CN116109899A - Ancient architecture repairing method, system, computer equipment and storage medium - Google Patents
Ancient architecture repairing method, system, computer equipment and storage medium Download PDFInfo
- Publication number
- CN116109899A CN116109899A CN202211605232.0A CN202211605232A CN116109899A CN 116109899 A CN116109899 A CN 116109899A CN 202211605232 A CN202211605232 A CN 202211605232A CN 116109899 A CN116109899 A CN 116109899A
- Authority
- CN
- China
- Prior art keywords
- target
- masonry
- building
- detail
- repair
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 68
- 238000013461 design Methods 0.000 claims abstract description 131
- 230000008439 repair process Effects 0.000 claims abstract description 127
- 238000004422 calculation algorithm Methods 0.000 claims description 68
- 238000010586 diagram Methods 0.000 claims description 55
- 238000013527 convolutional neural network Methods 0.000 claims description 53
- 238000012545 processing Methods 0.000 claims description 48
- 239000002023 wood Substances 0.000 claims description 44
- 238000013528 artificial neural network Methods 0.000 claims description 28
- 230000011218 segmentation Effects 0.000 claims description 26
- 230000000694 effects Effects 0.000 claims description 23
- 230000007246 mechanism Effects 0.000 claims description 16
- 238000004590 computer program Methods 0.000 claims description 13
- 238000010276 construction Methods 0.000 claims description 13
- 238000012549 training Methods 0.000 claims description 13
- 230000008569 process Effects 0.000 claims description 11
- 238000001514 detection method Methods 0.000 claims description 7
- 230000007547 defect Effects 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 3
- 238000011176 pooling Methods 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 7
- 238000003062 neural network model Methods 0.000 abstract description 5
- 238000011835 investigation Methods 0.000 description 8
- 238000011160 research Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 238000004088 simulation Methods 0.000 description 4
- 238000013473 artificial intelligence Methods 0.000 description 3
- 239000000203 mixture Substances 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 238000012827 research and development Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/04—Inference or reasoning models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
- G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Working Measures On Existing Buildindgs (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method and a system for repairing ancient architecture, which are based on the identification of surface damage of various neural network models. The gradually fine damage identification technology effectively solves the problems of multiple types of damage of the ancient building and high identification difficulty, provides a brand new solution for the damage identification of the ancient building, greatly reduces the manual participation, and simultaneously ensures the scientific and reasonable repair design scheme.
Description
Technical Field
The present application relates to the field of ancient architecture repair technology, and in particular, to a method, a system, a computer device, and a storage medium for repairing an ancient architecture.
Background
The great number and the importance of the ancient architecture in the key cultural insurance units in the whole country make the repair and digital protection work of the ancient architecture more and more important and urgent.
At present, the ancient architecture is repaired with the following problems:
(1) The ancient architecture surface damage cannot completely record damage information due to negligence, omission, experience deficiency and other reasons in the manual investigation process, and meanwhile, the repair design scheme is not matched with a real damage processing method due to insufficient service capacity of a designer and other reasons, so that damage cannot be repaired pertinently and effectively.
(2) The ancient architecture damage investigation adopts manual measurement, investigation and field data arrangement, and the operation mode is time-consuming, large in workload and low in investigation result precision.
(3) At present, ancient architecture damage repair design adopts designers to manually draw drawings according to industry standard requirements and combine design experience, then manually calculate engineering quantity, and the quality of design results depends on the professional level and personal experience of the designers, so that the quality is difficult to guarantee.
Disclosure of Invention
Based on the above, the method, the system, the computer equipment and the storage medium for repairing the ancient architecture are provided for solving the problems of complex identification of the damaged type of the existing ancient architecture, high difficulty of manual identification and strong specialization of repairing design scheme.
In a first aspect, a method of repairing an ancient building, the method comprising:
obtaining a target orthographic image of a target ancient building;
processing the target orthographic image of the target ancient architecture based on a dense connection convolutional neural network DenseNet combined with multi-scale characteristics and an attention mechanism algorithm, and realizing the structural segmentation of the detail of the ancient architecture to obtain a structural diagram of the detail of the target ancient architecture; wherein the detail structure diagram of the target ancient building comprises a masonry detail structure and a wood detail structure;
automatically detecting and cutting the masonry structure detail structure by adopting a FasterR-CNN algorithm to obtain a masonry structure basic structure unit;
automatically detecting and dividing the surface damage of the wood structure detail structure and the masonry basic structure unit by using a mask FasterR-CNN algorithm to obtain different types of damage parameters of the surfaces of the wood structure detail structure and the masonry basic structure unit;
Processing different types of damage parameters on the surfaces of the wood structure detail structure and masonry structure basic structure unit based on BP neural network and expert system algorithm to generate damage processing design parameters of the target building and repair strategy parameters of the target building;
modeling the damage processing design parameters of the target building based on a PKPM3D graphic platform, performing virtual roaming, and outputting a three-dimensional effect data graph of the target ancient building repair design;
based on the output module and combining the repair strategy parameters, outputting a repair design engineering drawing and a repair design engineering quantity table of the target ancient building;
and generating a target ancient building repair design result diagram by combining the three-dimensional effect data diagram of the target ancient building repair design, the repair design engineering diagram of the target ancient building and the repair design engineering quantity table.
In the above aspect, optionally, the acquiring the target orthographic image of the target ancient architecture includes acquiring the target orthographic image of the target ancient architecture through unmanned aerial vehicle oblique photogrammetry or close-up photogrammetry.
In the above scheme, further optionally, the dense connection convolutional neural network DenseNet-based multi-scale feature and attention mechanism algorithm is used for processing the target orthographic image of the target ancient architecture, so as to realize the structural segmentation of the detail of the ancient architecture, and obtain the detail structural diagram of the target ancient architecture, specifically: the target orthophoto map is subjected to multi-scale segmentation processing to obtain three-size target image data; inputting the segmented target image data into a densely connected convolutional neural network to respectively obtain first target feature images extracted by the neural network under different sizes;
Fusing the first target feature images together, and simultaneously sending the first target feature images into an attention model to obtain weight distribution of each pixel point of the target feature images;
and carrying out dot multiplication on the weight distribution of each pixel point and the fused second target feature map, and obtaining a detail structure map of the target ancient building through Soft-Max.
In the above scheme, further optionally, the method comprises automatically detecting and cutting the masonry structure detail structure by using a FasterR-CNN algorithm to obtain a masonry structure basic structure unit, specifically: inputting the original image of the masonry structure detail structure in the training set into a CNN network for network training to obtain a characteristic diagram of the original image of the masonry structure detail structure;
inputting a feature map of the masonry structure detail construction original image into an RPN network, generating a region of interest (ROI), and carrying out pooling treatment on the ROI; and (3) classifying and carrying out boundary regression on the masonry structure detail structure through a series of full-connection layers to obtain the masonry structure basic construction unit.
In the above solution, further optionally, the automatically detecting and dividing the surface damage of the wood structure detail structure and the masonry basic structure unit by using a maskpaster-CNN algorithm to obtain different types of damage parameters of the surface of the wood structure detail structure and the masonry basic structure unit, specifically: inputting the wood structure detail structure and masonry basic structure unit images in the training set into a pre-trained neural network to obtain a corresponding FeatureMap; setting a preset number of ROIs for each point in the Feature Map to obtain a plurality of candidate ROIs, sending the candidate ROIs into an RPN network to perform binary classification and BB regression, and filtering out part of candidate ROIs to obtain residual ROIs; and performing ROIALign operation on the residual ROI, and performing classification, BB regression and MASK generation on the residual ROI to obtain different types of damage parameters of the surface of the wood structure detail structure and the masonry basic structure unit.
In the above scheme, further optionally, the expert system stores the target historic building damage parameters, the intermediate running state and the reasoning process record and result information by creating a global database in advance;
storing experience knowledge based on industry design specification knowledge and experience sample learning based on BP neural network by pre-creating a knowledge base;
through pre-creating a neural network inference engine, each design element is inferred in a forward direction inference mode;
the system outputs the complete reasoning flow after completing the reasoning process by creating an interpreter in advance to realize the output of the system interpretation text for inquiring a certain problem by the user.
In the above solution, further optionally, the masonry basic building unit comprises: roof tiles, wall tiles and column base units.
In a second aspect, an ancient building repair system, the system comprising:
the acquisition module is used for: the method comprises the steps of obtaining a target orthographic image of a target ancient building;
a first processing module: the method is used for processing the target orthographic image of the target ancient building based on the dense connection convolutional neural network DenseNet combined with the multi-scale characteristics and the attention mechanism algorithm, so as to realize the structural segmentation of the detail of the ancient building and obtain a detail structural diagram of the target ancient building; wherein the detail structure diagram of the target ancient building comprises a masonry detail structure and a wood detail structure;
And a second processing module: the method is used for automatically detecting and cutting the masonry structure detail structure by adopting a FasterR-CNN algorithm to obtain a masonry structure basic structure unit;
and a third processing module: the method is used for automatically detecting and dividing the surface damage of the wood structure detail structure and the masonry basic structure unit through a mask FasterR-CNN algorithm to obtain different types of damage parameters of the surfaces of the wood structure detail structure and the masonry basic structure unit;
and a parameter generation module: the method is used for processing different types of damage parameters on the surfaces of the wood structure detail structure and masonry structure basic structure units based on a BP neural network and an expert system algorithm to generate damage processing design parameters of the target building and repair strategy parameters of the target building;
a first output module: the system is used for modeling the damage processing design parameters of the target building based on the PKPM3D graphic platform, performing virtual roaming, and outputting a three-dimensional effect data graph of the target ancient building repair design;
and a second output module: the repair strategy parameters are combined based on the output module, and a repair design engineering drawing and a repair design engineering quantity table of the target ancient building are output;
And a third output module: and the three-dimensional effect data graph is used for combining the target ancient architecture repair design, the target ancient architecture repair design engineering graph and the repair design engineering quantity table to generate a target ancient architecture repair design result graph.
In a third aspect, a computer device comprises a memory storing a computer program and a processor implementing the following steps when executing the computer program:
obtaining a target orthographic image of a target ancient building;
processing the target orthographic image of the target ancient architecture based on a dense connection convolutional neural network DenseNet combined with multi-scale characteristics and an attention mechanism algorithm, and realizing the structural segmentation of the detail of the ancient architecture to obtain a structural diagram of the detail of the target ancient architecture; wherein the detail structure diagram of the target ancient building comprises a masonry detail structure and a wood detail structure;
automatically detecting and cutting the masonry structure detail structure by adopting a FasterR-CNN algorithm to obtain a masonry structure basic structure unit;
automatically detecting and dividing the surface damage of the wood structure detail structure and the masonry basic structure unit by using a mask FasterR-CNN algorithm to obtain different types of damage parameters of the surfaces of the wood structure detail structure and the masonry basic structure unit;
Processing different types of damage parameters on the surfaces of the wood structure detail structure and masonry structure basic structure unit based on BP neural network and expert system algorithm to generate damage processing design parameters of the target building and repair strategy parameters of the target building;
modeling the damage processing design parameters of the target building based on a PKPM3D graphic platform, performing virtual roaming, and outputting a three-dimensional effect data graph of the target ancient building repair design;
based on the output module and combining the repair strategy parameters, outputting a repair design engineering drawing and a repair design engineering quantity table of the target ancient building;
and generating a target ancient building repair design result diagram by combining the three-dimensional effect data diagram of the target ancient building repair design, the repair design engineering diagram of the target ancient building and the repair design engineering quantity table.
In a fourth aspect, a computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:
obtaining a target orthographic image of a target ancient building;
processing the target orthographic image of the target ancient architecture based on a dense connection convolutional neural network DenseNet combined with multi-scale characteristics and an attention mechanism algorithm, and realizing the structural segmentation of the detail of the ancient architecture to obtain a structural diagram of the detail of the target ancient architecture; wherein the detail structure diagram of the target ancient building comprises a masonry detail structure and a wood detail structure;
Automatically detecting and cutting the masonry structure detail structure by adopting a FasterR-CNN algorithm to obtain a masonry structure basic structure unit;
automatically detecting and dividing the surface damage of the wood structure detail structure and the masonry basic structure unit by using a mask FasterR-CNN algorithm to obtain different types of damage parameters of the surfaces of the wood structure detail structure and the masonry basic structure unit;
processing different types of damage parameters on the surfaces of the wood structure detail structure and masonry structure basic structure unit based on BP neural network and expert system algorithm to generate damage processing design parameters of the target building and repair strategy parameters of the target building;
modeling the damage processing design parameters of the target building based on a PKPM3D graphic platform, performing virtual roaming, and outputting a three-dimensional effect data graph of the target ancient building repair design;
based on the output module and combining the repair strategy parameters, outputting a repair design engineering drawing and a repair design engineering quantity table of the target ancient building;
and generating a target ancient building repair design result diagram by combining the three-dimensional effect data diagram of the target ancient building repair design, the repair design engineering diagram of the target ancient building and the repair design engineering quantity table.
The invention has at least the following beneficial effects:
based on further analysis and research on the problems in the prior art, the invention realizes that the existing ancient architecture repair is mainly performed by manually measuring, investigating and arranging the ancient architecture damage data, and the repair scheme is manually provided, so that the problems of complex identification of the existing ancient architecture damage type, large manual identification difficulty and strong specialization of the repair design scheme are solved. The invention is based on the surface damage identification of various neural network models, the whole ancient architecture is firstly divided into detail structures, the detail structures are further divided into basic construction units according to the detail structure types, the damage in the basic construction units is extracted, further, the intelligent reasoning of scheme design is realized by adopting an algorithm based on the mixture of a neural network and an expert system, the reasoning parameters are modeled through a PKPM3D platform and virtual roaming is carried out, the visual three-dimensional effect display of the ancient architecture repair design is realized, the corresponding standard diagram and engineering quantity calculation table are output, and finally the target ancient architecture repair design result diagram is generated. The gradually fine damage identification technology effectively solves the problems of multiple types of damage of the ancient building and high identification difficulty, provides a brand new solution for the damage identification of the ancient building, greatly reduces the manual participation, and simultaneously ensures the scientific and reasonable repair design scheme.
Drawings
FIG. 1 is a schematic flow chart of a method for repairing an ancient architecture according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a first stage algorithm of an ancient architecture restoration method according to an embodiment of the present invention;
FIG. 3 is a schematic diagram illustrating a detailed structure division experiment of an ancient architecture repairing method according to an embodiment of the present invention;
FIG. 4 is a schematic flow chart of a second stage algorithm of the ancient architecture restoration method according to an embodiment of the present invention;
fig. 5 is a schematic diagram of a first basic structure division experiment result of an ancient architecture repairing method according to an embodiment of the present invention;
fig. 6 is a schematic diagram of a second basic structure division experiment result of the ancient architecture restoration method according to an embodiment of the present invention;
fig. 7 is a schematic diagram of a repair strategy experimental result of a repair method for an ancient architecture according to an embodiment of the present invention;
FIG. 8 is a schematic diagram of a parameterized modeling bottom layer effort of a historic building repair method according to an embodiment of the present invention;
FIG. 9 is a schematic diagram showing simulation of the repair design output results of the ancient architecture repair method according to one embodiment of the present invention;
FIG. 10 is a schematic flow diagram of an overall technical circuit diagram of an ancient architecture restoration method according to an embodiment of the present invention;
FIG. 11 is an internal block diagram of a computer device in one embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.
The great number and the importance of the ancient architecture in the key cultural insurance units in the whole country make the repair and digital protection work of the ancient architecture more and more important and urgent. The embodiment provides a whole set of algorithm based on artificial intelligence deep learning, and combines a big data expert system to enable a computer to accurately identify the damage of the ancient architecture and autonomously complete a whole set of ancient architecture repair design scheme comprising repair design drawings and engineering quantities. The project research and development content covers the ancient architecture surface damage identification and detection system, the ancient architecture repair design expert system and the parametric modeling and result output system, and an artificial intelligence system for the ancient architecture repair design is formed. The problems of complex ancient architecture damage type identification and great manual identification difficulty are effectively solved, and the problems of strong specialization and complex content of the repairing design scheme are solved. The quality and efficiency of repairing design are greatly improved, the input cost is greatly reduced, and a brand new artificial intelligence solution is provided for the repairing design of ancient buildings.
In one embodiment, as shown in fig. 1 and 10, there is provided a method for repairing an ancient architecture, comprising the steps of:
and obtaining a target orthographic image of the target ancient building.
Processing the target orthographic image of the target ancient architecture based on a dense connection convolutional neural network DenseNet combined with multi-scale characteristics and an attention mechanism algorithm, and realizing the structural segmentation of the detail of the ancient architecture to obtain a structural diagram of the detail of the target ancient architecture; wherein the detail structure diagram of the target ancient building comprises a masonry detail structure and a wood detail structure;
automatically detecting and cutting the masonry structure detail structure by adopting a FasterR-CNN algorithm to obtain a masonry structure basic structure unit;
and automatically detecting and dividing the surface damage of the wood structure detail structure and the masonry basic structure unit by a mask FasterR-CNN algorithm to obtain different types of damage parameters of the surface of the wood structure detail structure and the masonry basic structure unit.
In one embodiment, the process of identifying and detecting the surface damage using the neural network models includes the following three steps:
the first stage: because the sizes and distribution positions of various wood structures and masonry structures to be extracted in the historic building orthophoto data are greatly different, it is difficult to obtain object features of targets with different sizes in the same image at the same time; by utilizing the advantage that the multi-scale characteristics can improve the semantic segmentation capability of the model and combining the characteristic that the attention model allows the model to focus on the most relevant characteristics according to the needs, the algorithm that the densely connected convolutional neural network is combined with the multi-scale characteristics and the attention mechanism is provided, and the research shows that the method can effectively improve the segmentation capability and the segmentation effect of the model.
As shown in fig. 2, the timber structure details in the archetype orthophoto map are achieved by using a densely connected convolutional neural network DenseNet in combination with multi-scale features and attention mechanisms, such as: the method comprises the steps of dividing and extracting doors and windows, column shafts, purlins, roof boards, wall boards, ground boards, bench bases, steps, column bases and the like, and building blocks of masonry structures, wherein the surface damage detection is carried out on the masonry structures through a third-stage algorithm, basic building blocks are divided through a second-stage algorithm, and then the surface damage detection is carried out on the masonry structures through the third-stage algorithm. Fig. 3 is a detailed structure division experimental simulation diagram of the target building obtained as a result of the first stage experiment.
And a second stage: the existing method for identifying and extracting the damaged basic building units (such as roof tiles, wall tiles, floor tiles, column foundations and other basic units) of the masonry structure in the ancient architecture depends on manual identification and judgment one by one, and the method is time-consuming, labor-consuming and labor-consuming; through researches, the FasterR-CNN algorithm can realize the accurate segmentation of a single small target in a large scene, so that the FasterR-CNN algorithm is adopted to automatically detect and cut the detailed structure of the masonry structure to obtain a basic structural unit image, and the algorithm flow is shown in figure 4.
The first basic structure segmentation experimental result is shown in fig. 5, and the masonry structure detail structure extracted in the first stage is automatically detected and cut through a FasterR-CNN algorithm to obtain a basic structure unit which is used as a data set for detecting and segmenting the damage in the third stage.
And a third stage: through researches, most of target detection algorithms inherit and develop an R-CNN idea, and the algorithms can be effectively applied to the recognition and extraction of the ancient architecture surface damage; because the R-CNN algorithm has the defects of low model training efficiency, large calculation space occupied by training data and the like, the method of adding the feature extraction of the candidate region into the convolutional neural network training to realize direct end-to-end training and testing and estimating the segmentation Mask of each region of interest by adding a Mask branch is adopted to solve the defects, so that the Mask-CNN algorithm is provided for automatically detecting and segmenting the damage.
The second basic structure segmentation experimental result is shown in fig. 6, and the mask FasterR-CNN algorithm is used for automatically detecting and segmenting the surface damage of the wood structure detail structure and masonry structure basic structure unit, so that the accurate detection and measurement of different types of damage are realized, and the damage type information and the state parameters are used as a data set of the repair design expert system.
In the research and development process, the damage on the surface of the ancient building cannot be completely recorded due to negligence, omission, insufficient experience and other reasons in the manual investigation process, and meanwhile, the repair design scheme is not matched with the real damage processing method and cannot be pointedly and effectively repaired due to the reasons of insufficient business capability of a designer and the like. The BP neural network is an algorithm developed on the basis of simulating human brain neural tissue, has self-learning and self-adapting capabilities and the capability of applying learning results to new knowledge, can realize learning of design experience and be applied to subsequent design schemes by combining an expert system algorithm, learns experience parameters and updates a knowledge base, and solves the problem of difficult acquisition and update of the knowledge of the traditional expert system, so that the BP neural network combined expert system repairing design algorithm is provided.
The repair strategy is shown in fig. 7, and based on BP neural network and expert system algorithm, the automatic reasoning repair design scheme for different damage types is realized, and the repair scheme and damage processing design parameters are provided as the basic data of the parameterized modeling system.
The parametric modeling bottom layer result is shown in fig. 8, and the PKPM3D graphic platform has a live-action roaming function, so that a design scene can be truly displayed in any direction, a designer can feel the design scheme on the spot, and the parametric modeling bottom layer result has an important effect on verifying that the design scheme is scientific and reasonable.
The parameterized modeling system is based on a PKPM3D graphic platform, performs secondary development, models a repair scheme and damage processing design parameters given by a repair design expert system, and performs virtual roaming, so that three-dimensional effect display of the repair design of the ancient architecture is realized.
The simulation display of the repair design output result is shown in fig. 9, and the Python language has the characteristics of simplicity, clarity, easiness in mastering, rich standard library, strong compatibility and the like, and is widely applied to the field of automation office; the repairing design results are divided into scheme design drawings, damage treatment measure engineering quantity tables and the like, and have complex content and strong specialization; the repair design result chart output module based on the Python development platform can realize repair strategy parameters given by an expert system, call a database, output corresponding chart results such as a repair design engineering drawing, a repair design engineering quantity table and the like, and finally output a simulation display diagram of the repair design output result.
Processing different types of damage parameters on the surfaces of the wood structure detail structure and masonry structure basic structure unit based on BP neural network and expert system algorithm to generate damage processing design parameters of the target building and repair strategy parameters of the target building;
modeling the damage processing design parameters of the target building based on a PKPM3D graphic platform, performing virtual roaming, and outputting a three-dimensional effect data graph of the target ancient building repair design;
based on the output module and combining the repair strategy parameters, outputting a repair design engineering drawing and a repair design engineering quantity table of the target ancient building;
and generating a target ancient building repair design result diagram by combining the three-dimensional effect data diagram of the target ancient building repair design, the repair design engineering diagram of the target ancient building and the repair design engineering quantity table.
The surface damage recognition technology based on various neural network models is characterized in that the whole ancient architecture is divided into the detail structures, the detail structures are divided into the basic construction units according to the detail structure types, the damage in the basic construction units is extracted, and the gradually and fine damage recognition technology effectively solves the problems of multiple types of damage of the ancient architecture and high recognition difficulty and provides a brand-new solution for the damage recognition of the ancient architecture.
In one embodiment, the acquiring the target orthophoto map of the target monument comprises acquiring the target orthophoto map of the target monument by unmanned aerial vehicle oblique photogrammetry or proximity photogrammetry. In the prior art, the surface damage information of the ancient architecture is obtained through photographing and manual measurement, damage marking is carried out in a manual drawing mode, and then a drawing is drawn according to a repair principle and a repair design scheme is manufactured.
In one embodiment, the dense connection convolutional neural network DenseNet-based multi-scale feature and attention mechanism algorithm is used for processing the target orthographic image of the target ancient architecture to realize the structural segmentation of the detail of the ancient architecture, so as to obtain a structural diagram of the detail of the target ancient architecture, which is specifically as follows: the target orthophoto map is subjected to multi-scale segmentation processing to obtain three-size target image data; inputting the segmented target image data into a densely connected convolutional neural network to respectively obtain first target feature images extracted by the neural network under different sizes;
fusing the first target feature images together, and simultaneously sending the first target feature images into an attention model to obtain weight distribution of each pixel point of the target feature images;
And carrying out dot multiplication on the weight distribution of each pixel point and the fused second target feature map, and obtaining a detail structure map of the target ancient building through Soft-Max.
In one embodiment, the automatic detection and cutting of the masonry structure detail structure by using a FasterR-CNN algorithm is performed to obtain a masonry structure basic structure unit, which specifically comprises the following steps: inputting the original image of the masonry structure detail structure in the training set into a CNN network for network training to obtain a characteristic diagram of the original image of the masonry structure detail structure;
inputting a feature map of the masonry structure detail construction original image into an RPN network, generating a region of interest (ROI), and carrying out pooling treatment on the ROI; and (3) classifying and carrying out boundary regression on the masonry structure detail structure through a series of full-connection layers to obtain the masonry structure basic construction unit.
The dense connection convolutional neural network DenseNet is combined with the multi-scale characteristics and the attention mechanism to divide the model to be applied to the extraction of detail structures in the orthographic images of the ancient architecture, the novel efficient network model of the dense connection convolutional neural network DenseNet is combined with the multi-scale characteristics to obviously improve the semantic division capacity of the model, and the model can focus the most relevant characteristics according to the needs by combining with the attention mechanism so as to improve the output quality of the model; the segmentation effect and the segmentation capability of the DenseNet model combined with the multi-scale characteristics and the attention mechanism are greatly improved, the segmentation boundary is clearer, and a brand new algorithm is provided for accurately segmenting the detail structure.
In one embodiment, the method automatically detects and segments the surface damage of the wooden structure detail structure and the basic masonry unit by using a maskpister-CNN algorithm to obtain different types of damage parameters of the wooden structure detail structure and the basic masonry unit surface, specifically: inputting the wood structure detail structure and masonry basic structure unit images in the training set into a pre-trained neural network to obtain a corresponding FeatureMap; setting a preset number of ROIs for each point in the FeatureMap to obtain a plurality of candidate ROIs, sending the candidate ROIs into an RPN network to perform binary classification and BB regression, and filtering out part of candidate ROIs to obtain residual ROIs; and performing ROIALign operation on the residual ROI, and performing classification, BB regression and MASK generation on the residual ROI to obtain different types of damage parameters of the surface of the wood structure detail structure and the masonry basic structure unit.
The FasterR-CNN algorithm and the mask-CNN algorithm realize automatic segmentation and extraction of pixel-level damage, and provide a brand new solution for accurately identifying and measuring damage parameters of the surface of the basic construction unit.
In one embodiment, the expert system stores the target historic building damage parameters, intermediate operating states and reasoning process records and result information by creating a global database in advance;
storing experience knowledge based on industry design specification knowledge and experience sample learning based on BP neural network by pre-creating a knowledge base;
through pre-creating a neural network inference engine, each design element is inferred in a forward direction inference mode;
the system outputs the complete reasoning flow after completing the reasoning process by creating an interpreter in advance to realize the output of the system interpretation text for inquiring a certain problem by the user.
The ancient architecture repair design scheme of the embodiment has strong specialization and complex content, and the neural network is adopted to combine with the intelligent reasoning of the repair design scheme of the expert system, so that the establishment of the repair scheme is efficient and scientific; the method comprises the steps of learning experience parameters of designers through a neural network algorithm and updating a knowledge base, so that the problem that expert experience parameters are difficult to acquire and update is solved.
In one embodiment, the masonry basic building block comprises: roof tiles, wall tiles and column base units.
The ancient architecture damage investigation currently adopts manual measurement, investigation and arrangement of field data, and the operation mode is time-consuming, large in workload and low in investigation result precision, and the ancient architecture surface damage identification model of the embodiment can realize rapid and accurate acquisition of damage information, and greatly reduces manual participation.
At present, the ancient architecture damage repair design adopts the steps that a designer combines design experience according to industry standard requirements, manually draws a picture and then manually calculates engineering quantity, and the quality of a design result depends on the professional level and personal experience of the designer and is difficult to ensure; according to the embodiment, an algorithm based on the mixture of the neural network and the expert system is adopted, intelligent reasoning of scheme design is achieved, the inferred parameters are modeled through a PKPM3D platform and virtual roaming is carried out, visual three-dimensional effect display of ancient building repair design is achieved, corresponding standard diagrams and engineering quantity calculation tables are output, and the scientific and reasonable repair design scheme is guaranteed.
Based on further analysis and research on the problems in the prior art, the embodiment recognizes that the current ancient architecture repair is mainly performed by manually measuring, investigating and arranging the ancient architecture damage data, and the repair scheme is manually provided, so that the problems of complex identification of the existing ancient architecture damage type, high manual identification difficulty and strong specialization of the repair design scheme are solved. The invention is based on the surface damage identification of various neural network models, the whole ancient architecture is firstly divided into detail structures, the detail structures are further divided into basic construction units according to the detail structure types, the damage in the basic construction units is extracted, further, the intelligent reasoning of scheme design is realized by adopting an algorithm based on the mixture of a neural network and an expert system, the reasoning parameters are modeled through a PKPM3D platform and virtual roaming is carried out, the visual three-dimensional effect display of the ancient architecture repair design is realized, the corresponding standard diagram and engineering quantity calculation table are output, and finally the target ancient architecture repair design result diagram is generated. The gradually fine damage identification technology effectively solves the problems of multiple types of damage of the ancient building and high identification difficulty, provides a brand new solution for the damage identification of the ancient building, greatly reduces the manual participation, and simultaneously ensures the scientific and reasonable repair design scheme.
It should be understood that, although the steps in the flowchart of fig. 1 are shown in sequence as indicated by the arrows, the steps are not necessarily performed in sequence as indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least a portion of the steps in fig. 1 may include a plurality of steps or stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily sequential, but may be performed in rotation or alternatively with at least a portion of the steps or stages in other steps or other steps.
In one embodiment, an ancient building repair system is provided, comprising the following program modules:
the acquisition module is used for: the method comprises the steps of obtaining a target orthographic image of a target ancient building;
a first processing module: the method is used for processing the target orthographic image of the target ancient building based on the dense connection convolutional neural network DenseNet combined with the multi-scale characteristics and the attention mechanism algorithm, so as to realize the structural segmentation of the detail of the ancient building and obtain a detail structural diagram of the target ancient building; wherein the detail structure diagram of the target ancient building comprises a masonry detail structure and a wood detail structure;
And a second processing module: the method is used for automatically detecting and cutting the masonry structure detail structure by adopting a FasterR-CNN algorithm to obtain a masonry structure basic structure unit;
and a third processing module: the method is used for automatically detecting and dividing the surface damage of the wood structure detail structure and the masonry basic structure unit through a mask FasterR-CNN algorithm to obtain different types of damage parameters of the surfaces of the wood structure detail structure and the masonry basic structure unit;
and a parameter generation module: the method is used for processing different types of damage parameters on the surfaces of the wood structure detail structure and masonry structure basic structure units based on a BP neural network and an expert system algorithm to generate damage processing design parameters of the target building and repair strategy parameters of the target building;
a first output module: the system is used for modeling the damage processing design parameters of the target building based on the PKPM3D graphic platform, performing virtual roaming, and outputting a three-dimensional effect data graph of the target ancient building repair design;
and a second output module: the repair strategy parameters are combined based on the output module, and a repair design engineering drawing and a repair design engineering quantity table of the target ancient building are output;
And a third output module: and the three-dimensional effect data graph is used for combining the target ancient architecture repair design, the target ancient architecture repair design engineering graph and the repair design engineering quantity table to generate a target ancient architecture repair design result graph.
For specific limitations on the ancient building repair system, reference may be made to the above limitations on the ancient building repair method, and no further description is given here. The modules in the ancient building repair system can be realized in whole or in part by software, hardware and a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a terminal, and the internal structure thereof may be as shown in fig. 11. The computer device includes a processor, a memory, a communication interface, a display screen, and an input system connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless mode can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program when executed by a processor implements a method for repairing an ancient architecture. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input system of the computer equipment can be a touch layer covered on the display screen, can also be keys, a track ball or a touch pad arranged on the shell of the computer equipment, and can also be an external keyboard, a touch pad or a mouse and the like.
It will be appreciated by those skilled in the art that the structure shown in fig. 11 is merely a block diagram of a portion of the structure associated with the present application and is not limiting of the computer device to which the present application applies, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.
In one embodiment, a computer device is provided, including a memory and a processor, the memory having stored therein a computer program, involving all or part of the flow of the methods of the embodiments described above.
In one embodiment, a computer readable storage medium having a computer program stored thereon is provided, involving all or part of the flow of the methods of the embodiments described above.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile memory may include Read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, or the like. Volatile memory can include Random access memory (Random AccessMemory, RAM) or external cache memory. By way of illustration, and not limitation, RAM can take many forms, such as static random access memory (StaticRandomAccessMemory, SRAM) or dynamic random access memory (DynamicRandomAccessMemory, DRAM), among others.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above examples merely represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the invention. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application is to be determined by the claims appended hereto.
Claims (10)
1. A method of repairing an ancient building, the method comprising:
obtaining a target orthographic image of a target ancient building;
processing the target orthographic image of the target ancient architecture based on a dense connection convolutional neural network DenseNet combined with multi-scale characteristics and an attention mechanism algorithm, and realizing the structural segmentation of the detail of the ancient architecture to obtain a structural diagram of the detail of the target ancient architecture; wherein the detail structure diagram of the target ancient building comprises a masonry detail structure and a wood detail structure;
Automatically detecting and cutting the masonry structure detail structure by adopting a FasterR-CNN algorithm to obtain a masonry structure basic structure unit;
automatically detecting and dividing the surface damage of the wood structure detail structure and the masonry basic structure unit by using a mask FasterR-CNN algorithm to obtain different types of damage parameters of the surfaces of the wood structure detail structure and the masonry basic structure unit;
processing different types of damage parameters on the surfaces of the wood structure detail structure and masonry structure basic structure unit based on BP neural network and expert system algorithm to generate damage processing design parameters of the target building and repair strategy parameters of the target building;
modeling the damage processing design parameters of the target building based on a PKPM3D graphic platform, performing virtual roaming, and outputting a three-dimensional effect data graph of the target ancient building repair design;
based on the output module and combining the repair strategy parameters, outputting a repair design engineering drawing and a repair design engineering quantity table of the target ancient building;
and generating a target ancient building repair design result diagram by combining the three-dimensional effect data diagram of the target ancient building repair design, the repair design engineering diagram of the target ancient building and the repair design engineering quantity table.
2. The method of claim 1, wherein the acquiring the target orthophoto map of the target monument comprises acquiring the target orthophoto map of the target monument by unmanned aerial vehicle oblique photogrammetry or proximity photogrammetry.
3. The method of claim 1, wherein the dense connection convolutional neural network-based DenseNet is combined with a multi-scale feature and attention mechanism algorithm to process a target orthographic image of the target ancient building, so as to realize the structural segmentation of the ancient building details, and obtain a structural diagram of the target ancient building details, specifically comprising: the target orthophoto map is subjected to multi-scale segmentation processing to obtain three-size target image data; inputting the segmented target image data into a densely connected convolutional neural network to respectively obtain first target feature images extracted by the neural network under different sizes;
fusing the first target feature images together, and simultaneously sending the first target feature images into an attention model to obtain weight distribution of each pixel point of the target feature images;
and carrying out dot multiplication on the weight distribution of each pixel point and the fused second target feature map, and obtaining a detail structure map of the target ancient building through Soft-Max.
4. The method according to claim 1, wherein said automatically detecting and cutting said masonry structure details by using a FasterR-CNN algorithm yields masonry structure basic building blocks, in particular: inputting the original image of the masonry structure detail structure in the training set into a CNN network for network training to obtain a characteristic diagram of the original image of the masonry structure detail structure;
inputting a feature map of the masonry structure detail construction original image into an RPN network, generating a region of interest (ROI), and carrying out pooling treatment on the ROI; and (3) classifying and carrying out boundary regression on the masonry structure detail structure through a series of full-connection layers to obtain the masonry structure basic construction unit.
5. The method according to claim 1, wherein the automatic detection and segmentation of the surface defects of the wood structure detail structure and the masonry basic structure unit by the maskpaster-CNN algorithm is performed to obtain different types of defect parameters of the surfaces of the wood structure detail structure and the masonry basic structure unit, specifically: inputting the wood structure detail structure and masonry basic structure unit images in the training set into a pre-trained neural network to obtain a corresponding FeatureMap; setting a preset number of ROIs for each point in the FeatureMap to obtain a plurality of candidate ROIs, sending the candidate ROIs into an RPN network to perform binary classification and BB regression, and filtering out part of candidate ROIs to obtain residual ROIs; and performing ROIALign operation on the residual ROI, and performing classification, BB regression and MASK generation on the residual ROI to obtain different types of damage parameters of the surface of the wood structure detail structure and the masonry basic structure unit.
6. The method of claim 1, wherein the expert system stores the target historic building damage parameters, intermediate operating states and reasoning process records and outcome information by creating a global database in advance;
storing experience knowledge based on industry design specification knowledge and experience sample learning based on BP neural network by pre-creating a knowledge base;
through pre-creating a neural network inference engine, each design element is inferred in a forward direction inference mode;
the system outputs the complete reasoning flow after completing the reasoning process by creating an interpreter in advance to realize the output of the system interpretation text for inquiring a certain problem by the user.
7. The method of claim 1, wherein the masonry basic building block comprises: roof tiles, wall tiles and column base units.
8. A historic building repair system, the system comprising:
the acquisition module is used for: the method comprises the steps of obtaining a target orthographic image of a target ancient building;
a first processing module: the method is used for processing the target orthographic image of the target ancient building based on the dense connection convolutional neural network DenseNet combined with the multi-scale characteristics and the attention mechanism algorithm, so as to realize the structural segmentation of the detail of the ancient building and obtain a detail structural diagram of the target ancient building; wherein the detail structure diagram of the target ancient building comprises a masonry detail structure and a wood detail structure;
And a second processing module: the method is used for automatically detecting and cutting the masonry structure detail structure by adopting a FasterR-CNN algorithm to obtain a masonry structure basic structure unit;
and a third processing module: the method is used for automatically detecting and dividing the surface damage of the wood structure detail structure and the masonry basic structure unit through a mask FasterR-CNN algorithm to obtain different types of damage parameters of the surfaces of the wood structure detail structure and the masonry basic structure unit;
and a parameter generation module: the method is used for processing different types of damage parameters on the surfaces of the wood structure detail structure and masonry structure basic structure units based on a BP neural network and an expert system algorithm to generate damage processing design parameters of the target building and repair strategy parameters of the target building;
a first output module: the system is used for modeling the damage processing design parameters of the target building based on the PKPM3D graphic platform, performing virtual roaming, and outputting a three-dimensional effect data graph of the target ancient building repair design;
and a second output module: the repair strategy parameters are combined based on the output module, and a repair design engineering drawing and a repair design engineering quantity table of the target ancient building are output;
And a third output module: and the three-dimensional effect data graph is used for combining the target ancient architecture repair design, the target ancient architecture repair design engineering graph and the repair design engineering quantity table to generate a target ancient architecture repair design result graph.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 7 when the computer program is executed.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211605232.0A CN116109899B (en) | 2022-12-14 | Ancient architecture repairing method, system, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211605232.0A CN116109899B (en) | 2022-12-14 | Ancient architecture repairing method, system, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116109899A true CN116109899A (en) | 2023-05-12 |
| CN116109899B CN116109899B (en) | 2024-07-09 |
Family
ID=
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101567030A (en) * | 2009-01-13 | 2009-10-28 | 西安交通大学 | Method of constructing virtual preservation and restoration system for wooden structure of ancient architecture |
| CN109002841A (en) * | 2018-06-27 | 2018-12-14 | 淮阴工学院 | A kind of building element extracting method based on Faster-RCNN model |
| CN111126183A (en) * | 2019-12-09 | 2020-05-08 | 武汉大学 | Method for detecting damage of building after earthquake based on near-ground image data |
| JP2020080095A (en) * | 2018-11-14 | 2020-05-28 | 株式会社竹中工務店 | Building information processing device and building information processing model learning device |
| CN111612907A (en) * | 2020-04-29 | 2020-09-01 | 湖南南派古建园林工程有限公司 | Multidirectional repairing system and method for damaged ancient building column |
| CN113326863A (en) * | 2021-03-24 | 2021-08-31 | 广州大学 | Building structure health condition detection method and system and repairing scheme determination method |
| CN115249320A (en) * | 2022-01-15 | 2022-10-28 | 山东建筑大学 | Damage detection method and system for wood structure ancient building and storable medium |
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101567030A (en) * | 2009-01-13 | 2009-10-28 | 西安交通大学 | Method of constructing virtual preservation and restoration system for wooden structure of ancient architecture |
| CN109002841A (en) * | 2018-06-27 | 2018-12-14 | 淮阴工学院 | A kind of building element extracting method based on Faster-RCNN model |
| JP2020080095A (en) * | 2018-11-14 | 2020-05-28 | 株式会社竹中工務店 | Building information processing device and building information processing model learning device |
| CN111126183A (en) * | 2019-12-09 | 2020-05-08 | 武汉大学 | Method for detecting damage of building after earthquake based on near-ground image data |
| CN111612907A (en) * | 2020-04-29 | 2020-09-01 | 湖南南派古建园林工程有限公司 | Multidirectional repairing system and method for damaged ancient building column |
| CN113326863A (en) * | 2021-03-24 | 2021-08-31 | 广州大学 | Building structure health condition detection method and system and repairing scheme determination method |
| CN115249320A (en) * | 2022-01-15 | 2022-10-28 | 山东建筑大学 | Damage detection method and system for wood structure ancient building and storable medium |
Non-Patent Citations (2)
| Title |
|---|
| YANG YU: ""A novel deep learning-based method for damage identification of smart building structures"", 《STRUCTURAL HEALTH MONITORING》, 31 October 2018 (2018-10-31), pages 1 - 21 * |
| 赵鹏;赵雪峰;赵庆安;李生元;方腾伟;赵鑫如;: "基于人工智能机器视觉技术的古建筑表层损伤检测", 物联网技术, no. 09, 20 September 2017 (2017-09-20), pages 20 - 24 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112883799B (en) | Method for accurately identifying ramp in CAD ramp large-scale sample graph | |
| CN115409069A (en) | Village and town building identification method, classification method, device, electronic equipment and medium | |
| CN111402414A (en) | Point cloud map construction method, device, equipment and storage medium | |
| JP7333029B2 (en) | Program, information processing method and information processing apparatus | |
| CN104077447A (en) | Urban three-dimensional space vector modeling method based on paper plane data | |
| CN112257665A (en) | Image content recognition method, image recognition model training method, and medium | |
| CN110866922A (en) | Image semantic segmentation model and modeling method based on reinforcement learning and transfer learning | |
| CN113674216A (en) | Subway tunnel disease detection method based on deep learning | |
| CN113569852A (en) | Training method and device of semantic segmentation model, electronic equipment and storage medium | |
| CN115830399A (en) | Classification model training method, apparatus, device, storage medium, and program product | |
| CN114494168A (en) | Model determination, image recognition and industrial quality inspection method, equipment and storage medium | |
| CN116109899B (en) | Ancient architecture repairing method, system, computer equipment and storage medium | |
| CN110874170A (en) | Image area correction method, image segmentation method and device | |
| CN115984634B (en) | Image detection method, apparatus, device, storage medium, and program product | |
| CN116109899A (en) | Ancient architecture repairing method, system, computer equipment and storage medium | |
| CN116167910B (en) | Text editing method, text editing device, computer equipment and computer readable storage medium | |
| CN110674721A (en) | Method for automatically detecting test paper layout formula | |
| KR102416714B1 (en) | System and method for city-scale tree mapping using 3-channel images and multiple deep learning | |
| Fusco et al. | An application of artificial intelligence to support the discovering of roman centuriation remains | |
| CN111459050B (en) | Intelligent simulation type nursing teaching system and teaching method based on dual-network interconnection | |
| CN115170793A (en) | Small sample image segmentation self-calibration method for industrial product quality inspection | |
| CN112734778A (en) | Vehicle matting method, system, equipment and storage medium based on neural network | |
| CN112990131A (en) | Method, device, equipment and medium for acquiring working gear of voltage change-over switch | |
| CN116935168B (en) | Method, device, computer equipment and storage medium for target detection | |
| Perbet et al. | Homogeneous superpixels from Markov random walks |
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 | ||
| GR01 | Patent grant |