equipment and data storage device.

10. A computer device, characterized in that: It includes a data processor, which can assign the elevation information and geographical location information stored in the data storage device to the spectral data set according to the method described in claim 1 to form a three-dimensional model of the rural landscape with three-dimensional characteristics.

Images