CN110276756A - Road surface crack detection method, device and equipment - Google Patents
Road surface crack detection method, device and equipment Download PDFInfo
- Publication number
- CN110276756A CN110276756A CN201910555502.3A CN201910555502A CN110276756A CN 110276756 A CN110276756 A CN 110276756A CN 201910555502 A CN201910555502 A CN 201910555502A CN 110276756 A CN110276756 A CN 110276756A
- Authority
- CN
- China
- Prior art keywords
- image
- pavement
- disaggregated model
- model
- pavement image
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20021—Dividing image into blocks, subimages or windows
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Quality & Reliability (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
Abstract
The invention proposes a kind of road surface crack detection method, device and equipment, wherein method includes: to obtain the first pavement image to be detected；It is multiple second pavement images by the first pavement image cutting；Second pavement image is input in disaggregated model trained in advance and is handled, obtains the classification results of the second pavement image；The crack in the first pavement image is determined according to classification results.The accuracy of crack identification is improved as a result,.
Description
Technical field
The present invention relates to technical field of image processing more particularly to a kind of road surface crack detection methods, device and equipment.
Background technique
With constantly improve for highway infrastructures, the traffic transportation efficiency and transport capacity of highway are also constantly
It is promoted, the problem of elevator belt of highway mileage number and road transport capacity comes one and is hard to avoid: highway maintenance problem.It splits on road surface
Seam identification is most important for highway maintenance.
In the related technology, scheme one detects pavement crack based on the method that conspicuousness detects, and utilizes the neighboring area in crack
With the crack on certain significant difference (such as color difference) identification road surface in crack area face on the image, however the program
In be easy influenced by certain sundries on noise and road surface, cause crack identification accuracy to be improved.Scheme two is based on most
The scheme of small path selection detects pavement crack, by the geometry topology and rule in the image of crack, according to Topology Algorithm
To find non-closed curve, i.e., the form in general crack, crack knowledge of the program for some unconspicuous cracks and closure
Other effect is poor.
Summary of the invention
The present invention is directed to solve at least some of the technical problems in related technologies.
For this purpose, the first purpose of this invention is to propose a kind of road surface crack detection method, by by image to be detected
Cutting is multiple subgraphs, and determines the crack in image to be detected according to the classification results of subgraph, improves crack identification
Accuracy.
Second object of the present invention is to propose a kind of pavement crack detection device.
Third object of the present invention is to propose a kind of computer equipment.
Fourth object of the present invention is to propose a kind of computer readable storage medium.
First aspect present invention embodiment proposes a kind of road surface crack detection method, comprising:
Obtain the first pavement image to be detected；
It is multiple second pavement images by the first pavement image cutting；
Second pavement image is input in disaggregated model trained in advance and is handled, obtains second road surface
The classification results of image；
The crack in first pavement image is determined according to the classification results.
The road surface crack detection method of the embodiment of the present invention, by obtaining the first pavement image to be detected, and by first
Pavement image cutting is multiple second pavement images.In turn, the second pavement image is input in disaggregated model trained in advance
It is handled, obtains the classification results of the second pavement image.Further splitting in the first pavement image is determined according to classification results
Seam.As a result, by being multiple subgraphs by image to be detected cutting, and image to be detected is determined according to the classification results of subgraph
In crack, realize using deep neural network carry out crack identification, improve the accuracy of crack identification.
In addition, road surface crack detection method according to the above embodiment of the present invention can also have following supplementary technology special
Sign:
Optionally, before being handled in second pavement image to be input to disaggregated model trained in advance, also
It include: to obtain road surface sample image；It by the road surface sample image cutting is multiple target images by the method for sliding window,
Wherein, whether the target image mark is with crannied image；According to the target image of mark training depth nerve
The processing parameter of network generates the disaggregated model.
Optionally, during training deep neural network, the learning rate of the deep neural network is in preset range
Periodically variation.
Optionally, described be input to second pavement image in disaggregated model trained in advance is handled, comprising:
Model compression processing is carried out to the disaggregated model；According to compressed disaggregated model to second pavement image at
Reason.
Optionally, the method for the model compression includes at least one of beta pruning, quantization.
Optionally, described be input to second pavement image in disaggregated model trained in advance is handled, comprising:
Model optimization processing is carried out to the disaggregated model, wherein the model optimization processing includes the volume for merging the disaggregated model
Lamination and pond layer；Second pavement image is handled according to the disaggregated model after optimization.
Second aspect of the present invention embodiment proposes a kind of pavement crack detection device, comprising:
Module is obtained, for obtaining the first pavement image to be detected；
Cutting module, for being multiple second pavement images by the first pavement image cutting；
Categorization module is handled for second pavement image to be input in disaggregated model trained in advance, is obtained
Take the classification results of second pavement image；
Detection module, for determining the crack in first pavement image according to the classification results.
The pavement crack detection device of the embodiment of the present invention, by obtaining the first pavement image to be detected, and by first
Pavement image cutting is multiple second pavement images.In turn, the second pavement image is input in disaggregated model trained in advance
It is handled, obtains the classification results of the second pavement image.Further splitting in the first pavement image is determined according to classification results
Seam.As a result, by being multiple subgraphs by image to be detected cutting, and image to be detected is determined according to the classification results of subgraph
In crack, realize using deep neural network carry out crack identification, improve the accuracy of crack identification.
In addition, pavement crack detection device according to the above embodiment of the present invention can also have following supplementary technology special
Sign:
Optionally, the device further include: training module, for obtaining road surface sample image；Pass through sliding window
The road surface sample image cutting is multiple target images by method, wherein whether the target image mark is with crack
Image；According to the processing parameter of the target image of mark training deep neural network, the disaggregated model is generated.
Optionally, during training deep neural network, the learning rate of the deep neural network is in preset range
Periodically variation.
Optionally, the categorization module is specifically used for: carrying out model compression processing to the disaggregated model；After compression
Disaggregated model second pavement image is handled.
Optionally, the method for the model compression includes at least one of beta pruning, quantization.
Optionally, the categorization module is specifically used for: carrying out model optimization processing to the disaggregated model, wherein described
Model optimization processing includes the convolutional layer and pond layer for merging the disaggregated model；According to the disaggregated model after optimization to described
Two pavement images are handled.
Third aspect present invention embodiment proposes a kind of computer equipment, including processor and memory；Wherein, described
Processor is corresponding with the executable program code to run by reading the executable program code stored in the memory
Program, for realizing the road surface crack detection method as described in first aspect embodiment.
Fourth aspect present invention embodiment proposes a kind of computer readable storage medium, is stored thereon with computer journey
Sequence realizes the road surface crack detection method as described in first aspect embodiment when the program is executed by processor.
The additional aspect of the present invention and advantage will be set forth in part in the description, and will partially become from the following description
Obviously, or practice through the invention is recognized.
Detailed description of the invention
Fig. 1 is a kind of flow diagram of road surface crack detection method provided by the embodiment of the present invention；
Fig. 2 is the schematic diagram of the first pavement image；
Fig. 3 is the schematic diagram of the second pavement image；
Fig. 4 is the flow diagram of another kind road surface crack detection method provided by the embodiment of the present invention；
Fig. 5 is the flow diagram of another kind road surface crack detection method provided by the embodiment of the present invention；
Fig. 6 is a kind of structural schematic diagram of pavement crack detection device provided by the embodiment of the present invention；
Fig. 7 is the structural schematic diagram of another kind pavement crack detection device provided by the embodiment of the present invention；
Fig. 8 shows the block diagram for being suitable for the exemplary computer device for being used to realize the embodiment of the present invention.
Specific embodiment
The embodiment of the present invention is described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning to end
Same or similar label indicates same or similar element or element with the same or similar functions.Below with reference to attached
The embodiment of figure description is exemplary, it is intended to is used to explain the present invention, and is not considered as limiting the invention.
Below with reference to the accompanying drawings the road surface crack detection method, device and equipment of the embodiment of the present invention are described.
Fig. 1 is a kind of flow diagram of road surface crack detection method provided by the embodiment of the present invention, as shown in Figure 1,
This method comprises:
Step 101, the first pavement image to be detected is obtained.
In the present embodiment, when carrying out pavement crack detection, pavement image to be detected can be first obtained.For example, can be with
The image collecting devices such as camera are set on detection vehicle, when detection vehicle on road when driving, pass through image collector
Acquisition pavement image is set as the first pavement image to be detected.
It step 102, is multiple second pavement images by the first pavement image cutting.
It can be multiple by the first pavement image cutting after obtaining the first pavement image to be detected in the present embodiment
Subgraph, the subgraph that cutting is obtained is as the second pavement image.It is alternatively possible to which the image for presetting subgraph is big
It is small, and then after obtaining the first pavement image, it is the subgraph of multiple default sizes by the first pavement image cutting.Optionally,
The quantity of subgraph can also be preset, and then after obtaining the first pavement image, according to preset quantity and the first road surface figure
The size of picture determines the size of each subgraph, to be multiple subgraphs of preset quantity by the first pavement image cutting.
As an example, the size of pavement image to be detected is 3000*2000 pixel, passes through the method for sliding window
The subgraph of 600 100*100 pixels is obtained, according to the pavement image with 600 subgraphs further obtained according to cutting
Classify respectively.It should be noted that the implementation of above-mentioned image cutting is only a kind of example, do not limit specifically herein
System.
Step 103, the second pavement image is input in disaggregated model trained in advance and is handled, obtain the second road surface
The classification results of image.
In one embodiment of the invention, it is carried out in the second pavement image to be input to disaggregated model trained in advance
Before processing, it is also based on deep neural network train classification models.It is alternatively possible to road surface sample image is obtained, such as
Pavement image can be obtained from associated internet platform for another example by image acquisition device pavement of road image.
It in turn, is multiple target images by road surface sample image cutting by the method for sliding window, wherein whether target image mark
For with crannied image.Further, according to the processing parameter of the target image of mark training deep neural network, classification is generated
Model exports the classification for the image so that disaggregated model input is image, and wherein image category includes crack image and normal
Image.
It as an example, is multiple target images by road surface sample image cutting after obtaining road surface sample image, and
Image category is marked to each target image, to generate according to the target image training deep neural network of mark image category
Disaggregated model.Referring to attached drawing, Fig. 2 is road surface sample image, and Fig. 3 is the partial target image after cutting, is wherein mark on the left of Fig. 3
For note with crannied image, the right side Fig. 3 is the normal pavement image of mark.
It is appreciated that in practical applications, the form of pavement crack is varied, the thickness length in crack all there may be
Difference is difficult to be learnt by neural network if pavement crack image is larger, is difficult to meet crack when carrying out crack identification
The requirement of the accuracy of identification.It therefore, can be multiple subgraphs by image cutting, fracture pattern compares mostly in each subgraph
It is more single, be conducive to neural network learning.
Road surface sample image can also be judged after obtaining road surface sample image as a kind of possible implementation
The road surface sample image cutting is multiple target images, with basis if the image size is greater than preset threshold by image size
Multiple target pavement images are trained.Otherwise, it determines without carrying out cutting and subsequent being instructed according to the road surface sample image
Practice.
It should be noted that the above-mentioned implementation according to road surface sample image training identification model is only a kind of example,
It is not specifically limited herein.
In the present embodiment, the second pavement image can be input in disaggregated model trained in advance and be handled, obtained
The classification results of second pavement image.
As an example, cutting is carried out for the first pavement image to be detected, N number of second pavement image is obtained, by N
A second pavement image is separately input to be handled in the disaggregated model trained in advance, and each second pavement image is exported
Image category corresponding with the image, wherein image category includes crack image and normal picture, thus obtain it is N number of respectively with N
The corresponding classification results of a second pavement image.
Step 104, the crack in the first pavement image is determined according to classification results.
In the present embodiment, after the classification results for obtaining multiple second pavement images respectively, it can determine that classification results are
Several second pavement images of crack image, and then determine region position of above-mentioned second pavement image in the first pavement image
It sets, and using above-mentioned zone position as the region in crack in the first pavement image, to realize on the first road surface to be detected
Crack is identified in image and determines the location information in crack.
The road surface crack detection method of the embodiment of the present invention, by obtaining the first pavement image to be detected, and by first
Pavement image cutting is multiple second pavement images.In turn, the second pavement image is input in disaggregated model trained in advance
It is handled, obtains the classification results of the second pavement image.Further splitting in the first pavement image is determined according to classification results
Seam.As a result, by being multiple subgraphs by image to be detected cutting, and image to be detected is determined according to the classification results of subgraph
In crack, realize using deep neural network carry out crack identification, improve the accuracy of crack identification.
It based on the above embodiment, further, can be with root in order to meet the requirement of real-time of pavement crack detection project
According to the method for model compression and model optimization, the efficiency of crack identification is improved.
Fig. 4 is the flow diagram of another kind road surface crack detection method provided by the embodiment of the present invention, such as Fig. 4 institute
Show, this method comprises:
Step 201, model compression processing is carried out to disaggregated model.
Wherein, the method for model compression includes but is not limited to beta pruning, quantization etc..
As an example, by taking beta pruning as an example, the corresponding weighted value of available each branch of model, and by weighted value and pre-
If threshold value is compared, the branch that weighted value is more than or equal to preset threshold is retained, weighted value is less than to the branch of preset threshold
Delete, and then construct new model according to remaining branch, so that implementation model compresses, can reduce model storage space and
Accelerate the inference speed of model.
As another example, by taking quantization as an example, model parameter range is 1-100, is indicated by 32 floating numbers.It will ginseng
Number range is quantized to 1-10, so as to be indicated by 16 floating numbers, it is possible thereby to which the compression of implementation model, can reduce mould
The storage space of type and the inference speed for accelerating model.
Step 202, the second pavement image is handled according to compressed disaggregated model.
In the present embodiment, by carrying out model compression processing to disaggregated model, and compressed disaggregated model is obtained.Into
And the second pavement image can be handled according to compressed disaggregated model, obtain corresponding with the second pavement image point
Class is as a result, promote the inference speed of disaggregated model, to promote the efficiency of pavement crack identification.Wherein, above-mentioned model compression side
Method, which can according to need, to be combined.
Fig. 5 is the flow diagram of another kind road surface crack detection method provided by the embodiment of the present invention, such as Fig. 5 institute
Show, this method comprises:
Step 301, model optimization processing is carried out to disaggregated model.
As a kind of possible implementation, optimized using reasoning process of the TensorRT to model, it specifically, can
To be merged to network layer annexable in model.For example, preset convolutional layer in available model, and by preset volume
Lamination merges, wherein preset convolutional layer can be greater than the convolutional layer of preset value for similarity.And the pond in acquisition model
Change layer and convolutional layer, and pond layer and convolutional layer are merged, to generate the disaggregated model of TensorRT form.
Step 302, the second pavement image is handled according to the disaggregated model after optimization.
In the present embodiment, by carrying out model optimization processing to disaggregated model, and the disaggregated model after optimization is obtained.Into
And the second pavement image can be handled according to the disaggregated model after optimization, obtain corresponding with the second pavement image point
Class is as a result, promote the inference speed of disaggregated model, to promote the efficiency of pavement crack identification
It is alternatively possible to crack identification is carried out for example, by using MobileNet using the deep neural network of lightweight, from
And while guaranteeing accuracy rate lift scheme the speed of service.
In one embodiment of the invention, during according to target image training deep neural network, depth mind
The cyclically-varying within a preset range of learning rate through network.It is alternatively possible to preset the maximum value and minimum of learning rate
Value, Schistosomiasis control rate cyclically-varying between the maximum value and minimum value during training neural network, to be promoted
The convergence rate of model training, prevents model from falling into saddle point.
In one embodiment of the invention, deformable convolutional coding structure can be increased in deep neural network, for example, can
Increase deformable convolutional coding structure with upper one layer of convolutional layer in a model, to consider horizontally and vertically inclined when calculating convolution
It moves.It is directed to the crack of different shape as a result, by the way that deformable convolution to be added in neural network, so as to preferably identify
Different crack information improves model for the recognition accuracy in the crack of different shape scale.
The road surface crack detection method of the embodiment of the present invention improves classification by the method for model compression and model optimization
The inference speed of model meets the requirement of real-time of pavement crack identification to promote the efficiency of pavement crack identification.
In order to realize above-described embodiment, the present invention also proposes a kind of pavement crack detection device.
Fig. 6 is a kind of structural schematic diagram of pavement crack detection device provided by the embodiment of the present invention, as shown in fig. 6,
The device includes: to obtain module 100, cutting module 200, categorization module 300, detection module 400.
Wherein, module 100 is obtained, for obtaining the first pavement image to be detected.
Cutting module 200, for being multiple second pavement images by the first pavement image cutting.
Categorization module 300 is handled for the second pavement image to be input in disaggregated model trained in advance, is obtained
The classification results of second pavement image.
Detection module 400, for determining the crack in the first pavement image according to classification results.
On the basis of Fig. 6, device shown in Fig. 7 further include: training module 500.
Wherein, training module 500, for obtaining road surface sample image；By the method for sliding window by road surface sample graph
As cutting is multiple target images, wherein whether target image mark is with crannied image；According to the target image of mark
The processing parameter of training deep neural network, generates disaggregated model.
In one embodiment of the invention, categorization module 300 is specifically used for: carrying out at model compression to disaggregated model
Reason；The second pavement image is handled according to compressed disaggregated model.
In one embodiment of the invention, categorization module 300 is specifically used for: carrying out at model optimization to disaggregated model
Reason, wherein model optimization processing includes the convolutional layer and pond layer for merging disaggregated model；According to the disaggregated model after optimization to
Two pavement images are handled.
Optionally, during training deep neural network, the learning rate of deep neural network is all within a preset range
The variation of phase property.
Optionally, the method for model compression includes at least one of beta pruning, quantization.
It should be noted that the explanation of previous embodiment road pavement crack detection method is equally applicable to the present embodiment
Pavement crack detection device, details are not described herein again.
The pavement crack detection device of the embodiment of the present invention, by obtaining the first pavement image to be detected, and by first
Pavement image cutting is multiple second pavement images.In turn, the second pavement image is input in disaggregated model trained in advance
It is handled, obtains the classification results of the second pavement image.Further splitting in the first pavement image is determined according to classification results
Seam.As a result, by being multiple subgraphs by image to be detected cutting, and image to be detected is determined according to the classification results of subgraph
In crack, realize using deep neural network carry out crack identification, improve the accuracy of crack identification.In addition, will
Image cutting is multiple subgraphs, and fracture pattern is mostly relatively simple in each subgraph, such as the crack in subgraph is most
For the form that a line passes through, be conducive to neural network learning.
In order to realize above-described embodiment, the present invention also proposes a kind of computer equipment, including processor and memory；Its
In, processor runs journey corresponding with executable program code by reading the executable program code stored in memory
Sequence, for realizing the road surface crack detection method as described in aforementioned any embodiment.
In order to realize above-described embodiment, the present invention also proposes a kind of computer program product, when in computer program product
Instruction the road surface crack detection method as described in aforementioned any embodiment is realized when being executed by processor.
In order to realize above-described embodiment, the present invention also proposes a kind of computer readable storage medium, is stored thereon with calculating
Machine program realizes the road surface crack detection method as described in aforementioned any embodiment when the program is executed by processor.
Fig. 8 shows the block diagram for being suitable for the exemplary computer device for being used to realize the embodiment of the present invention.The meter that Fig. 8 is shown
Calculating machine equipment 12 is only an example, should not function to the embodiment of the present invention and use scope bring any restrictions.
As shown in figure 8, computer equipment 12 is showed in the form of universal computing device.The component of computer equipment 12 can be with
Including but not limited to: one or more processor or processing unit 16, system storage 28 connect different system components
The bus 18 of (including system storage 28 and processing unit 16).
Bus 18 indicates one of a few class bus structures or a variety of, including memory bus or Memory Controller,
Peripheral bus, graphics acceleration port, processor or the local bus using any bus structures in a variety of bus structures.It lifts
For example, these architectures include but is not limited to industry standard architecture (Industry Standard
Architecture；Hereinafter referred to as: ISA) bus, microchannel architecture (Micro Channel Architecture；Below
Referred to as: MAC) bus, enhanced isa bus, Video Electronics Standards Association (Video Electronics Standards
Association；Hereinafter referred to as: VESA) local bus and peripheral component interconnection (Peripheral Component
Interconnection；Hereinafter referred to as: PCI) bus.
Computer equipment 12 typically comprises a variety of computer system readable media.These media can be it is any can be by
The usable medium that computer equipment 12 accesses, including volatile and non-volatile media, moveable and immovable medium.
Memory 28 may include the computer system readable media of form of volatile memory, such as random access memory
Device (Random Access Memory；Hereinafter referred to as: RAM) 30 and/or cache memory 32.Computer equipment 12 can be with
It further comprise other removable/nonremovable, volatile/non-volatile computer system storage mediums.Only as an example,
Storage system 34 can be used for reading and writing immovable, non-volatile magnetic media, and (Fig. 8 do not show, commonly referred to as " hard drive
Device ").Although being not shown in Fig. 8, the disk for reading and writing to removable non-volatile magnetic disk (such as " floppy disk ") can be provided and driven
Dynamic device, and to removable anonvolatile optical disk (such as: compact disc read-only memory (Compact Disc Read Only
Memory；Hereinafter referred to as: CD-ROM), digital multi CD-ROM (Digital Video Disc Read Only
Memory；Hereinafter referred to as: DVD-ROM) or other optical mediums) read-write CD drive.In these cases, each driving
Device can be connected by one or more data media interfaces with bus 18.Memory 28 may include that at least one program produces
Product, the program product have one group of (for example, at least one) program module, and it is each that these program modules are configured to perform the application
The function of embodiment.
Program/utility 40 with one group of (at least one) program module 42 can store in such as memory 28
In, such program module 42 include but is not limited to operating system, one or more application program, other program modules and
It may include the realization of network environment in program data, each of these examples or certain combination.Program module 42 is usual
Execute the function and/or method in embodiments described herein.
Computer equipment 12 can also be with one or more external equipments 14 (such as keyboard, sensing equipment, display 24
Deng) communication, the equipment interacted with the computer system/server 12 can be also enabled a user to one or more to be communicated, and/
Or with enable the computer system/server 12 and one or more of the other any equipment (example for being communicated of calculating equipment
Such as network interface card, modem etc.) communication.This communication can be carried out by input/output (I/O) interface 22.Also, it calculates
Machine equipment 12 can also pass through network adapter 20 and one or more network (such as local area network (Local Area
Network；Hereinafter referred to as: LAN), wide area network (Wide Area Network；Hereinafter referred to as: WAN) and/or public network, example
Such as internet) communication.As shown, network adapter 20 is communicated by bus 18 with other modules of computer equipment 12.It answers
When understanding, although not shown in the drawings, other hardware and/or software module can be used in conjunction with computer equipment 12, including but not
Be limited to: microcode, device driver, redundant processing unit, external disk drive array, RAID system, tape drive and
Data backup storage system etc..
Processing unit 16 by the program that is stored in system storage 28 of operation, thereby executing various function application and
Data processing, such as realize the method referred in previous embodiment.
In the description of the present invention, it is to be understood that, term " first ", " second " are used for description purposes only, and cannot
It is interpreted as indication or suggestion relative importance or implicitly indicates the quantity of indicated technical characteristic.Define as a result, " the
One ", the feature of " second " can explicitly or implicitly include at least one of the features.In the description of the present invention, " multiple "
It is meant that at least two, such as two, three etc., unless otherwise specifically defined.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means specific features, structure, material or spy described in conjunction with this embodiment or example
Point is included at least one embodiment or example of the invention.In the present specification, schematic expression of the above terms are not
It must be directed to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described can be in office
It can be combined in any suitable manner in one or more embodiment or examples.In addition, without conflicting with each other, the skill of this field
Art personnel can tie the feature of different embodiments or examples described in this specification and different embodiments or examples
It closes and combines.
Although the embodiments of the present invention has been shown and described above, it is to be understood that above-described embodiment is example
Property, it is not considered as limiting the invention, those skilled in the art within the scope of the invention can be to above-mentioned
Embodiment is changed, modifies, replacement and variant.
Claims (12)
1. a kind of road surface crack detection method characterized by comprising
Obtain the first pavement image to be detected；
It is multiple second pavement images by the first pavement image cutting；
Second pavement image is input in disaggregated model trained in advance and is handled, obtains second pavement image
Classification results；
The crack in first pavement image is determined according to the classification results.
2. the method as described in claim 1, which is characterized in that in point that second pavement image is input to training in advance
Before being handled in class model, further includes:
Obtain road surface sample image；
It by the road surface sample image cutting is multiple target images by the method for sliding window, wherein the target image
Whether mark is with crannied image；
According to the processing parameter of the target image of mark training deep neural network, the disaggregated model is generated.
3. method according to claim 2, which is characterized in that during training deep neural network, the depth mind
The cyclically-varying within a preset range of learning rate through network.
4. the method as described in claim 1, which is characterized in that described that second pavement image is input to training in advance
It is handled in disaggregated model, comprising:
Model compression processing is carried out to the disaggregated model；
Second pavement image is handled according to compressed disaggregated model.
5. method as claimed in claim 4, which is characterized in that the method for the model compression include beta pruning, in quantization extremely
Few one kind.
6. the method as described in claim 1, which is characterized in that described that second pavement image is input to training in advance
It is handled in disaggregated model, comprising:
Model optimization processing is carried out to the disaggregated model, wherein the model optimization processing includes merging the disaggregated model
Convolutional layer and pond layer；
Second pavement image is handled according to the disaggregated model after optimization.
7. a kind of pavement crack detection device characterized by comprising
Module is obtained, for obtaining the first pavement image to be detected；
Cutting module, for being multiple second pavement images by the first pavement image cutting；
Categorization module is handled for second pavement image to be input in disaggregated model trained in advance, obtains institute
State the classification results of the second pavement image；
Detection module, for determining the crack in first pavement image according to the classification results.
8. device as claimed in claim 7, which is characterized in that further include:
Training module, for obtaining road surface sample image；
It by the road surface sample image cutting is multiple target images by the method for sliding window, wherein the target image
Whether mark is with crannied image；
According to the processing parameter of the target image of mark training deep neural network, the disaggregated model is generated.
9. device as claimed in claim 7, which is characterized in that the categorization module is specifically used for:
Model compression processing is carried out to the disaggregated model；
Second pavement image is handled according to compressed disaggregated model.
10. device as claimed in claim 7, which is characterized in that the categorization module is specifically used for:
Model optimization processing is carried out to the disaggregated model, wherein the model optimization processing includes merging the disaggregated model
Convolutional layer and pond layer；
Second pavement image is handled according to the disaggregated model after optimization.
11. a kind of computer equipment, which is characterized in that including processor and memory；
Wherein, the processor is run by reading the executable program code stored in the memory can be performed with described
The corresponding program of program code, for realizing road surface crack detection method such as of any of claims 1-6.
12. a kind of non-transitorycomputer readable storage medium, is stored thereon with computer program, which is characterized in that the program
Such as road surface crack detection method of any of claims 1-6 is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910555502.3A CN110276756A (en) | 2019-06-25 | 2019-06-25 | Road surface crack detection method, device and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910555502.3A CN110276756A (en) | 2019-06-25 | 2019-06-25 | Road surface crack detection method, device and equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110276756A true CN110276756A (en) | 2019-09-24 |
Family
ID=67963154
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910555502.3A Pending CN110276756A (en) | 2019-06-25 | 2019-06-25 | Road surface crack detection method, device and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110276756A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111784667A (en) * | 2020-06-30 | 2020-10-16 | 北京海益同展信息科技有限公司 | Crack identification method and device |
| CN112037205A (en) * | 2020-09-01 | 2020-12-04 | 湖北微特传感物联研究院有限公司 | Crane crack detection method based on deep learning |
| CN112215819A (en) * | 2020-10-13 | 2021-01-12 | 中国民航大学 | Airport pavement crack detection method based on depth feature fusion |
| CN113128664A (en) * | 2021-03-16 | 2021-07-16 | 广东电力信息科技有限公司 | Neural network compression method, device, electronic equipment and storage medium |
| CN114429464A (en) * | 2022-01-25 | 2022-05-03 | 众安在线财产保险股份有限公司 | Screen-breaking identification method of terminal and related equipment |
| CN114565793A (en) * | 2022-02-28 | 2022-05-31 | 湖南北斗微芯产业发展有限公司 | Road traffic crack monitoring method and system |
| CN115830007A (en) * | 2023-02-06 | 2023-03-21 | 山东省滨州公路工程有限公司 | Pavement void detection method and system based on image recognition |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160104056A1 (en) * | 2014-10-09 | 2016-04-14 | Microsoft Technology Licensing, Llc | Spatial pyramid pooling networks for image processing |
| CN106097315A (en) * | 2016-06-03 | 2016-11-09 | 河海大学常州校区 | A kind of underwater works crack extract method based on sonar image |
| CN106127802A (en) * | 2016-06-16 | 2016-11-16 | 南京邮电大学盐城大数据研究院有限公司 | A kind of movement objective orbit method for tracing |
| CN106910186A (en) * | 2017-01-13 | 2017-06-30 | 陕西师范大学 | A kind of Bridge Crack detection localization method based on CNN deep learnings |
| CN106934795A (en) * | 2017-01-23 | 2017-07-07 | 陕西师范大学 | The automatic testing method and Forecasting Methodology of a kind of glue into concrete beam cracks |
| CN108197671A (en) * | 2018-02-02 | 2018-06-22 | 石家庄铁道大学 | A kind of distress in concrete detection method and device represented based on compressive features |
| CN108765386A (en) * | 2018-05-16 | 2018-11-06 | 中铁科学技术开发公司 | A kind of tunnel slot detection method, device, electronic equipment and storage medium |
| CN109829371A (en) * | 2018-12-26 | 2019-05-31 | 深圳云天励飞技术有限公司 | A kind of method for detecting human face and device |
-
2019
- 2019-06-25 CN CN201910555502.3A patent/CN110276756A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160104056A1 (en) * | 2014-10-09 | 2016-04-14 | Microsoft Technology Licensing, Llc | Spatial pyramid pooling networks for image processing |
| CN106097315A (en) * | 2016-06-03 | 2016-11-09 | 河海大学常州校区 | A kind of underwater works crack extract method based on sonar image |
| CN106127802A (en) * | 2016-06-16 | 2016-11-16 | 南京邮电大学盐城大数据研究院有限公司 | A kind of movement objective orbit method for tracing |
| CN106910186A (en) * | 2017-01-13 | 2017-06-30 | 陕西师范大学 | A kind of Bridge Crack detection localization method based on CNN deep learnings |
| CN106934795A (en) * | 2017-01-23 | 2017-07-07 | 陕西师范大学 | The automatic testing method and Forecasting Methodology of a kind of glue into concrete beam cracks |
| CN108197671A (en) * | 2018-02-02 | 2018-06-22 | 石家庄铁道大学 | A kind of distress in concrete detection method and device represented based on compressive features |
| CN108765386A (en) * | 2018-05-16 | 2018-11-06 | 中铁科学技术开发公司 | A kind of tunnel slot detection method, device, electronic equipment and storage medium |
| CN109829371A (en) * | 2018-12-26 | 2019-05-31 | 深圳云天励飞技术有限公司 | A kind of method for detecting human face and device |
Non-Patent Citations (1)
| Title |
|---|
| 王功鹏 等: "基于卷积神经网络的随机梯度下降算法", 《计算机工程与设计》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111784667A (en) * | 2020-06-30 | 2020-10-16 | 北京海益同展信息科技有限公司 | Crack identification method and device |
| CN111784667B (en) * | 2020-06-30 | 2023-09-26 | 京东科技信息技术有限公司 | Crack identification method and device |
| CN112037205A (en) * | 2020-09-01 | 2020-12-04 | 湖北微特传感物联研究院有限公司 | Crane crack detection method based on deep learning |
| CN112215819A (en) * | 2020-10-13 | 2021-01-12 | 中国民航大学 | Airport pavement crack detection method based on depth feature fusion |
| CN112215819B (en) * | 2020-10-13 | 2023-06-30 | 中国民航大学 | Airport pavement crack detection method based on depth feature fusion |
| CN113128664A (en) * | 2021-03-16 | 2021-07-16 | 广东电力信息科技有限公司 | Neural network compression method, device, electronic equipment and storage medium |
| CN114429464A (en) * | 2022-01-25 | 2022-05-03 | 众安在线财产保险股份有限公司 | Screen-breaking identification method of terminal and related equipment |
| CN114565793A (en) * | 2022-02-28 | 2022-05-31 | 湖南北斗微芯产业发展有限公司 | Road traffic crack monitoring method and system |
| CN114565793B (en) * | 2022-02-28 | 2023-05-23 | 湖南北斗微芯产业发展有限公司 | Road traffic crack monitoring method and system |
| CN115830007A (en) * | 2023-02-06 | 2023-03-21 | 山东省滨州公路工程有限公司 | Pavement void detection method and system based on image recognition |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110276756A (en) | Road surface crack detection method, device and equipment | |
| Lim et al. | Real-time traffic sign recognition based on a general purpose GPU and deep-learning | |
| US9965719B2 (en) | Subcategory-aware convolutional neural networks for object detection | |
| WO2019228211A1 (en) | Lane-line-based intelligent driving control method and apparatus, and electronic device | |
| CN111582008B (en) | Device and method for training classification model and device for classifying by using classification model | |
| CN109409252A (en) | A kind of traffic multi-target detection method based on modified SSD network | |
| CN111899515B (en) | Vehicle detection system based on wisdom road edge calculates gateway | |
| CN111242015A (en) | Method for predicting driving danger scene based on motion contour semantic graph | |
| CN110598703B (en) | OCR (optical character recognition) method and device based on deep neural network | |
| CN110580499B (en) | Deep learning target detection method and system based on crowdsourcing repeated labels | |
| Saleh et al. | Traffic signs recognition and distance estimation using a monocular camera | |
| CN111611933A (en) | Information extraction method and system for document image | |
| Jensen et al. | Traffic light detection at night: Comparison of a learning-based detector and three model-based detectors | |
| CN113724286A (en) | Method and device for detecting saliency target and computer-readable storage medium | |
| CN116982089A (en) | Method and system for image semantic enhancement | |
| Song et al. | A novel deep learning network for accurate lane detection in low-light environments | |
| Hassan et al. | Learning deep feature fusion for traffic light detection | |
| WO2024012234A1 (en) | Target detection method, computer device, computer-readable storage medium and vehicle | |
| CN103136536A (en) | System and method for detecting target and method for exacting image features | |
| Alam et al. | Faster RCNN based robust vehicle detection algorithm for identifying and classifying vehicles | |
| Sadakatul Bari et al. | Performance evaluation of convolution neural network based object detection model for Bangladeshi traffic vehicle detection | |
| CN116246128B (en) | Training method and device of detection model crossing data sets and electronic equipment | |
| Maity et al. | An approach for detecting circular callouts in architectural, engineering and constructional drawing documents | |
| Fadili et al. | Fastest Moroccan License Plate Recognition Using a Lightweight Modified YOLOv5 Model | |
| CN114639037B (en) | Method for determining vehicle saturation of high-speed service area and electronic equipment |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20211014 Address after: 105 / F, building 1, No. 10, Shangdi 10th Street, Haidian District, Beijing 100085 Applicant after: Apollo Intelligent Technology (Beijing) Co.,Ltd. Address before: 100085 Baidu Building, 10 Shangdi Tenth Street, Haidian District, Beijing Applicant before: BAIDU ONLINE NETWORK TECHNOLOGY (BEIJING) Co.,Ltd. |
|
| TA01 | Transfer of patent application right |