Green, blue and grey infrastructure classification methods, devices, systems and media

Abstract

The present invention discloses a method, device, system, and medium for classifying green, blue, and gray infrastructure. The method comprises: obtaining a multispectral photograph corresponding to a target area, and obtaining a target light band image set and a color orthophoto map based on the multispectral photograph; obtaining a sample file based on the color orthophoto map, and obtaining a classification result of green, blue, and gray infrastructure corresponding to the target area based on the target light band image set and the sample file. The present invention obtains a target light band image set and a color orthophoto map based on the multispectral photograph corresponding to the target area, and combines the sample file and the target light band image set obtained based on the color orthophoto map to obtain a classification result of green, blue, and gray infrastructure corresponding to the target area, thereby improving the accuracy and efficiency of green, blue, and gray infrastructure classification.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a method, device, system and medium for classifying green, blue and gray infrastructure. Background Art

[0002] With the continuous development of cities, urban infrastructure often includes water bodies, trees, grasslands, bare land, buildings, roads, and more. Different types of infrastructure have significant differences in their ability to utilize rainwater. Therefore, urban planners need to assist in city planning and construction based on the rainwater utilization performance of different types of infrastructure. However, existing methods for classifying urban infrastructure generally suffer from low accuracy and efficiency. Improving the accuracy and efficiency of green, blue, and gray infrastructure classification is an urgent issue. Summary of the Invention

[0003] The main purpose of the present invention is to propose a green, blue and gray infrastructure classification method, device, system and medium, aiming to solve the problem of how to improve the accuracy and efficiency of green, blue and gray infrastructure classification.

[0004] To achieve the above objectives, the present invention provides a method for classifying green, blue, and gray infrastructure, comprising the following steps:

[0005] Obtaining a multispectral photograph corresponding to the target area, and obtaining a target light band image set and a color orthophoto based on the multispectral photograph;

[0006] A sample file is obtained based on the color orthophoto, and green, blue and gray infrastructure classification results corresponding to the target area are obtained according to the target light band image set and the sample file.

[0007] Preferably, the step of obtaining a target light band image set and a color orthophoto based on the multispectral photograph includes:

[0008] A two-dimensional reconstruction operation is performed on the multispectral photograph, and a target light band image set and a color orthophoto are obtained based on the multispectral photograph after the two-dimensional reconstruction operation and a preset resolution.

[0009] Preferably, the sample file includes a training sample and a verification sample, and the step of obtaining the sample file based on the color orthophoto includes:

[0010] Generate a grid image according to a preset grid spacing, and overlap the grid image with the color orthophoto to obtain a color grid orthophoto;

[0011] Identifying a grid point attribute corresponding to each grid point in the color grid orthophoto, and obtaining the training sample according to the grid point attribute;

[0012] The grid image in the color grid orthophoto is offset to obtain the verification sample.

[0013] Preferably, the step of obtaining green, blue and gray infrastructure classification results corresponding to the target area according to the target light band image set and the sample file includes:

[0014] Segmenting the target light band image set according to spectral similarity to obtain shape objects, and calculating green, blue, and gray infrastructure pre-classification results based on the training samples in the sample file and the corresponding shape object light band image values;

[0015] The green, blue and gray infrastructure pre-classification results are evaluated for accuracy based on the verification sample in the sample file to obtain an accuracy evaluation result, and the green, blue and gray infrastructure classification results are determined based on the accuracy evaluation result.

[0016] Preferably, the step of determining green, blue and grey infrastructure classification results based on the accuracy evaluation result comprises:

[0017] Comparing the accuracy evaluation result with the preset accuracy to obtain a comparison result;

[0018] If the comparison result shows that the accuracy evaluation result is not less than the preset accuracy, the green, blue and gray infrastructure pre-classification results are determined as green, blue and gray infrastructure classification results;

[0019] If the comparison result is that the accuracy evaluation result is less than the preset accuracy, the grid image and the target light band image set are subjected to preset processing to obtain an optimal grid image and an optimal target light band image set, and the step of overlapping the grid image with the color orthophoto map to obtain a color grid orthophoto map is executed.

[0020] Preferably, the step of performing preset processing on the grid image and the target light band image set to obtain an optimal grid image and an optimal target light band image set includes:

[0021] performing a scaling operation on the grid spacing corresponding to the grid images to obtain a first preset number of grid images, and performing the first preset operation on each grid image to obtain an optimal grid image;

[0022] An increase or decrease operation is performed on the target light band images in the target light band image set to obtain a second preset number of target light band image sets, and a second preset operation is performed on each target light band image set based on the optimal grid image to obtain an optimal target light band image set.

[0023] Preferably, after the steps of obtaining a sample file based on the color grid orthophoto, and obtaining green, blue, and gray infrastructure classification results corresponding to the target area according to the target light band image set and the sample file, the green, blue, and gray infrastructure classification method further comprises:

[0024] A classification map corresponding to the target area is generated based on the green, blue and gray infrastructure classification results, and rainwater utilization analysis data corresponding to the target area is provided based on the classification map.

[0025] In addition, to achieve the above objectives, the present invention further provides a green, blue and gray infrastructure classification device, the green, blue and gray infrastructure classification device comprising:

[0026] An acquisition module is used to acquire a multispectral photograph corresponding to the target area, and obtain a target light band image set and a color orthophoto based on the multispectral photograph;

[0027] The classification module is used to obtain a sample file based on the color orthophoto, and obtain green, blue and gray infrastructure classification results corresponding to the target area according to the target light band image set and the sample file.

[0028] Furthermore, the acquisition module further includes a two-dimensional reconstruction module, and the two-dimensional reconstruction module is used to:

[0029] A two-dimensional reconstruction operation is performed on the multispectral photograph, and a target light band image set and a color orthophoto are obtained based on the multispectral photograph after the two-dimensional reconstruction operation and a preset resolution. Furthermore, the classification module also includes a generation module, which is used to:

[0030] Generate a grid image according to a preset grid spacing, and overlap the grid image with the color orthophoto to obtain a color grid orthophoto;

[0031] Identifying a grid point attribute corresponding to each grid point in the color grid orthophoto, and obtaining the training sample according to the grid point attribute;

[0032] The grid image in the color grid orthophoto is offset to obtain the verification sample.

[0033] Furthermore, the classification module is also used for:

[0034] Segmenting the target light band image set according to spectral similarity to obtain shape objects, and calculating green, blue, and gray infrastructure pre-classification results based on the training samples in the sample file and the corresponding shape object light band image values;

[0035] The green, blue and gray infrastructure pre-classification results are evaluated for accuracy based on the verification sample in the sample file to obtain an accuracy evaluation result, and the green, blue and gray infrastructure classification results are determined based on the accuracy evaluation result.

[0036] Furthermore, the classification module is also used for:

[0037] Comparing the accuracy evaluation result with the preset accuracy to obtain a comparison result;

[0038] If the comparison result shows that the accuracy evaluation result is not less than the preset accuracy, the green, blue and gray infrastructure pre-classification results are determined as green, blue and gray infrastructure classification results;

[0039] If the comparison result is that the accuracy evaluation result is less than the preset accuracy, the grid image and the target light band image set are subjected to preset processing to obtain an optimal grid image and an optimal target light band image set, and the step of overlapping the grid image with the color orthophoto map to obtain a color grid orthophoto map is executed.

[0040] Furthermore, the classification module further includes an optimization module, and the optimization module is used to:

[0041] performing a scaling operation on the grid spacing corresponding to the grid images to obtain a first preset number of grid images, and performing the first preset operation on each grid image to obtain an optimal grid image;

[0042] An increase or decrease operation is performed on the target light band images in the target light band image set to obtain a second preset number of target light band image sets, and a second preset operation is performed on each target light band image set based on the optimal grid image to obtain an optimal target light band image set.

[0043] Furthermore, the classification module further includes an analysis module, which is used to:

[0044] A classification map corresponding to the target area is generated based on the green, blue and gray infrastructure classification results, and rainwater utilization analysis data corresponding to the target area is provided based on the classification map.

[0045] In addition, to achieve the above-mentioned objectives, the present invention also provides a green, blue and gray infrastructure classification system, which includes: a memory, a processor and a green, blue and gray infrastructure classification program stored on the memory and executable on the processor, wherein the green, blue and gray infrastructure classification program implements the steps of the green, blue and gray infrastructure classification method described above when executed by the processor.

[0046] In addition, to achieve the above-mentioned purpose, the present invention also provides a medium, which is a computer-readable storage medium, and the computer-readable storage medium stores a green, blue and gray infrastructure classification program, and when the green, blue and gray infrastructure classification program is executed by a processor, it implements the steps of the green, blue and gray infrastructure classification method as described above.

[0047] The green, blue, and gray infrastructure classification method proposed in the present invention obtains a multispectral photograph corresponding to a target area and, based on the multispectral photograph, obtains a target light band image set and a color orthophoto map. Based on the color orthophoto map, a sample file is obtained, and based on the target light band image set and the sample file, a classification result for the green, blue, and gray infrastructure corresponding to the target area is obtained. The present invention obtains a target light band image set and a color orthophoto map based on the multispectral photograph corresponding to the target area, and combines the sample file and the target light band image set obtained based on the color orthophoto map to obtain a classification result for the green, blue, and gray infrastructure corresponding to the target area, thereby improving the accuracy and efficiency of green, blue, and gray infrastructure classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 Schematic diagram of the device structure of the hardware operating environment involved in the embodiment of the present invention;

[0049] Figure 2 Schematic diagram of the flow of the first embodiment of the green, blue and grey infrastructure classification method of the present invention;

[0050] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0051] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0052] like Figure 1 As shown, Figure 1 It is a schematic diagram of the device structure of the hardware operating environment involved in the embodiment of the present invention.

[0053] The device in the embodiment of the present invention may be a PC or a server device.

[0054] like Figure 1 As shown, the device may include: a processor 1001, such as a CPU, a network interface 1004, a user interface 1003, a memory 1005, and a communication bus 1002. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory, or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0055] Those skilled in the art will understand that Figure 1 The device structure shown in the figure does not constitute a limitation of the device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0056] like Figure 1 As shown, the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and green, blue, and gray infrastructure classification programs.

[0057] Among them, the operating system is a program that manages and controls portable green, blue and gray infrastructure classification devices and software resources, and supports the operation of network communication modules, user interface modules, green, blue and gray infrastructure classification programs and other programs or software; the network communication module is used to manage and control the network interface 1002; the user interface module is used to manage and control the user interface 1003.

[0058] exist Figure 1 In the green, blue and gray infrastructure classification device shown, the green, blue and gray infrastructure classification device calls the green, blue and gray infrastructure classification program stored in the memory 1005 through the processor 1001, and performs the operations in each embodiment of the green, blue and gray infrastructure classification method described below.

[0059] Based on the above hardware structure, an embodiment of the green, blue and grey infrastructure classification method of the present invention is proposed.

[0060] Reference Figure 2 , Figure 2This is a flow chart of a first embodiment of a method for classifying green, blue, and gray infrastructure according to the present invention, the method comprising:

[0061] Step S10, obtaining a multispectral photograph corresponding to the target area, and obtaining a target light band image set and a color orthophoto based on the multispectral photograph;

[0062] Step S20 , obtaining a sample file based on the color orthophoto, and obtaining green, blue and gray infrastructure classification results corresponding to the target area according to the target light band image set and the sample file.

[0063] The green, blue and gray infrastructure classification method of this embodiment is applied to the green, blue and gray infrastructure classification device of the urban planning agency. The green, blue and gray infrastructure classification device can be a terminal or PC device. For the convenience of description, the green, blue and gray infrastructure classification device is taken as an example for description; the green, blue and gray infrastructure classification device obtains the multispectral photos corresponding to the target area, and performs a two-dimensional reconstruction operation on the multispectral photos, and obtains a target light band image set and a color orthophoto according to the multispectral photos after the two-dimensional reconstruction operation and the preset resolution; the green, blue and gray infrastructure classification device generates a grid image according to the preset grid spacing, and overlaps the grid image with the color orthophoto to obtain a color grid orthophoto ... obtains a multispectral photo corresponding to the target area, and performs a two-dimensional reconstruction operation on the multispectral photo and the preset resolution; the green, blue and gray infrastructure classification device generates a grid image according to the preset grid spacing, and overlaps the grid image with the color orthophoto to obtain a color grid orthophoto; the green, blue and gray infrastructure classification device obtains a multispectral photo corresponding to the target area, and performs a two-dimensional reconstruction operation on the multispectral photo The blue and gray infrastructure classification devices identify the grid attributes corresponding to each grid point in the color grid orthophoto, obtain training samples based on the grid attributes, and then offset the grid image in the color grid orthophoto to obtain the verification samples. The green, blue, and gray infrastructure classification devices segment the target optical band image set based on spectral similarity to obtain shape objects. Based on the grid attributes in the training samples in the sample file and the corresponding shape object optical band image values, they calculate the green, blue, and gray infrastructure pre-classification results. Based on the green, blue, and gray infrastructure pre-classification results and the verification samples in the sample file, they perform an accuracy evaluation to obtain an accuracy evaluation result, and then determine the green, blue, and gray infrastructure classification results based on the accuracy evaluation result. It should be noted that the sample file includes training samples and verification samples. The infrastructure types are set by the relevant R&D personnel as seven categories: water bodies, trees and shrubs, grasslands, green roofs, bare land, buildings, and roads. Water bodies are blue infrastructure, trees and shrubs, grasslands, green roofs, and bare land are green infrastructure, and buildings and roads are gray infrastructure.

[0064] The green, blue, and gray infrastructure classification method of this embodiment obtains a multispectral photograph corresponding to the target area and, based on the multispectral photograph, obtains a target light band image set and a color orthophoto. Based on the color orthophoto, a sample file is obtained, and based on the target light band image set and the sample file, a classification result for the green, blue, and gray infrastructure corresponding to the target area is obtained. The present invention obtains a target light band image set and a color orthophoto based on the multispectral photograph corresponding to the target area, and combines the sample file and the target light band image set obtained based on the color orthophoto to obtain a classification result for the green, blue, and gray infrastructure corresponding to the target area, thereby improving the accuracy and efficiency of green, blue, and gray infrastructure classification.

[0065] The following describes each step in detail:

[0066] Step S10, obtaining a multispectral photograph corresponding to the target area, and obtaining a target light band image set and a color orthophoto based on the multispectral photograph;

[0067] In this embodiment, the green, blue and gray infrastructure classification equipment uses a drone with a multispectral photo shooting function to shoot the target area under suitable weather conditions to obtain multispectral photos corresponding to the target area, and obtains a target light band image set and a color orthophoto based on the multispectral photos; for example, relevant researchers determine the target area to be studied according to actual conditions, and use a drone with a multispectral photo shooting function to fly to a preset height above the ground of the target area to shoot the target area, and send the multispectral photos corresponding to the target area taken by the drone to the green, blue and gray infrastructure classification equipment. When the green, blue and gray infrastructure classification equipment obtains the multispectral photos corresponding to the target area, it obtains the target area corresponding to the multispectral photos based on the multispectral photos. The target light band image set and color orthophoto are included. It should be noted that multispectral photos refer to photos containing many bands, sometimes only 3 bands (such as color images), but sometimes many more bands, even hundreds of bands. Each band is a grayscale image, which represents the scene brightness obtained according to the sensitivity of the sensor used to generate the band. In multispectral photos, each pixel is associated with a numerical string of pixels in different bands, that is, a vector; the target light band image set includes: blue light band images, green light band images, red light band images, red edge light band images, near-infrared light band images, NDVI images and DSM images. Among them, NDVI images are images for detecting vegetation growth status, vegetation coverage and eliminating some radiation errors. DSM (Digital Surface Model) refers to a ground elevation model that includes the heights of surface buildings, bridges, and trees, that is, DSM images refer to digital surface model images; color orthophoto refers to a color bird's-eye view of the target area taken by a drone from above.

[0068] Specifically, step S10 includes:

[0069] Step a: performing a two-dimensional reconstruction operation on the multispectral photograph, and obtaining a target light band image set and a color orthophoto according to the multispectral photograph after the two-dimensional reconstruction operation and a preset resolution.

[0070] In this step, the green, blue, and gray infrastructure classification equipment performs a two-dimensional reconstruction operation on the multispectral photos corresponding to the target area, and obtains a target light band image set and a color orthophoto based on the multispectral photos after the two-dimensional reconstruction operation and the preset resolution, where the preset resolution can be 6 cm, 6.5 cm, 8 cm, or 10 cm, etc. For example, relevant researchers set the preset resolution to 6 cm, and the green, blue, and gray infrastructure classification equipment inputs the multispectral photos corresponding to the target area into DJI Terra or similar image stitching software, performs a two-dimensional reconstruction operation on the multispectral photos, and obtains blue light band images, green light band images, red light band images, red edge light band images, near-infrared light band images, NDVI images, DSM images, and color orthophotos with a resolution of 6 cm, respectively.

[0071] Step S20 , obtaining a sample file based on the color orthophoto, and obtaining green, blue and gray infrastructure classification results corresponding to the target area according to the target light band image set and the sample file.

[0072] In this embodiment, the green, blue, and gray infrastructure classification device uses a color orthophoto map and obtains a sample file using eCognition and Arc GIS software. It then selects one or more images from the target light band image collection: a blue light band image, a green light band image, a red light band image, a red-edge light band image, a near-infrared light band image, an NDVI image, or a DSM image. These images are then combined with the sample file to obtain the green, blue, and gray infrastructure classification results corresponding to the target area. It should be noted that eCognition is intelligent image analysis software that uses an object-oriented information extraction method, fully utilizing object information (hue, shape, texture, and hierarchy) and inter-class information (characteristics related to neighboring objects, child objects, and parent objects) for analysis. Arc GIS software is software for collecting, organizing, managing, analyzing, communicating, and publishing geographic information. The sample file includes training samples and validation samples.

[0073] Specifically, step S20 further includes:

[0074] Step b: generating a grid image according to a preset grid spacing, and overlapping the grid image with the color orthophoto to obtain a color grid orthophoto;

[0075] In this step, the green, blue and gray infrastructure classification devices generate grid images through the eCognition software according to the preset grid spacing, and overlap the grid images with the color orthophoto corresponding to the target area to obtain a color grid orthophoto; for example, relevant researchers set the sampling interval to 11.6 meters based on experience, and proportionally reduce the sampling interval based on the size corresponding to the grid image to be generated to obtain the preset grid spacing. The green, blue and gray infrastructure classification devices generate grid images through the eCognition software according to the preset grid spacing, and overlay the grid images on the color orthophoto surface so that the grid image overlaps with the color orthophoto to obtain a color grid orthophoto. It can be understood that the color grid orthophoto is obtained by overlaying the grid image on the color orthophoto surface. The color grid orthophoto surface has a grid, wherein each grid point on the grid corresponds to a different infrastructure on the color grid orthophoto.

[0076] Step c, identifying the grid point attributes corresponding to each grid point in the color grid orthophoto, and obtaining the training sample according to the grid point attributes;

[0077] In this step, after obtaining the color grid orthophoto, the green, blue and gray infrastructure classification equipment uses Arc GIS software to identify the grid attributes corresponding to each grid point in the color grid orthophoto, and obtains training samples based on the grid attributes corresponding to each grid point; it can be understood that the grid point refers to the intersection formed by the intersection of two line segments in the grid, and each grid point in the color grid orthophoto will correspond to a different infrastructure on the color grid orthophoto. The grid attribute is the type of infrastructure corresponding to the grid point. The types of infrastructure are set by relevant R&D personnel into seven categories: water bodies, trees and shrubs, grasslands, green roofs, bare land, buildings, and roads. Among them, water bodies are blue infrastructure, trees and shrubs, grasslands, green roofs, and bare land are green infrastructure, and buildings and roads are gray infrastructure.

[0078] Step d: offset the grid image in the color grid orthophoto to obtain the verification sample.

[0079] In this step, after obtaining the training sample, the green, blue and gray infrastructure classification equipment offsets the grid in the color grid orthophoto. Specifically, the grid as a whole can be offset in the upward, downward, left, right, upper left, upper right and other directions, so that the grid point attributes corresponding to each grid point in the grid are different from the grid point attributes corresponding to each grid point in the training sample, thereby obtaining a verification sample. For example: the green, blue and gray infrastructure classification equipment offsets the grid in the color grid orthophoto as a whole downward by a preset grid spacing, and identifies the grid attributes corresponding to each grid point after the offset, thereby obtaining the corresponding verification sample. Optionally, the grid in the color grid orthophoto can be offset in multiple directions as a whole, thereby obtaining multiple verification samples.

[0080] Step e: performing a segmentation operation on the target light band image set according to spectral similarity to obtain shape objects, and calculating green, blue and gray infrastructure pre-classification results based on the training samples in the sample file and the corresponding shape object light band image values;

[0081] In this step, the green, blue, and gray infrastructure classification equipment uses eCognition software to segment the target light band image set. Based on the band value similarity of the blue light band image, green light band image, red light band image, red-edge light band image, near-infrared light band image, NDVI image, and DSM image in the target light band image set, shape objects are segmented and the grid attributes of the training samples in the sample file are assigned to each shape object. Then, through algorithms such as random forest, fuzzy classification, and Bayesian algorithm, the attribute values ​​of other attributeless shape objects are calculated to obtain the green, blue, and gray infrastructure pre-classification results.

[0082] Step f: performing accuracy evaluation on the green, blue and gray infrastructure pre-classification results based on the verification samples in the sample file to obtain accuracy evaluation results, and determining green, blue and gray infrastructure classification results based on the accuracy evaluation results.

[0083] In this step, the green, blue and gray infrastructure classification device performs accuracy evaluation on the green, blue and gray infrastructure pre-classification results based on the grid point attributes corresponding to each grid point in the verification sample in the sample file, obtains an accuracy evaluation result, and determines the green, blue and gray infrastructure classification results based on the accuracy evaluation result.

[0084] Furthermore, based on the accuracy evaluation result, the step of determining the classification results of green, blue and gray infrastructure includes:

[0085] Step f1, comparing the accuracy evaluation result with the preset accuracy to obtain a comparison result;

[0086] In this step, the green, blue and gray infrastructure classification equipment compares the accuracy evaluation result with the preset accuracy to obtain a comparison result. For example, relevant researchers set the preset accuracy to 0.8 based on actual conditions. If the accuracy evaluation result of the green, blue and gray infrastructure pre-classification result obtained by the green, blue and gray infrastructure classification equipment is 0.7, then the comparison result obtained by comparing the accuracy evaluation result with the preset accuracy is that the accuracy evaluation result is less than the preset accuracy. If the accuracy evaluation result of the green, blue and gray infrastructure pre-classification result obtained by the green, blue and gray infrastructure classification equipment is 0.85, then the comparison result obtained by comparing the accuracy evaluation result with the preset accuracy is that the accuracy evaluation result is not less than the preset accuracy.

[0087] Step f2: if the comparison result shows that the accuracy evaluation result is not less than the preset accuracy, the green, blue and gray infrastructure pre-classification results are determined as green, blue and gray infrastructure classification results;

[0088] In this step, if the comparison result obtained by the green, blue and gray infrastructure classification equipment is that the accuracy evaluation result is not less than the preset accuracy, the green, blue and gray infrastructure pre-classification result is determined as the green, blue and gray infrastructure classification result, and the green, blue and gray infrastructure classification result is used to provide rainwater utilization analysis data corresponding to the target area.

[0089] Step f3: If the comparison result shows that the accuracy evaluation result is less than the preset accuracy, the grid image and the target light band image set are subjected to preset processing to obtain an optimal grid image and an optimal target light band image set, and the step of overlapping the grid image with the color orthophoto to obtain a color grid orthophoto is executed.

[0090] In this step, if the comparison result obtained by the green, blue and gray infrastructure classification equipment is that the accuracy evaluation result is less than the preset accuracy, the grid image and the target light band image set are preset processed to obtain the optimal grid image and the optimal target light band image set, and the grid image is overlapped with the color orthophoto to obtain the color grid orthophoto and subsequent steps until the accuracy corresponding to the green, blue and gray infrastructure pre-classification results is not less than the preset accuracy.

[0091] Furthermore, the step of performing preset processing on the grid image and the target light band image set to obtain an optimal grid image and an optimal target light band image set includes:

[0092] performing a scaling operation on the grid spacing corresponding to the grid images to obtain a first preset number of grid images, and performing the first preset operation on each grid image to obtain an optimal grid image;

[0093] In this step, the green, blue, and gray infrastructure classification devices scale the grid spacing corresponding to the grid images to obtain a first preset number of grid images, and perform the first preset operation on each of the first preset number of grid images to obtain an optimal grid image. For example, the green, blue, and gray infrastructure classification devices may scale the grid spacing corresponding to the grid images based on the preset grid spacing to obtain the first preset number of grid images. Optionally, the green, blue, and gray infrastructure classification devices may increase or decrease the grid spacing corresponding to the grid images based on the preset grid spacing according to instructions from relevant researchers. Each increase or decrease in the grid spacing is performed once to obtain a corresponding grid image, and the process stops when the first preset number of grid images is obtained. Optionally, the green, blue, and gray infrastructure classification devices may intelligently set the first preset number and intelligently increase or decrease the grid spacing corresponding to the grid images based on the preset grid spacing to obtain the first preset number of grid images. After obtaining a first preset number of grid images, the green, blue and gray infrastructure classification device performs a first preset operation on each grid image, that is, overlapping the grid image with the color orthophoto for each grid image to obtain a color grid orthophoto and subsequent steps, to obtain the accuracy evaluation results corresponding to the green, blue and gray infrastructure pre-classification results corresponding to each grid image, and then compares the accuracy evaluation results corresponding to each grid image with the preset accuracy, screens out the grid images corresponding to the accuracy evaluation results not less than the preset accuracy, and screens out the grid images corresponding to the accuracy evaluation results not less than the preset accuracy as the optimal grid images.

[0094] An increase or decrease operation is performed on the target light band images in the target light band image set to obtain a second preset number of target light band image sets, and a second preset operation is performed on each target light band image set based on the optimal grid image to obtain an optimal target light band image set.

[0095] In this step, the green, blue, and gray infrastructure classification devices perform addition and subtraction operations on the target light band images in the target light band image set to obtain a second preset number of target light band image sets. Based on the optimal grid image, the second preset operation is performed on each target light band image set to obtain the optimal target light band image set. For example, the green, blue, and gray infrastructure classification devices perform addition and subtraction operations on the target light band images in the target light band image set to obtain the second preset number of target light band image sets. Optionally, the green, blue, and gray infrastructure classification devices, based on instructions from relevant researchers, add or subtract target light band images from the target light band image set until the second preset number of target light band image sets is obtained and then stop. Optionally, the green, blue, and gray infrastructure classification devices can intelligently set the second preset number and intelligently add or subtract target light band images from the target light band image set to obtain the second preset number of target light band image sets. Specifically, the target light band image set may include: blue light band images, green light band images, NDVI images, DSM images; may include blue light band images, green light band images, red light band images, NDVI images, DSM images; may also include blue light band images, green light band images, red light band images, near-infrared light band images, NDVI images, DSM images, etc. The green, blue and gray infrastructure classification device obtains training samples and verification samples based on the optimal grid image and the color orthophoto corresponding to the target area, and performs a second preset operation on each target light band image set in the second preset number, that is, performs a segmentation operation on each target light band image set in the second preset number to obtain a shape object, and calculates the green, blue and gray infrastructure pre-classification results and subsequent steps based on the training samples and shape objects in the sample file, so as to obtain the accuracy evaluation results corresponding to the green, blue and gray infrastructure pre-classification results corresponding to each target light band image set, and screens out the target light band image set with the highest accuracy evaluation from each target light band image set whose accuracy evaluation result is not less than the preset accuracy as the optimal target light band image set.

[0096] In the green, blue and gray infrastructure classification method of this embodiment, the green, blue and gray infrastructure classification device obtains the multispectral photos corresponding to the target area, and performs a two-dimensional reconstruction operation on the multispectral photos, and obtains a target light band image set and a color orthophoto according to the multispectral photos after the two-dimensional reconstruction operation and a preset resolution; the green, blue and gray infrastructure classification device generates a grid image according to a preset grid spacing, and overlaps the grid image with the color orthophoto to obtain a color grid orthophoto; the green, blue and gray infrastructure classification device identifies the grid point attributes corresponding to each grid point in the color grid orthophoto, and obtains a training image according to the grid point attributes. The training samples are obtained, and then the grid images in the color grid orthophoto are offset to obtain verification samples; the green, blue and gray infrastructure classification equipment performs a segmentation operation on the target light band image set to obtain shape objects, and calculates the green, blue and gray infrastructure pre-classification results based on the grid attributes and shape objects in the training samples in the sample file, and then performs accuracy evaluation based on the green, blue and gray infrastructure pre-classification results and the verification samples in the sample file to obtain accuracy evaluation results, and determines the green, blue and gray infrastructure classification results based on the accuracy evaluation results, thereby improving the accuracy and efficiency of green, blue and gray infrastructure classification.

[0097] Furthermore, based on the first embodiment of the green, blue and gray infrastructure classification method of the present invention, a second embodiment of the green, blue and gray infrastructure classification method of the present invention is proposed.

[0098] The second embodiment of the green, blue, and gray infrastructure classification method differs from the first embodiment of the green, blue, and gray infrastructure classification method in that, after step S20, the green, blue, and gray infrastructure classification method further includes:

[0099] Step g: generating a classification map corresponding to the target area based on the green, blue and gray infrastructure classification results, and providing rainwater utilization analysis data corresponding to the target area based on the classification map.

[0100] In this embodiment, the green, blue and gray infrastructure classification device generates a classification map corresponding to the target area based on the green, blue and gray infrastructure classification results obtained by the grid image corresponding to the preset grid spacing and the target light band image set consisting of blue light band image, green light band image, red light band image, red edge light band image, near infrared light band image, NDVI image and DSM image, or the green, blue and gray infrastructure classification results obtained by the optimal grid image and the optimal target light band image set. The classification map includes seven types of infrastructure in the target area: water bodies, trees and shrubs, grasslands, green roofs, bare land, buildings, and roads, and provides rainwater utilization analysis data corresponding to the target area based on the classification map.

[0101] The green, blue and gray infrastructure classification device of this embodiment generates a classification map corresponding to the target area based on the final green, blue and gray infrastructure classification results of the target area. The classification map includes seven types of infrastructure types of the target area: water bodies, trees and shrubs, grasslands, green roofs, bare land, buildings, and roads. Based on the classification map, rainwater utilization analysis data corresponding to the target area is provided, so that the green, blue and gray infrastructure classification results can provide data for rainwater utilization analysis of the target area.

[0102] The present invention also provides a green, blue and gray infrastructure classification device. The green, blue and gray infrastructure classification device of the present invention comprises:

[0103] An acquisition module is used to acquire a multispectral photograph corresponding to the target area, and obtain a target light band image set and a color orthophoto based on the multispectral photograph;

[0104] The classification module is configured to obtain a sample file based on the color orthophoto, and obtain the green, blue, and gray infrastructure classification results corresponding to the target area based on the target light band image set and the sample file. Furthermore, the acquisition module also includes a two-dimensional reconstruction module, which is configured to:

[0105] A two-dimensional reconstruction operation is performed on the multispectral photograph, and a target light band image set and a color orthophoto are obtained based on the multispectral photograph after the two-dimensional reconstruction operation and a preset resolution. Furthermore, the classification module also includes a generation module, which is used to:

[0106] Generate a grid image according to a preset grid spacing, and overlap the grid image with the color orthophoto to obtain a color grid orthophoto;

[0107] Identifying a grid point attribute corresponding to each grid point in the color grid orthophoto, and obtaining the training sample according to the grid point attribute;

[0108] The grid image in the color grid orthophoto is offset to obtain the verification sample.

[0109] Furthermore, the classification module is also used for:

[0110] Segmenting the target light band image set according to spectral similarity to obtain shape objects, and calculating green, blue, and gray infrastructure pre-classification results based on the training samples in the sample file and the corresponding shape object light band image values;

[0111] The green, blue and gray infrastructure pre-classification results are evaluated for accuracy based on the verification sample in the sample file to obtain an accuracy evaluation result, and the green, blue and gray infrastructure classification results are determined based on the accuracy evaluation result.

[0112] Furthermore, the classification module is also used for:

[0113] Comparing the accuracy evaluation result with the preset accuracy to obtain a comparison result;

[0114] If the comparison result shows that the accuracy evaluation result is not less than the preset accuracy, the green, blue and gray infrastructure pre-classification results are determined as green, blue and gray infrastructure classification results;

[0115] If the comparison result is that the accuracy evaluation result is less than the preset accuracy, the grid image and the target light band image set are subjected to preset processing to obtain an optimal grid image and an optimal target light band image set, and the step of overlapping the grid image with the color orthophoto map to obtain a color grid orthophoto map is executed.

[0116] Furthermore, the classification module further includes an optimization module, and the optimization module is used to:

[0117] performing a scaling operation on the grid spacing corresponding to the grid images to obtain a first preset number of grid images, and performing the first preset operation on each grid image to obtain an optimal grid image;

[0118] An increase or decrease operation is performed on the target light band images in the target light band image set to obtain a second preset number of target light band image sets, and a second preset operation is performed on each target light band image set based on the optimal grid image to obtain an optimal target light band image set.

[0119] Furthermore, the classification module further includes an analysis module, which is used to:

[0120] A classification map corresponding to the target area is generated based on the green, blue and gray infrastructure classification results, and rainwater utilization analysis data corresponding to the target area is provided based on the classification map.

[0121] The present invention also provides a medium.

[0122] The medium of the present invention is a computer-readable storage medium, which stores a green, blue and gray infrastructure classification program. When the green, blue and gray infrastructure classification program is executed by a processor, the steps of the green, blue and gray infrastructure classification method described above are implemented.

[0123] Among them, the method implemented when the green, blue and gray infrastructure classification program running on the processor is executed can refer to the various embodiments of the green, blue and gray infrastructure classification method of the present invention, and will not be repeated here.

[0124] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or system comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or system. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or system comprising the element.

[0125] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0126] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better embodiment. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0127] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for classifying green, blue and grey infrastructure, characterized in that: The green, blue, and grey infrastructure classification method includes the following steps: Obtaining a multispectral photograph corresponding to the target area, and obtaining a target light band image set and a color orthophoto based on the multispectral photograph; Based on the color orthophoto, a sample file is obtained, and according to the target light band image set and the sample file, green, blue and gray infrastructure classification results corresponding to the target area are obtained; The sample file includes a training sample and a verification sample. The step of obtaining the sample file based on the color orthophoto includes: Generate a grid image according to a preset grid spacing, and overlap the grid image with the color orthophoto to obtain a color grid orthophoto; identify the grid point attributes corresponding to each grid point in the color grid orthophoto, and obtain the training sample according to the grid point attributes; offset the grid image in the color grid orthophoto to obtain the verification sample; The step of obtaining green, blue, and gray infrastructure classification results corresponding to the target area according to the target light band image set and the sample file includes: Performing a segmentation operation on the target light band image set based on spectral similarity to obtain shape objects, and calculating green, blue, and gray infrastructure pre-classification results based on the training samples in the sample file and the corresponding shape object light band image values; performing an accuracy evaluation on the green, blue, and gray infrastructure pre-classification results based on the verification samples in the sample file to obtain an accuracy evaluation result, and determining green, blue, and gray infrastructure classification results based on the accuracy evaluation result; The step of determining green, blue, and gray infrastructure classification results based on the accuracy evaluation result includes: The accuracy evaluation result is compared with the preset accuracy to obtain a comparison result; if the comparison result is that the accuracy evaluation result is not less than the preset accuracy, the green, blue and gray infrastructure pre-classification results are determined as green, blue and gray infrastructure classification results; if the comparison result is that the accuracy evaluation result is less than the preset accuracy, the grid image and the target light band image set are preset processed to obtain the optimal grid image and the optimal target light band image set, and the step of overlapping the grid image with the color orthophoto to obtain a color grid orthophoto.

2. The green, blue and grey infrastructure classification method according to claim 1, characterized in that: The step of obtaining a target light band image set and a color orthophoto based on the multispectral photograph comprises: A two-dimensional reconstruction operation is performed on the multispectral photograph, and a target light band image set and a color orthophoto are obtained based on the multispectral photograph after the two-dimensional reconstruction operation and a preset resolution.

3. The green, blue and grey infrastructure classification method according to claim 1, characterized in that: The step of performing preset processing on the grid image and the target light band image set to obtain an optimal grid image and an optimal target light band image set includes: performing a scaling operation on the grid spacing corresponding to the grid images to obtain a first preset number of grid images, and performing the first preset operation on each grid image to obtain an optimal grid image; An increase or decrease operation is performed on the target light band images in the target light band image set to obtain a second preset number of target light band image sets, and a second preset operation is performed on each target light band image set based on the optimal grid image to obtain an optimal target light band image set.

4. The green, blue and grey infrastructure classification method according to claim 1, wherein: After the steps of obtaining a sample file based on the color orthophoto, and obtaining green, blue, and gray infrastructure classification results corresponding to the target area according to the target light band image set and the sample file, the green, blue, and gray infrastructure classification method further includes: A classification map corresponding to the target area is generated based on the green, blue and gray infrastructure classification results, and rainwater utilization analysis data corresponding to the target area is provided based on the classification map.

5. A green, blue and grey infrastructure classification device, characterized in that The green, blue and grey infrastructure categories include: An acquisition module is used to acquire a multispectral photograph corresponding to the target area, and obtain a target light band image set and a color orthophoto based on the multispectral photograph; A classification module is configured to obtain a sample file based on the color orthophoto, and obtain green, blue, and gray infrastructure classification results corresponding to the target area based on the target light band image set and the sample file; The classification module is further configured to generate a grid image according to a preset grid spacing, and overlap the grid image with the color orthophoto to obtain a color grid orthophoto; identify the grid point attributes corresponding to each grid point in the color grid orthophoto, and obtain training samples based on the grid point attributes; and offset the grid image in the color grid orthophoto to obtain a verification sample. The classification module is further configured to perform a segmentation operation on the target light band image set based on spectral similarity to obtain shape objects, and calculate green, blue, and gray infrastructure pre-classification results based on the training samples in the sample file and the corresponding shape object light band image values; perform an accuracy evaluation on the green, blue, and gray infrastructure pre-classification results based on the verification samples in the sample file to obtain an accuracy evaluation result; and determine the green, blue, and gray infrastructure classification results based on the accuracy evaluation result; The classification module is further used to compare the accuracy evaluation result with the preset accuracy to obtain a comparison result; if the comparison result is that the accuracy evaluation result is not less than the preset accuracy, the green, blue and gray infrastructure pre-classification results are determined as green, blue and gray infrastructure classification results; if the comparison result is that the accuracy evaluation result is less than the preset accuracy, the grid image and the target light band image set are preset processed to obtain the optimal grid image and the optimal target light band image set, and the step of overlapping the grid image with the color orthophoto map to obtain a color grid orthophoto map is executed.

6. A classification system for green, blue and grey infrastructure, characterized by: The green, blue and grey infrastructure classification system includes: a memory, a processor and a green, blue and grey infrastructure classification program stored on the memory and executable on the processor. When the green, blue and grey infrastructure classification program is executed by the processor, the steps of the green, blue and grey infrastructure classification method according to any one of claims 1 to 4 are implemented.

7. A medium, characterized in that The medium is a computer-readable storage medium, on which a green, blue and gray infrastructure classification program is stored. When the green, blue and gray infrastructure classification program is executed by a processor, the steps of the green, blue and gray infrastructure classification method according to any one of claims 1 to 4 are implemented.

Images