Riverbed sediment measurement method and device based on image and storage medium

Abstract

The invention relates to the technical field of riverbed sediment measurement, in particular to an image-based riverbed sediment measurement method, device and equipment and a computer storage medium. The method comprises the following steps: acquiring a first target riverbed sediment picture sampled by an unmanned aerial vehicle at a height lower than the highest flight height of a first unmanned aerial vehicle under the condition of low flight of the unmanned aerial vehicle for riverbed sediment particle size detection; and acquiring a second target riverbed sediment picture sampled by the unmanned aerial vehicle at a height lower than the highest flight height of the second unmanned aerial vehicle under the condition of no low flight of the unmanned aerial vehicle, inputting the target sediment image super-resolution model to obtain a target super-resolution riverbed sediment image, and carrying out riverbed sediment particle size detection. According to the invention, the accuracy of riverbed sediment measurement is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of riverbed sediment measurement, and in particular to an image-based riverbed sediment measurement method, device, equipment and computer storage medium. Background Art

[0002] While image-based riverbed sediment measurement technology can quickly obtain geometric parameters such as sediment particle size distribution and shape, it suffers from poor accuracy when measuring fine-grained sediment. This is because fine-grained sediment typically occupies only a few pixels in an image, making its texture and edge features difficult to clearly display, resulting in inaccurate measurement.

[0003] The boundary particle size between gravel and sand is 2 mm (according to the "River Sediment Particle Analysis Procedure" SL42-2010). Accurately measuring millimeter-level sediment particles is very important for the management of riverbed erosion and siltation, the management of aquatic habitats, and flood prevention and disaster reduction.

[0004] However, the currently commonly used drone mapping technology generally flies at an altitude of more than 20 meters (to ensure the efficiency and safety of the drone and avoid collisions with trees, telephone poles, etc.), resulting in the minimum sediment particle size that can be identified from drone images being usually tens of millimeters or hundreds of millimeters, which cannot meet the measurement needs of millimeter-level fine-grained sediment.

[0005] Currently, there is a lack of effective methods to overcome the influence of image resolution on the measurement accuracy of fine-grained sediment, so as to realize the measurement and identification of millimeter-scale sediment particles using UAV sediment image measurement technology. Summary of the Invention

[0006] To this end, the technical problem to be solved by the present invention is to overcome the problem of low measurement and identification accuracy of millimeter-level sediment particles based on drone images in the prior art.

[0007] To solve the above technical problems, the present invention provides an image-based riverbed sediment measurement method, comprising:

[0008] Obtain the first riverbed sediment image samples taken by the drone at different flight altitudes, and create a scatter plot of the drone's flight altitude and image resolution;

[0009] Fitting the scatter plot to obtain a fitting relationship between the UAV flight altitude and the image resolution;

[0010] Determine the resolution corresponding to the first minimum sediment particle size pixel value that can be identified in the original resolution image, and obtain the maximum flight altitude of the first UAV based on the fitting relationship;

[0011] Under the condition of low-flying drone, obtain the first target riverbed sediment image sampled by the drone at a lower than the maximum flight altitude of the first drone, and perform riverbed sediment particle size detection.

[0012] Preferably, the fitting relationship is:

[0013] R=0.1·H 1.15

[0014] Among them, H is the flight altitude of the drone and R is the image resolution.

[0015] Preferably, the first minimum sediment particle size pixel value is 4.

[0016] Preferably, the image-based riverbed sediment measurement method further includes:

[0017] Based on the fitting relationship, a second riverbed sediment image sample with a resolution higher than a preset threshold is collected, and the image sample is upsampled n times to obtain a third riverbed sediment sample to construct a training set;

[0018] Taking the third riverbed sediment sample as input and the second riverbed sediment image sample as output, constructing a sediment image super-resolution model;

[0019] Training the sediment image super-resolution model based on a deep learning algorithm to obtain a trained sediment image super-resolution model;

[0020] Testing and analyzing the trained sediment image super-resolution model to determine the optimal upsampling factor and obtain a target sediment image super-resolution model;

[0021] Determining the resolution corresponding to the second minimum sediment particle size pixel value that can be identified in the super-resolution image processed by the target sediment image super-resolution model, and obtaining the maximum flight altitude of the second UAV based on the fitting relationship;

[0022] In the absence of low-flying UAV conditions, obtain the second target riverbed sediment image sampled by the UAV at a lower altitude than the second UAV's maximum flight altitude;

[0023] The second target riverbed sediment image is input into the target sediment image super-resolution model to obtain a target super-resolution riverbed sediment image, and riverbed sediment particle size detection is performed on the target super-resolution riverbed sediment image.

[0024] Preferably, the upsampling adopts a nearest neighbor sampling algorithm.

[0025] Preferably, the optimal upsampling factor is 4.

[0026] Preferably, the second minimum sediment particle size pixel value is 1.

[0027] The present invention also provides an image-based riverbed sediment measurement device, comprising:

[0028] A scatter plot construction module is used to obtain the first riverbed sediment image samples sampled by the drone at different flight altitudes and to create a scatter plot of the drone's flight altitude and image resolution;

[0029] A fitting relationship building module is used to fit the scatter plot to obtain a fitting relationship between the UAV flight altitude and the image resolution;

[0030] a first UAV maximum flight altitude determination module, configured to determine a resolution corresponding to a first minimum sediment particle size pixel value that can be identified in the original resolution image, and obtain the first UAV maximum flight altitude based on the fitting relationship;

[0031] The sediment measurement module is used to obtain the first target riverbed sediment image sampled by the drone at a lower altitude than the first drone's maximum flight altitude under the condition of low-flying drone, and to detect the riverbed sediment particle size.

[0032] The present invention also provides an image-based riverbed sediment measurement device, comprising:

[0033] memory for storing computer programs;

[0034] The processor is configured to implement the steps of the above-mentioned image-based riverbed sediment measurement method when executing the computer program.

[0035] The present invention also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the above-mentioned image-based riverbed sediment measurement method are implemented.

[0036] The above technical solution of the present invention has the following advantages over the prior art:

[0037] The image-based riverbed sediment measurement method of the present invention determines the resolution corresponding to the first minimum sediment particle size pixel value that can be identified in the original resolution image based on the fitting relationship between the drone flight altitude and the image resolution, obtains the first drone's maximum flight altitude, obtains the first target riverbed sediment image sampled by the drone at a lower altitude than the first drone's maximum flight altitude, and performs riverbed sediment particle size detection; when the drone does not fly low, determines the resolution corresponding to the second minimum sediment particle size pixel value that can be identified in the super-resolution image processed by the target sediment image super-resolution model, and obtains the second drone's maximum flight altitude based on the fitting relationship, obtains the second target riverbed sediment image sampled by the drone at a lower altitude than the second drone's maximum flight altitude, inputs the second target riverbed sediment image into the target sediment image super-resolution model to obtain a target super-resolution riverbed sediment image, and performs riverbed sediment particle size detection on the target super-resolution riverbed sediment image. The present invention effectively improves the accuracy of riverbed sediment measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to make the content of the present invention more clearly understood, the present invention is further described in detail below based on specific embodiments of the present invention in conjunction with the accompanying drawings, wherein:

[0039] Figure 1 This is a flow chart for implementing an image-based riverbed sediment measurement method provided by the present invention;

[0040] Figure 2 It is a scatter plot of the resolution of riverbed sediment photos taken by drone at different flight altitudes;

[0041] Figure 3 are riverbed sediment images of different resolutions;

[0042] Figure 4 Schematic diagram of the sediment image super-resolution model training process;

[0043] Figure 5 Newly sampled sediment photos and corresponding 4x SRGAN super-resolution model output photos. DETAILED DESCRIPTION

[0044] The core of the present invention is to provide an image-based riverbed sediment measurement method, device, equipment and computer storage medium, which effectively improves the accuracy of riverbed sediment measurement.

[0045] In order to enable those skilled in the art to better understand the present invention, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0046] Please refer to Figure 1 , Figure 1 This is a flow chart of an implementation of an image-based riverbed sediment measurement method provided by the present invention; the specific operation steps are as follows:

[0047] S101: Obtain first riverbed sediment image samples sampled by the drone at different flight altitudes, and create a scatter plot of the drone's flight altitude and image resolution;

[0048] S102: Fitting the scatter plot to obtain a fitting relationship between the UAV flight altitude and the image resolution;

[0049] S103: Determine the resolution corresponding to the first minimum sediment particle size pixel value that can be identified in the original resolution image, and obtain the maximum flight altitude of the first UAV according to the fitting relationship;

[0050] S104: Under the condition of low-flying drone, obtain a first target riverbed sediment image sampled by the drone at a lower altitude than the maximum flight altitude of the first drone, and perform riverbed sediment particle size detection.

[0051] Based on the above embodiment, the present invention provides a detailed description of step S101:

[0052] In a specific embodiment, a drone equipped with a 45-megapixel camera is used to sample riverbed sediment images at different heights for typical mountain river types with different riverbed morphologies and different gradations, to obtain a first riverbed sediment image sample, and a scatter plot of the drone's flight altitude (H, in meters) and the image resolution (R, in mm / pixel) is obtained; Figure 2 , Figure 2 It shows the resolution differences of riverbed sediment photos taken by drones at different flight altitudes, intuitively reflects the impact of flight altitude on image resolution, and provides actual image data support for the drawing and fitting of scatter plots.

[0053] Based on the above embodiment, the present invention provides a detailed description of step S102:

[0054] In a specific embodiment, a power fitting is performed on the scatter plot to obtain a fitting relationship between the UAV flight altitude and the image resolution:

[0055] R=0.1·H 1.15

[0056] This fitting equation reveals the quantitative relationship between flight altitude and image resolution, providing a basis for subsequent determination of the appropriate flight altitude.

[0057] Based on the above embodiment, the present invention provides a detailed description of step S103:

[0058] According to the Nyquist sampling theorem, if you need to distinguish a sediment particle, the sample length of the particle in the one-dimensional signal needs to be 2. Since RGB images are three-dimensional signals, the pixel length required for the human eye to distinguish a sediment particle is 2+1+1=4 pixels. Based on a lot of experience in manually identifying sediment images, we can also get 4 pixels as the minimum sediment particle size pixel value to be identified from the image, such as Figure 3 As shown. Therefore, 4 pixels are used as the first minimum sediment particle size pixel value that can be identified from the image. In summary, if it is necessary to identify millimeter-level sediment particles, the resolution is required to be at least 10÷4=2.5mm / pixel. Further, according to R=0.1·H 1.15 The relationship requires a maximum flight altitude (i.e., the maximum flight altitude of the first drone) of about 16 meters. Therefore, in open areas with conditions for drones to fly low, sampling should be carried out at a flight altitude below 16 meters.

[0059] Based on the above embodiments, the image-based riverbed sediment measurement method of the present invention further includes:

[0060] S105: Based on the fitting relationship, collect a second riverbed sediment image sample with a resolution higher than a preset threshold, and perform n-fold upsampling on the second riverbed sediment image sample to obtain a third riverbed sediment sample to construct a training set;

[0061] S106: Using the third riverbed sediment sample as input and the second riverbed sediment image sample as output, constructing a sediment image super-resolution model;

[0062] S107: Training the sediment image super-resolution model based on a deep learning algorithm to obtain a trained sediment image super-resolution model;

[0063] S108: Testing and analyzing the trained sediment image super-resolution model to determine the optimal upsampling factor and obtain a target sediment image super-resolution model;

[0064] S109: Determine the resolution corresponding to the second minimum sediment particle size pixel value that can be identified in the super-resolution image processed by the target sediment image super-resolution model, and obtain the maximum flight altitude of the second UAV according to the fitting relationship;

[0065] S110: When the drone does not have a low-flying condition, obtain a second target riverbed sediment image sampled by the drone at a lower altitude than the second drone's maximum flight altitude;

[0066] S111: Inputting the second target riverbed sediment image into a target sediment image super-resolution model to obtain a target super-resolution riverbed sediment image, and performing riverbed sediment particle size detection on the target super-resolution riverbed sediment image.

[0067] Based on the above embodiment, this embodiment describes step S105 in detail:

[0068] In a specific embodiment, Figure 3 As shown, based on the fitting relationship of step S102, a large number of original high-resolution riverbed sediment photo datasets with a resolution of X (i.e., the second riverbed sediment image sample) are collected. The dataset needs to cover different river slopes, different riverbed morphology types, different river widths, different lithologies, and different sediment gradations. The drone flight altitude covers 10 meters to 200 meters to ensure data diversity and representativeness.

[0069] In a specific embodiment, a nearest neighbor sampling algorithm is used to sample the original high-resolution photo at a resolution of X / n by reducing the resolution by a factor of n to obtain a resampled low-resolution photo dataset (ie, the third riverbed sediment image sample).

[0070] Based on the above embodiment, this embodiment describes step S107 in detail:

[0071] Train a deep learning sediment image super-resolution model. Model algorithms may include SRCNN, SRGAN, etc. These algorithms can learn the mapping relationship from low-resolution images to high-resolution images, thereby improving image resolution.

[0072] Based on the above embodiment, this embodiment describes step S108 in detail:

[0073] In a specific embodiment, the optimal upsampling factor is 4 times.

[0074] Based on the above embodiment, this embodiment describes step S109 in detail:

[0075] In a specific embodiment, the human eye can recognize most of the 1-pixel sediment particles in the image processed by the super-resolution model, such as Figure 4 This means that after super-resolution processing, the image resolution has been significantly improved, which can better identify fine sediment particles. If millimeter-level sediment needs to be identified, the resolution must be at least 10÷1=10mm / pixel. 1.15Relationship, the maximum flight altitude of the second drone is required to be about 55 meters.

[0076] In summary, if low-flight conditions (less than 16 meters) are unavailable, riverbed sediment should first be sampled at an altitude below 55 meters. The sampled samples are then fed into a sediment image super-resolution model. Finally, millimeter-sized sediment particles on the bed surface can be identified from the output image of the sediment image super-resolution model. This solution provides a solution for situations where low-flight sampling is impossible in practical applications, compensating for the insufficient resolution caused by flight altitude restrictions through super-resolution technology.

[0077] Based on the above embodiments, the image-based riverbed sediment measurement of the present invention determines the resolution corresponding to the first minimum sediment particle size pixel value that can be identified in the original resolution image based on the fitting relationship between the drone flight altitude and the image resolution, obtains the first drone's maximum flight altitude, obtains the first target riverbed sediment image sampled by the drone at a lower altitude than the first drone's maximum flight altitude when the drone is in low flight, and performs riverbed sediment particle size detection; when the drone is not in low flight, determines the resolution corresponding to the second minimum sediment particle size pixel value that can be identified in the super-resolution image processed by the target sediment image super-resolution model, and obtains the second drone's maximum flight altitude based on the fitting relationship, obtains the second target riverbed sediment image sampled by the drone at a lower altitude than the second drone's maximum flight altitude, inputs the second target riverbed sediment image into the target sediment image super-resolution model to obtain a target super-resolution riverbed sediment image, and performs riverbed sediment particle size detection on the target super-resolution riverbed sediment image. The present invention effectively improves the accuracy of riverbed sediment measurement.

[0078] The embodiment of the present invention further provides an image-based riverbed sediment measurement device; the specific device may include:

[0079] A scatter plot construction module is used to obtain the first riverbed sediment image samples sampled by the drone at different flight altitudes and to create a scatter plot of the drone's flight altitude and image resolution;

[0080] A fitting relationship building module is used to fit the scatter plot to obtain a fitting relationship between the UAV flight altitude and the image resolution;

[0081] a first UAV maximum flight altitude determination module, configured to determine a resolution corresponding to a first minimum sediment particle size pixel value that can be identified in the original resolution image, and obtain the first UAV maximum flight altitude based on the fitting relationship;

[0082] The sediment measurement module is used to obtain the first target riverbed sediment image sampled by the drone at a lower altitude than the first drone's maximum flight altitude under the condition of low-flying drone, and to detect the riverbed sediment particle size.

[0083] The image-based riverbed sediment measurement device of this embodiment is used to implement the aforementioned image-based riverbed sediment measurement method. Therefore, the specific implementation methods of the image-based riverbed sediment measurement device can be seen in the embodiment part of the image-based riverbed sediment measurement method mentioned above. For example, the scatter plot construction module, the fitting relationship construction module, the first UAV maximum flight altitude determination module, and the sediment measurement module are respectively used to implement steps S101, S102, S103, and S104 in the above-mentioned image-based riverbed sediment measurement method. Therefore, its specific implementation methods can refer to the descriptions of the corresponding embodiments of each part and will not be repeated here.

[0084] Based on the above embodiments, the image-based riverbed sediment measurement device of the present invention further includes:

[0085] The training set module is used to collect second riverbed sediment image samples with a resolution higher than a preset threshold based on the fitting relationship, and perform n-fold upsampling on the second riverbed sediment image samples to obtain third riverbed sediment samples to construct a training set;

[0086] a model building module, configured to build a sediment image super-resolution model using the third riverbed sediment sample as input and the second riverbed sediment image sample as output;

[0087] A model training module is used to train the sediment image super-resolution model based on a deep learning algorithm to obtain a trained sediment image super-resolution model;

[0088] A target model determination module is used to test and analyze the trained sediment image super-resolution model, determine the optimal upsampling multiple, and obtain a target sediment image super-resolution model;

[0089] a second UAV maximum flight altitude determination module, configured to determine a resolution corresponding to a second minimum sediment particle size pixel value that can be identified in the super-resolution image processed by the target sediment image super-resolution model, and to obtain the second UAV maximum flight altitude based on the fitting relationship;

[0090] The sediment measurement module is also used to:

[0091] In the absence of low-flying drone conditions, a second target riverbed sediment image sampled by the drone at a lower altitude than the second drone's maximum flight altitude is obtained, the second target riverbed sediment image is input into the target sediment image super-resolution model to obtain a target super-resolution riverbed sediment image, and riverbed sediment particle size detection is performed on the target super-resolution riverbed sediment image.

[0092] A specific embodiment of the present invention further provides an image-based riverbed sediment measurement device, comprising: a memory for storing a computer program; and a processor for implementing the steps of the above-mentioned image-based riverbed sediment measurement method when executing the computer program.

[0093] A specific embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the above-mentioned image-based riverbed sediment measurement method are implemented.

[0094] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0095] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0096] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0097] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0098] Obviously, the above embodiments are merely examples for clarity of explanation and are not intended to limit the implementation methods. Those skilled in the art will appreciate that other variations or modifications can be made based on the above description. It is not necessary and impossible to enumerate all implementation methods here. Obvious variations or modifications arising therefrom remain within the scope of protection of the present invention.

Claims

1. A riverbed sediment measurement method based on image, characterized in that: include: Obtain the first riverbed sediment image samples taken by the drone at different flight altitudes, and create a scatter plot of the drone's flight altitude and image resolution; Fitting the scatter plot to obtain a fitting relationship between the UAV flight altitude and the image resolution; Determine the resolution corresponding to the first minimum sediment particle size pixel value that can be identified in the original resolution image, and obtain the maximum flight altitude of the first UAV based on the fitting relationship; Under the condition of low-flying drone, obtain the first target riverbed sediment image sampled by the drone at a lower altitude than the first drone's maximum flight altitude, and perform riverbed sediment particle size detection.

2. The image-based riverbed sediment measurement method according to claim 1, characterized in that: The fitting relationship is: R=0.1·H 1.15 Among them, H is the flight altitude of the drone and R is the image resolution.

3. The image-based riverbed sediment measurement method according to claim 2, characterized in that: The first minimum sediment particle size pixel value is 4.

4. The image-based riverbed sediment measurement method according to claim 2, characterized in that: Also includes: Based on the fitting relationship, a second riverbed sediment image sample with a resolution higher than a preset threshold is collected, and the image sample is upsampled n times to obtain a third riverbed sediment sample to construct a training set; Taking the third riverbed sediment sample as input and the second riverbed sediment image sample as output, constructing a sediment image super-resolution model; Training the sediment image super-resolution model based on a deep learning algorithm to obtain a trained sediment image super-resolution model; Testing and analyzing the trained sediment image super-resolution model to determine the optimal upsampling factor and obtain a target sediment image super-resolution model; Determining the resolution corresponding to the second minimum sediment particle size pixel value that can be identified in the super-resolution image processed by the target sediment image super-resolution model, and obtaining the maximum flight altitude of the second UAV based on the fitting relationship; In the absence of low-flying UAV conditions, obtain the second target riverbed sediment image sampled by the UAV at a lower altitude than the second UAV's maximum flight altitude; The second target riverbed sediment image is input into the target sediment image super-resolution model to obtain a target super-resolution riverbed sediment image, and riverbed sediment particle size detection is performed on the target super-resolution riverbed sediment image.

5. The image-based riverbed sediment measurement method according to claim 4, characterized in that: The up-sampling adopts the nearest neighbor sampling algorithm.

6. The image-based riverbed sediment measurement method according to claim 4 or 5, characterized in that: The optimal upsampling factor is 4.

7. The image-based riverbed sediment measurement method according to claim 6, characterized in that: The second minimum sediment particle size pixel value is 1.

8. An image-based riverbed sediment measurement device, characterized in that: include: A scatter plot construction module is used to obtain the first riverbed sediment image samples sampled by the drone at different flight altitudes and to create a scatter plot of the drone's flight altitude and image resolution; A fitting relationship building module is used to fit the scatter plot to obtain a fitting relationship between the UAV flight altitude and the image resolution; a first UAV maximum flight altitude determination module, configured to determine a resolution corresponding to a first minimum sediment particle size pixel value that can be identified in the original resolution image, and obtain the first UAV maximum flight altitude based on the fitting relationship; The sediment measurement module is used to obtain the first target riverbed sediment image sampled by the drone at a lower altitude than the first drone's maximum flight altitude under the condition of low-flying drone, and to detect the riverbed sediment particle size.

9. An image-based riverbed sediment measurement device, characterized in that: include: Memory for storing computer programs; A processor is configured to implement the steps of the image-based riverbed sediment measurement method as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the image-based riverbed sediment measurement method according to any one of claims 1 to 7 are implemented.

Images