CN113743227A - Rock mass fracture intelligent identification method and system based on proximity photography - Google Patents
Rock mass fracture intelligent identification method and system based on proximity photography Download PDFInfo
- Publication number
- CN113743227A CN113743227A CN202110901662.6A CN202110901662A CN113743227A CN 113743227 A CN113743227 A CN 113743227A CN 202110901662 A CN202110901662 A CN 202110901662A CN 113743227 A CN113743227 A CN 113743227A
- Authority
- CN
- China
- Prior art keywords
- rock mass
- image
- fracture
- aerial vehicle
- unmanned aerial
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 239000011435 rock Substances 0.000 title claims abstract description 108
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000007781 pre-processing Methods 0.000 claims abstract description 20
- 238000013459 approach Methods 0.000 claims abstract description 16
- 238000003709 image segmentation Methods 0.000 claims abstract description 15
- 238000003062 neural network model Methods 0.000 claims description 14
- 238000001914 filtration Methods 0.000 claims description 7
- 230000009466 transformation Effects 0.000 claims description 7
- 238000011176 pooling Methods 0.000 claims description 5
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 230000000007 visual effect Effects 0.000 claims description 2
- 238000013528 artificial neural network Methods 0.000 abstract description 4
- RZVHIXYEVGDQDX-UHFFFAOYSA-N 9,10-anthraquinone Chemical compound C1=CC=C2C(=O)C3=CC=CC=C3C(=O)C2=C1 RZVHIXYEVGDQDX-UHFFFAOYSA-N 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000011218 segmentation Effects 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 2
- 238000005260 corrosion Methods 0.000 description 2
- 230000007797 corrosion Effects 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000012535 impurity Substances 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000008961 swelling Effects 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The utility model provides a rock mass fracture intelligent identification method and system based on an approximation photography, comprising: acquiring topographic data of a region to be identified of a rock mass fracture by using an unmanned aerial vehicle, and acquiring a surface plane model by fitting the topographic data; generating a three-dimensional route based on the obtained ground surface plane model; controlling the unmanned aerial vehicle to automatically approach the surface of the rock body to fly through the three-dimensional air route, and shooting images according to a preset position; carrying out data preprocessing on the shot image, and carrying out image segmentation on the preprocessed image; and inputting the segmented image into a pre-trained rock mass fracture identification model to obtain the identification result of the rock mass fracture. According to the scheme, the unmanned aerial vehicle is tightly attached to the surface of the rock body to make a video recording, so that a fracture image is automatically obtained, accurate identification of the rock body fracture is achieved through the LinkNet neural network, and meanwhile, the fracture identification efficiency is effectively improved.
Description
Technical Field
The disclosure belongs to the technical field of rock mass fracture identification, and particularly relates to a rock mass fracture intelligent identification method and system based on approximation photography.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
Rock fracture identification is the basis for realizing rock fracture network modeling, and the occurrence data of the rock can be obtained through the fracture identification to know the structural characteristics of the rock. For example: rock mass fracture identification is carried out on the slope zone, a slope deformation mechanism can be judged, and disasters such as slope landslide and landslide are predicted; the crack identification is carried out on the exposed ore deposit, the crack network modeling is further carried out, the simulation prediction is carried out on disasters such as water burst, the occurrence of the disasters is reduced, and the safety of constructors is ensured; the method can be used for identifying the fractures of the field exposed rock mass and providing design parameters for the subsequent construction of the engineering. Therefore, rock mass fracture identification is very important.
The traditional crack identification method mainly depends on manual work to measure and record rock cracks one by means of a compass and a ruler, is simple to operate, but consumes time, is low in efficiency, and is greatly influenced by natural conditions such as weather and geographical environment. The method mainly utilizes methods such as threshold segmentation, edge detection, region growth and the like to carry out binarization processing on an original image, namely, a processed rock image is a black-white image, and then fracture identification is carried out. Therefore, a rock mass fracture identification method which has relatively high identification precision, can liberate labor force and overcome geographical condition limitation is urgently needed.
Disclosure of Invention
In order to solve the problems, the invention provides a rock mass fracture intelligent identification method and system based on approach photography.
According to a first aspect of the disclosed embodiments, there is provided a rock mass fracture intelligent identification method based on proximity photography, including:
acquiring topographic data of a region to be identified of a rock mass fracture by using an unmanned aerial vehicle, and acquiring a surface plane model by fitting the topographic data;
generating a three-dimensional route based on the obtained ground surface plane model;
controlling the unmanned aerial vehicle to automatically approach the surface of the rock body to fly through the three-dimensional air route, and shooting images according to a preset position;
carrying out data preprocessing on the shot image, and carrying out image segmentation on the preprocessed image;
and inputting the segmented image into a pre-trained rock mass fracture identification model to obtain the identification result of the rock mass fracture.
Further, fitting the surface plane of the area to be identified of the rock mass fracture by using the unmanned aerial vehicle specifically comprises: controlling the unmanned aerial vehicle to enter a rock mass fracture area to be identified and fly close to the ground for a preset distance, so as to obtain topographic data; and fitting a ground surface plane model through route planning software based on the acquired topographic data.
Further, the image shooting according to the preset position specifically includes:
after the shooting position is determined, vertical shooting and left-right deflection preset angles are performed by automatically adjusting the shooting equipment to be vertical to the fitted terrain plane.
Further, the data preprocessing of the captured image specifically includes: carrying out geometric transformation and rotation scaling on the image to unify the scale of the shot image; removing noise in the shot image through image filtering; the visual effect of the shot image is improved through the image enhancement operation.
Furthermore, the rock mass fracture identification model adopts a Link-Net neural network model, and the neural network model comprises a convolution layer, a pooling layer, a coding block, a decoding block and a deconvolution layer which are sequentially connected.
According to a second aspect of the embodiments of the present disclosure, there is provided a rock mass fracture intelligent identification system based on an approach photography, including:
the terrain data acquisition unit is used for acquiring the terrain data of the area to be identified of the rock mass fracture by using the unmanned aerial vehicle and acquiring a ground surface plane model by fitting the terrain data;
the three-dimensional route generation unit is used for generating a three-dimensional route based on the acquired ground surface plane model;
the approach photography unit is used for controlling the unmanned aerial vehicle to automatically approach the surface of the rock body to fly through the three-dimensional route and shooting images according to a preset position;
the image preprocessing unit is used for carrying out data preprocessing on the shot image and carrying out image segmentation on the preprocessed image;
and the fracture identification unit is used for inputting the segmented image into a pre-trained rock mass fracture identification neural network model to obtain the identification result of the rock mass fracture.
Compared with the prior art, the beneficial effect of this disclosure is:
(1) the scheme of the invention provides an intelligent rock mass fracture identification method based on the approach photography, and the method comprises the steps of carrying out camera shooting by using an unmanned aerial vehicle to cling to the surface of a rock mass, thereby automatically obtaining fracture images; the constructed Link-Net neural network model is used for identifying the rock mass fracture, and the rock mass fracture image realizes intelligent identification through self circulation, self feedback and self learning, so that the accuracy of identifying the rock mass fracture is improved;
(2) compared with the traditional network model, the Link-Net neural network model in the scheme disclosed by the disclosure adopts a Resnet18 encoder, and simultaneously, the encoder is directly connected with the decoder, so that the spatial dimension of image data is not reduced, the characteristic information of a rock fracture image can be more effectively extracted, and meanwhile, the scheme is very effective for reducing the phenomena of gradient explosion and the like in the network by adding a ReLU activation function and a hyperparametric loss function;
(3) according to the scheme, the unmanned aerial vehicle is used for close shooting, automatic flight and automatic image data acquisition of the unmanned aerial vehicle can be realized by means of three-dimensional route planning, and an image self-learning network is added, so that intelligent identification of rock mass cracks is really realized, and the automation degree is high;
(4) according to the scheme, the unmanned aerial vehicle carries a camera, an image preprocessing process and a crack identification method based on deep learning, so that automatic identification of rock mass cracks can be realized; labor force is liberated, and crack identification efficiency is improved; meanwhile, crack identification can be carried out on areas which cannot be reached by technicians, the limitation of geographical conditions is overcome, and the danger of the technicians is reduced.
Advantages of additional aspects of the disclosure will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the disclosure.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure and are not to limit the disclosure.
FIG. 1 is a flowchart of an approach aerial photography process according to a first embodiment of the present disclosure;
FIG. 2 is a flowchart illustrating image preprocessing according to a first embodiment of the disclosure;
FIG. 3 is a diagram of a Link-Net neural network for fracture identification according to a first embodiment of the disclosure;
FIG. 4 is a schematic diagram of a rock mass fracture extraction process based on deep learning according to a first embodiment of the disclosure;
FIG. 5(a) is a schematic structural diagram of a Link-Net neural network encoder according to a first embodiment of the disclosure;
fig. 5(b) is a schematic structural diagram of a Link-Net neural network decoder according to the first embodiment of the disclosure.
Detailed Description
The present disclosure is further described with reference to the following drawings and examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present disclosure. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
The embodiments and features of the embodiments in the present disclosure may be combined with each other without conflict.
The first embodiment is as follows:
the embodiment aims to provide an intelligent rock mass fracture identification method based on the approximating photography.
A rock mass fracture intelligent identification method based on proximity photography comprises the following steps:
acquiring topographic data of a region to be identified of a rock mass fracture by using an unmanned aerial vehicle, and acquiring a surface plane model by fitting the topographic data;
generating a three-dimensional route based on the obtained ground surface plane model;
controlling the unmanned aerial vehicle to automatically approach the surface of the rock body to fly through the three-dimensional air route, and shooting images according to a preset position;
carrying out data preprocessing on the shot image, and carrying out image segmentation on the preprocessed image;
and inputting the segmented image into a pre-trained rock mass fracture identification model to obtain the identification result of the rock mass fracture.
Specifically, for ease of understanding, the embodiments of the present disclosure are described in detail below with reference to the accompanying drawings:
step 1: acquiring topographic data of a region to be identified of a rock mass fracture by using an unmanned aerial vehicle, and acquiring a surface plane model by fitting the topographic data; generating a three-dimensional route based on the obtained ground surface plane model;
firstly, the intelligent rock mass fracture identification method disclosed by the disclosure needs to prepare an unmanned aerial vehicle provided with a camera system, a course planning software, a flight control system and an image storage system.
Unmanned aerial vehicle can adopt rotor unmanned aerial vehicle, and rotor unmanned aerial vehicle compares in other unmanned aerial vehicle comparatively steady, can satisfy the requirement of photographic image high definition. The camera system is arranged below the rotor unmanned aerial vehicle, and the flight path planning software (in the embodiment, Pix4d aerial survey software is adopted, and is software for generating an orthoimage of the unmanned aerial vehicle through three-dimensional reconstruction, is installed in the unmanned aerial vehicle system and depends on image acquisition equipment and other sensor settings on the unmanned aerial vehicle to realize the functions in the step 1, has the functions of data acquisition and the like, and can complete corresponding flight according to specific software requirements), and a flight control system (namely, utilizes a control system of the unmanned aerial vehicle and interacts with the flight path planning software) is arranged at the position of the aircraft body. The camera system, the air route planning software and the flight control system are connected through Bluetooth, pictures and pictures shot by the camera system can be transmitted to the air route planning software, and the air route software can plan routes generated according to the pictures and can also be transmitted to the unmanned aerial vehicle flight control system to control the unmanned aerial vehicle to fly.
Step 2: controlling the unmanned aerial vehicle to automatically approach the surface of the rock body to fly through the three-dimensional route, and shooting images according to a preset period;
the main process is that when a rotor unmanned aerial vehicle flies for the first time through a camera system, the rough terrain of a rock mass fracture recognition area is obtained, a plane which is 3-15 m away from the ground surface plane is roughly fitted by means of flight path planning software, the three-dimensional flight path planning is further generated by the flight path planning software, and the unmanned aerial vehicle can fly automatically without manual operation. When the shooting target is determined, the unmanned aerial vehicle does not need to land, and the camera automatically and perpendicularly shoots a picture on a fitting plane and deflects a preset angle left and right to shoot a picture. The camera system is provided with a flash lamp, so that the camera system can still shoot in rainy days or in shadow with poor light, and a clear picture can be obtained.
And step 3: carrying out data preprocessing on the shot image, and carrying out image segmentation on the preprocessed image;
the shot picture can not be directly used for image fracture identification, the shot rock mass fracture image needs to be subjected to image preprocessing operation, the processing process is as shown in figure 2, and the image is subjected to preprocessing measures such as geometric transformation, rotation scaling, image filtering, image enhancement and the like in sequence.
(1) Geometric transformation
Determining the geometric scale proportion of the image and the real object through geometric transformation, and geometrically transforming the input image according to the following formula:
when the three-dimensional route planning is carried out, a rough plan view of a shooting area is obtained, corresponding position coordinates of each shooting object are obtained, X, Y are transverse and longitudinal axis coordinates corresponding to a real rock mass and can be obtained in flight control software, X 'and Y' shooting center coordinates can also be obtained in the flight control software, and a is a coefficient corresponding to the focal length of a camera lens. The geometric proportion of the image and the real object can be obtained through the formula, and the subsequent statistics is facilitated to obtain the geometric parameters of the rock mass fracture.
(2) Rotational zoom
The second process of preprocessing is rotation, flipping, scaling, etc. When a training set is acquired, in order to acquire a large number of data sets, the operation is performed by rotating and scaling the data sets to regularize the model, so that the image types in the small data sets are balanced, a representative use case is generated, and the data sets are also expanded. After the above-mentioned rock mass fracture picture is treated,
(3) filtering and enhancement
The image filtering and the image enhancement are carried out, the image filtering is carried out to inhibit the noise in the image target under the condition of ensuring the picture details as much as possible, a filter consisting of mathematical morphology basic operation is adopted to carry out the operation, the structure information of the object can be obtained, and the essential form of the object can be visually obtained through some operation between the object and the structure elements. Mainly comprises 4 basic operations: swelling, corrosion, opening, closing. The image can be enlarged by means of expansion, and unconnected cracks in the picture are bridged; by corroding and shrinking the image, eliminating the impurity points smaller than the structural elements and the like; the open and close operations smooth the picture. The morphological filter may be implemented by Matlab coding. Picture enhancement to enhance the realism of pictures taken under severe conditions. And (3) sharpening the picture by using a Laplace operator, so that the part of the rock fracture picture edge jumping is enhanced, the picture becomes clearer, and a sharpening template is shown as follows, wherein alpha is a sharpening strength coefficient, and the larger the alpha is, the stronger the sharpening degree is.
(4) Image segmentation
The image segmentation adopts a K-means segmentation algorithm, and analysis and research are carried out in order to extract target information from background information. The image segmentation directly influences the rock mass fracture identification and fracture characteristic information extraction in the next step. The principle of the K-means segmentation algorithm is that a part of points are selected as the centers of initial clustering points, the distances between other points and the part of points are calculated, a sample point closest to a certain initial point and the initial point are classified into one class, the average value of each class is calculated, the average value is used as the center of a new sample point, and classification is finished until the sample center point is calculated to be not changed for two times continuously. The classification formula is as follows:
wherein: l isijInto class iJth object, miIs the clustering center of class i, NiIs the number of i types.
And 4, step 4: and inputting the segmented image into a pre-trained rock mass fracture identification model to obtain the identification result of the rock mass fracture. The rock mass fracture identification model adopts a Link-Net neural network model, and the neural network model comprises a convolution layer, a pooling layer, a coding block, a decoding block and a deconvolution layer which are sequentially connected.
Specifically, the preprocessed image is imported into a Link-Net image semantic segmentation network for training, and the neural network model structure is shown in FIG. 3. The coding block i output information is added with the decoding block i +1 output information by the direct connection of the coder and the decoder, and the addition result is used as the input of the decoding block i-1 layer. The encoder and the decoder can enjoy the fracture information extracted from the target rock mass at the same time, so that the time for relearning is saved, and the efficiency is improved. The left half part is an encoder, the encoder extracts fracture characteristic information of a rock mass, the encoder can meet fracture identification accuracy for ResNet18, information execution efficiency can be guaranteed, the number of encoder layers is set to be 4, the encoder starts from an initial block, image data with the input image size of 7 × 7 and the step size of 2 is convolved, and the block performs space maximum pool operation with the step size of 2 in a 3 × 3 region. Each encoder structure is shown in fig. 5 (a). The decoder restores the feature information and spatial dimension of the image through a deconvolution operation, as shown in fig. 5(b), which can reduce the loss of spatial information between the encoder and the decoder. The decoder corresponds to the encoder and is also a 4-layer decoding block, the output of the coding block is the input of the decoding block, and the complete convolution layer is expressed as (k × k) (im, om) to represent the size of the kernel, the input size and the output size. Fig. 3 x 3(64,32) shows that the kernel size is 3 x 3, the input size 64, and the output size 32.
Further, the concrete steps of identifying the rock mass fracture by using the identification method are as follows:
(1) based on the flight control system arranged on the unmanned aerial vehicle, the unmanned aerial vehicle is manually controlled to reach an area needing rock crack identification, a camera is opened for recording, the unmanned aerial vehicle flies for a circle 3-15 m away from a rock plane, and course planning software is used for fitting the rock plane.
(2) According to the fitted rock mass plane, selecting a proper distance (set according to actual requirements) between the vertical distances of 3-15 m, and generating a three-dimensional course by course planning software.
(3) And determining a fixed point position of the rock mass fracture to be shot according to the video, guiding the three-dimensional route into an unmanned aerial vehicle flight control system through route planning software, and guiding the unmanned aerial vehicle to reach the rock mass to be subjected to fracture identification.
(4) And opening the camera to shoot, firstly shooting the image perpendicular to the surface of the rock mass, then shooting the image by deflecting a certain angle, and transmitting the shot rock mass fracture image to an image storage system to realize the construction of an image data set.
(5) Preprocessing and image segmentation operations are carried out on images in the data set, geometric transformation is carried out on the images by means of rock mass position coordinates, and geometric parameters of rock mass cracks are determined; and the image after geometric transformation needs image filtering, and comprises four steps of expansion, corrosion, opening and closing, so that the miscellaneous points in the image are removed, and the image crack is bridged, so that the image is clearer. Performing image enhancement treatment, and performing sharp treatment on the rock fracture edge; performing the last step of image preprocessing, and rotating and scaling the image; and finally, carrying out image segmentation according to a K-means algorithm, and carrying out data set expansion.
(6) A Link-Net neural network model is adopted to construct a rock mass fracture identification model, and the neural network model comprises a convolution layer, a pooling layer, a coding block, a decoding block and a deconvolution layer which are sequentially connected. When the model is used for the first time, a data set is divided into a training set and a testing set, the training set is put into a Link-net network model, network parameters are initialized, the model is trained, the model is continuously adjusted, and the optimal model is determined. (data set needs manual annotation of fracture in advance, and is convenient for network model learning)
(7) Model parameters are fine-tuned through a test set.
(8) After the model is trained, inputting collected rock mass fracture data into a Link-Net fracture network, and performing rock mass fracture identification through a convolution layer, a maximum pooling layer, a coding block, a decoding block and a deconvolution layer.
Example two:
the purpose of this embodiment is to provide a rock mass crack intelligent recognition system based on it is photographic to cut close.
A rock mass crack intelligent identification system based on approach photography comprises:
the terrain data acquisition unit is used for acquiring the terrain data of the area to be identified of the rock mass fracture by using the unmanned aerial vehicle and acquiring a ground surface plane model through fitting;
the three-dimensional route generation unit is used for generating a three-dimensional route based on the acquired ground surface plane model;
the approach photography unit is used for controlling the unmanned aerial vehicle to automatically approach the surface of the rock body to fly through the three-dimensional route and shooting images according to a preset position;
the image preprocessing unit is used for carrying out data preprocessing on the shot image and carrying out image segmentation on the preprocessed image;
and the fracture identification unit is used for inputting the segmented image into a pre-trained rock mass fracture identification neural network model to obtain the identification result of the rock mass fracture.
The rock mass fracture intelligent identification method and system based on the approximating photography can be achieved, and have wide application prospects.
The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure, and various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.
Claims (10)
1. A rock mass fracture intelligent identification method based on an approximation photography is characterized by comprising the following steps:
acquiring topographic data of a region to be identified of a rock mass fracture by using an unmanned aerial vehicle, and acquiring a surface plane model by fitting the topographic data;
generating a three-dimensional route based on the obtained ground surface plane model;
controlling the unmanned aerial vehicle to automatically approach the surface of the rock body to fly through the three-dimensional air route, and shooting images according to a preset position;
carrying out data preprocessing on the shot image, and carrying out image segmentation on the preprocessed image;
and inputting the segmented image into a pre-trained rock mass fracture identification model to obtain the identification result of the rock mass fracture.
2. The method for intelligently identifying the rock mass fracture based on the approximating photography as claimed in claim 1, wherein the fitting of the earth surface plane of the region to be identified of the rock mass fracture by the unmanned aerial vehicle is specifically as follows: controlling the unmanned aerial vehicle to enter a rock mass fracture area to be identified and fly close to the ground for a preset distance, so as to obtain topographic data; and fitting a ground surface plane model through route planning software based on the acquired topographic data.
3. The method for intelligently identifying rock mass fractures based on the approximation photography as claimed in claim 1, wherein the three-dimensional route is generated based on the acquired topographic data, and specifically comprises the following steps: and generating a three-dimensional air route for the unmanned aerial vehicle to fly by utilizing air route planning software based on the obtained ground surface plane model.
4. The method for intelligently identifying rock mass fractures based on the approximating photography as claimed in claim 1, wherein the image shooting is carried out according to a preset position, specifically:
after the shooting position is determined, vertical shooting and left-right deflection preset angles are performed by automatically adjusting the shooting equipment to be vertical to the fitted terrain plane.
5. The method for intelligently identifying rock mass fractures based on the approximating photography as claimed in claim 1, wherein the data preprocessing is performed on the shot images, and specifically comprises the following steps: carrying out geometric transformation and rotation scaling on the image to unify the scale of the shot image; removing noise in the shot image through image filtering; the visual effect of the shot image is improved through the image enhancement operation.
6. The method for intelligently identifying rock mass fractures based on the approximating photography as claimed in claim 1, wherein the preprocessed image is subjected to image segmentation, specifically to image segmentation by using a k-means algorithm.
7. The method for intelligently identifying the rock mass fracture based on the approximating photography as claimed in claim 1, wherein the rock mass fracture identification model adopts a Link-Net neural network model, and the neural network model comprises a convolutional layer, a pooling layer, a coding block, a decoding block and an anti-convolutional layer which are connected in sequence.
8. The utility model provides a rock mass crack intelligent recognition system based on it is photographic to cut close which characterized in that includes:
the terrain data acquisition unit is used for acquiring the terrain data of the area to be identified of the rock mass fracture by using the unmanned aerial vehicle and acquiring a ground surface plane model by fitting the terrain data;
the three-dimensional route generation unit is used for generating a three-dimensional route based on the acquired ground surface plane model;
the approach photography unit is used for controlling the unmanned aerial vehicle to automatically approach the surface of the rock body to fly through the three-dimensional route and shooting images according to a preset position;
the image preprocessing unit is used for carrying out data preprocessing on the shot image and carrying out image segmentation on the preprocessed image;
and the fracture identification unit is used for inputting the segmented image into a pre-trained rock mass fracture identification neural network model to obtain the identification result of the rock mass fracture.
9. The system for intelligently identifying rock mass fractures based on the approximating photography as claimed in claim 8, wherein the unmanned aerial vehicle carries a camera device, and the camera device is used for realizing multi-degree-of-freedom adjustment by means of a holder module.
10. The system for intelligently identifying rock mass fractures based on the approximating photography as claimed in claim 8, wherein the camera device is provided with a flash lamp.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110901662.6A CN113743227A (en) | 2021-08-06 | 2021-08-06 | Rock mass fracture intelligent identification method and system based on proximity photography |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110901662.6A CN113743227A (en) | 2021-08-06 | 2021-08-06 | Rock mass fracture intelligent identification method and system based on proximity photography |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113743227A true CN113743227A (en) | 2021-12-03 |
Family
ID=78730289
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110901662.6A Pending CN113743227A (en) | 2021-08-06 | 2021-08-06 | Rock mass fracture intelligent identification method and system based on proximity photography |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113743227A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114419088A (en) * | 2022-01-24 | 2022-04-29 | 重庆中环建设有限公司 | Tunnel surrounding rock information extraction method based on digital image |
| CN115147615A (en) * | 2022-07-01 | 2022-10-04 | 河海大学 | Rock image classification method and device based on metric learning network |
| CN115217084A (en) * | 2022-07-22 | 2022-10-21 | 中国华能集团清洁能源技术研究院有限公司 | Reservoir area expansive soil surface crack rate detection method and system |
| CN115546113A (en) * | 2022-09-15 | 2022-12-30 | 山东大学 | Method and system for predicting parameters of tunnel face crack image and front three-dimensional structure |
| CN115861848A (en) * | 2023-03-01 | 2023-03-28 | 成都理工大学 | Method and device for processing rock mass image |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110110661A (en) * | 2019-05-07 | 2019-08-09 | 西南石油大学 | A kind of rock image porosity type recognition methods based on unet segmentation |
| CN111442762A (en) * | 2020-04-27 | 2020-07-24 | 长江空间信息技术工程有限公司(武汉) | Unmanned aerial vehicle close photography method facing hydraulic engineering |
| CN111649723A (en) * | 2020-05-29 | 2020-09-11 | 武汉大学 | Complex terrain-oriented close-up photography three-dimensional track and attitude planning method and device |
| CN111861982A (en) * | 2020-06-07 | 2020-10-30 | 中国葛洲坝集团第一工程有限公司 | Visual image monitoring and identifying system |
| CN112529901A (en) * | 2020-12-31 | 2021-03-19 | 江西飞尚科技有限公司 | Crack identification method in complex environment |
| CN113012111A (en) * | 2021-02-25 | 2021-06-22 | 北京麦飞科技有限公司 | Rice lodging degree grading monitoring method based on edge calculation |
-
2021
- 2021-08-06 CN CN202110901662.6A patent/CN113743227A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110110661A (en) * | 2019-05-07 | 2019-08-09 | 西南石油大学 | A kind of rock image porosity type recognition methods based on unet segmentation |
| CN111442762A (en) * | 2020-04-27 | 2020-07-24 | 长江空间信息技术工程有限公司(武汉) | Unmanned aerial vehicle close photography method facing hydraulic engineering |
| CN111649723A (en) * | 2020-05-29 | 2020-09-11 | 武汉大学 | Complex terrain-oriented close-up photography three-dimensional track and attitude planning method and device |
| CN111861982A (en) * | 2020-06-07 | 2020-10-30 | 中国葛洲坝集团第一工程有限公司 | Visual image monitoring and identifying system |
| CN112529901A (en) * | 2020-12-31 | 2021-03-19 | 江西飞尚科技有限公司 | Crack identification method in complex environment |
| CN113012111A (en) * | 2021-02-25 | 2021-06-22 | 北京麦飞科技有限公司 | Rice lodging degree grading monitoring method based on edge calculation |
Non-Patent Citations (3)
| Title |
|---|
| 范留明 等: "基于数码摄影技术的岩体裂隙测量方法初探", 《岩石力学与工程学报》, vol. 24, no. 5, 31 March 2005 (2005-03-31), pages 792 - 797 * |
| 陆玲 等: "《图像目标分割方法》", 30 November 2016, 哈尔滨工程大学出版社, pages: 58 - 63 * |
| 韩冰 等: "《数字音视频处理》", 31 October 2018, 西安电子科技大学出版社, pages: 167 - 168 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114419088A (en) * | 2022-01-24 | 2022-04-29 | 重庆中环建设有限公司 | Tunnel surrounding rock information extraction method based on digital image |
| CN115147615A (en) * | 2022-07-01 | 2022-10-04 | 河海大学 | Rock image classification method and device based on metric learning network |
| CN115217084A (en) * | 2022-07-22 | 2022-10-21 | 中国华能集团清洁能源技术研究院有限公司 | Reservoir area expansive soil surface crack rate detection method and system |
| CN115217084B (en) * | 2022-07-22 | 2023-07-28 | 中国华能集团清洁能源技术研究院有限公司 | Method and system for detecting surface fracture rate of expansive soil in reservoir area |
| CN115546113A (en) * | 2022-09-15 | 2022-12-30 | 山东大学 | Method and system for predicting parameters of tunnel face crack image and front three-dimensional structure |
| CN115861848A (en) * | 2023-03-01 | 2023-03-28 | 成都理工大学 | Method and device for processing rock mass image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113743227A (en) | Rock mass fracture intelligent identification method and system based on proximity photography | |
| CN106909886B (en) | A kind of high-precision method for traffic sign detection and system based on deep learning | |
| CN109255317B (en) | Aerial image difference detection method based on double networks | |
| CN110889324A (en) | Thermal infrared image target identification method based on YOLO V3 terminal-oriented guidance | |
| CN111985274B (en) | Remote sensing image segmentation method based on convolutional neural network | |
| CN114549956B (en) | Deep learning-assisted inclined model building outer elevation target recognition method | |
| CN115205692B (en) | Typical feature intelligent identification and extraction method based on generation of countermeasure network | |
| CN114004977A (en) | Aerial photography data target positioning method and system based on deep learning | |
| CN115497015A (en) | River floating pollutant identification method based on convolutional neural network | |
| CN116385958A (en) | Edge intelligent detection method for power grid inspection and monitoring | |
| CN111831010A (en) | Unmanned aerial vehicle obstacle avoidance flight method based on digital space slice | |
| CN111027436A (en) | Northeast black fungus disease and pest image recognition system based on deep learning | |
| CN114445467A (en) | Specific target identification and tracking system of quad-rotor unmanned aerial vehicle based on vision | |
| CN114037895A (en) | Unmanned aerial vehicle pole tower inspection image identification method | |
| CN113781431A (en) | Green vision rate calculation method based on urban point cloud data | |
| CN113610009A (en) | Flood disaster unmanned aerial vehicle image information extraction system | |
| CN113033386A (en) | High-resolution remote sensing image-based transmission line channel hidden danger identification method and system | |
| CN116664545A (en) | Offshore benthos quantitative detection method and system based on deep learning | |
| CN116543322A (en) | Intelligent property routing inspection method based on community potential safety hazards | |
| CN114659499B (en) | Smart city 3D map model photography establishment method based on unmanned aerial vehicle technology | |
| CN115661673A (en) | Image target detection method based on YOLOv4 and attention mechanism | |
| CN114694048A (en) | Sparse shrub species identification method and system based on unmanned aerial vehicle remote sensing technology | |
| CN114463628A (en) | Deep learning remote sensing image ship target identification method based on threshold value constraint | |
| Lari et al. | Automatic extraction of building features from high resolution satellite images using artificial neural networks | |
| Skinner et al. | Detection and segmentation of underwater archaeological sites surveyed with stereo-vision platforms |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |