CN115512322A - Multidimensional pavement damage data processing method - Google Patents
Multidimensional pavement damage data processing method Download PDFInfo
- Publication number
- CN115512322A CN115512322A CN202211193198.0A CN202211193198A CN115512322A CN 115512322 A CN115512322 A CN 115512322A CN 202211193198 A CN202211193198 A CN 202211193198A CN 115512322 A CN115512322 A CN 115512322A
- Authority
- CN
- China
- Prior art keywords
- dimensional
- image
- dimensional image
- damage
- elevation
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
-
- 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
- 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
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Image Processing (AREA)
Abstract
The invention relates to a multidimensional pavement damage data processing method, which comprises the following steps: processing the acquired two-dimensional gray level image by using a two-dimensional image preprocessing algorithm to obtain a processed two-dimensional image; processing the acquired three-dimensional elevation image by using a three-dimensional image preprocessing algorithm to obtain a three-dimensional image enhancement output result; and taking the processed two-dimensional image and the three-dimensional image enhanced output result as the input of a pavement damage disease identification model, and positioning and automatically and quickly detecting a pavement damage area. Compared with the prior art, the method and the device solve the problems of image data loss, complex data characteristics and nonuniform data distribution, and effectively improve the accuracy and efficiency of disease identification.
Description
Technical Field
The invention belongs to the technical field of pavement information acquisition and detection, and relates to a multidimensional pavement damage data processing method.
Background
At present, road traffic is more and more crowded and complex, so that different types of diseases such as cracks, pits, tracks, crack pouring, block repairing, loosening and the like can occur on the road surface. In order to avoid affecting daily traffic and driving safety of people, the damage condition of the road surface needs to be regularly detected, and diseases need to be repaired in time. In the prior art, a one-way line structured light scanning technology is generally adopted for intelligent detection of road surface damage, but the limitation from the technical principle exists, on one hand, the optical characteristics of adjacent longitudinal sections are different, and on the other hand, the elevation of the adjacent longitudinal sections is superposed with vehicle vibration displacement, so that the data characteristics of two-dimensional gray scale data and three-dimensional elevation data output by a system tend to be complex. The road surface data with irregular characteristics greatly increase the difficulty of intelligent identification of road surface diseases. In addition, a relatively complete evaluation system is not formed in the aspect of pavement damage detection, and the automatic disease classification method cannot achieve satisfactory effects in the aspects of identification effect, category coverage, real-time processing and the like.
Disclosure of Invention
The invention aims to provide a multidimensional road surface damage data processing method to overcome the problems of image data missing and low recognition accuracy and efficiency.
The purpose of the invention can be realized by the following technical scheme:
a multidimensional pavement damage data processing method comprises the following steps:
processing the acquired two-dimensional gray level image by using a two-dimensional image preprocessing algorithm to obtain a processed two-dimensional image;
processing the acquired three-dimensional elevation image by using a three-dimensional image preprocessing algorithm to obtain a three-dimensional image enhancement output result;
and taking the processed two-dimensional image and the three-dimensional image enhanced output result as the input of a pavement damage disease identification model, and positioning and automatically and quickly detecting a pavement damage area.
Further, the two-dimensional image preprocessing algorithm comprises:
step S310, calculating an initial preprocessing coefficient C 0 ,
Wherein, the first and the second end of the pipe are connected with each other,is the initial overall mean value, benchmark is the set reference gray value Benchmark;
step S320, obtaining a two-dimensional road gray image I, judging whether the two-dimensional road gray image contains abnormal gray information or not based on a set threshold, if so, rejecting the image, otherwise, adding the image into an image row to obtain a gray average Ave of the image row current ;
Step S330, based on the preprocessing coefficient and the gray average value Ave current Acquiring a processed two-dimensional image I of a two-dimensional gray level image I of the road surface new ,
I new =I×C
Wherein C is a preprocessing coefficient, and if I is the first image to be processed, C is C 0 Otherwise, C is updated by the following formula:
Ave Back =αAve Back +(1-α)Ave current
And step S340, returning to step S320 until all image data processing is finished.
acquiring a two-dimensional gray image of a road surface, eliminating the image containing abnormal gray information based on a set threshold value, and extracting N images which can be used for calculating a preprocessing coefficient C 0 Calculating the gray average of N images as the initial overall average
Further, the three-dimensional data preprocessing algorithm comprises:
s410, processing the relative elevation data of the road surface obtained through data calibration by adopting a pixel left-right adjacent interpolation method to obtain a three-dimensional image elevation value point set after filtering;
and step S420, processing the elevation value point set of the original three-dimensional image and the filtered elevation value point set of the three-dimensional image by using a three-dimensional image enhancement function to obtain a three-dimensional image enhancement output result.
Furthermore, the data calibration establishes a conversion relation between pixels and space coordinates for the two-dimensional gray scale image and the three-dimensional elevation image by means of a line structured light principle, and generates a calibration parameter file for obtaining relative elevation data of the road surface.
Furthermore, the method for interpolating the adjacent pixels on the left and right sides comprises the following steps,
wherein, X = { X 1 ,X 2 ,...,X w Is the per-row cross-sectional pixel matrix, X w =0 (W =1,2,..., W) is the outlier set, X w Not equal to 0 (W =1,2.,. W) is the normal set of points, X '= { X' 1 ,X′ 2 ,...,X′ w Is a set of elevation value points after filtering, W is cross section elevationThe number of dots.
Further, the calculation formula of the three-dimensional image enhancement function is as follows:
A′(x,y)=A+MFF k |A|+MFF k |A(x-I,y-j)|
wherein, A represents a matrix of an original three-dimensional image elevation value point set, A 'represents a matrix of a filtered three-dimensional image elevation value point set, MFF is an adaptive mean value filtering technology, K is a filtering coefficient, and (2K + 1) × (2K + 1) is taken as a filtering template, A' min Represents the minimum value of the set A ', A' max Represents the maximum value of the A 'set, A' represents the enhanced output result of the three-dimensional image, i, j belongs to [ -K, K [ ]]。
Further, the construction method of the road surface damage disease identification model comprises the following steps:
step S510, marking the damage of a high-quality two-dimensional and three-dimensional image containing the damage of the pavement;
s520, superposing the marked two-dimensional and three-dimensional images to obtain a two-channel training sample database;
and S530, training a deep learning network model by using a dual-channel training sample database to obtain a road surface damage and disease identification model.
Further, the road surface damage and disease identification model is superposed with a conditional random field module and used for improving the precision of the road surface damage and disease identification model.
Further, the damage diseases of the pavement comprise cracks, pot holes, ruts, crack pouring, block repairing and loosening.
Compared with the prior art, the invention has the following characteristics:
1. the invention respectively develops a two-dimensional image preprocessing algorithm and a three-dimensional image preprocessing algorithm aiming at two-dimensional gray level images and three-dimensional elevation data, and can effectively solve the problems of image data loss, complex data characteristics and non-uniform data distribution caused by vehicle jolting, mechanical vibration, uneven laser intensity distribution and the like.
2. According to the method, the deep learning network model is trained by using the two-channel training sample database to obtain the pavement damage disease identification model, and various pavement damage diseases such as cracks, pits, tracks, crack pouring, block repairing and loosening can be quickly, accurately and widely identified.
3. The invention establishes the conversion relation between the pixels and the space by means of the structured light principle to calibrate the data, and realizes higher precision of the obtained pavement relative elevation data.
4. The method uses the conditional random field module to be superposed on the pavement damage and disease identification model, and improves the precision of the pavement damage and disease identification model.
Drawings
FIG. 1 is a flow chart of a multi-dimensional road surface damage data processing method.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. The present embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation manner and a specific operation process are given, but the scope of the present invention is not limited to the following embodiments.
The embodiment is as follows:
the embodiment provides a multidimensional road surface damage data processing method, as shown in fig. 1, including the following steps:
s1, data calibration: based on a line structured light principle, establishing a conversion relation between pixels and space coordinates for a two-dimensional gray image and a three-dimensional elevation image, and generating a calibration parameter file for obtaining relative elevation data of a pavement;
s2, data acquisition: after data calibration, continuously acquiring multi-dimensional data of the cross section of the asphalt pavement in a vehicle-mounted measuring mode, and loading a calibration parameter file to obtain relative elevation data of the pavement;
s3, preprocessing a two-dimensional image: processing the acquired two-dimensional gray image by using a two-dimensional image preprocessing algorithm, eliminating the conditions of data gray unevenness and data loss, and obtaining a processed two-dimensional image;
s4, three-dimensional image preprocessing: executing the steps S3 simultaneously, processing the acquired three-dimensional elevation image by using a three-dimensional image preprocessing algorithm, performing abnormal correction and elevation calibration on the relative elevation data of the road surface, eliminating the conditions of uneven depth distribution and data loss of the data and obtaining a three-dimensional image enhancement output result;
step S5: and taking the processed two-dimensional image and the three-dimensional image enhanced output result as the input of a pavement damage disease identification model, and positioning and automatically and quickly detecting a pavement damage area.
The step S2 includes the following steps:
and step S210, triggering a camera by using a vehicle-mounted transverse line structure light imaging system and using a wheel universal encoder, and converting the driving mileage and the driving speed in unit time to be used as a file name for storing multi-source data.
Step S220, the obtained multi-source data has two channels in total, wherein the channel 1 is a two-dimensional gray image, the channel 2 is a three-dimensional elevation image, the two data sources have the same size and are both M H *N z Obtaining the relative elevation data of the road surface after calibration by loading the calibration parameter file of the step S210,
the conversion relation between the pixels and the space is established by means of the structured light principle to carry out data calibration, and the obtained relative elevation data of the road surface is high in precision.
The step S3 includes the following steps:
step S310, calculating an initial preprocessing coefficient C 0 ,
Wherein, the first and the second end of the pipe are connected with each other,is the initial overall mean value, benchmark is the set reference gray value Benchmark;
step S320, acquiring a two-dimensional gray level image I of the road surface, and judging based on a set threshold valueWhether the two-dimensional gray level image of the broken road surface contains abnormal gray level information or not is judged, if yes, the image is removed, if not, the image is added into the image row, and the gray level average value Ave of the image row is obtained current ;
Step S330, based on the preprocessing coefficient and the average value Ave of the gray level current Acquiring a processed two-dimensional image I of a two-dimensional gray level image I of the road surface new ,
I new =I×C
Wherein C is a preprocessing coefficient, and if I is the first image to be processed, C is C 0 Otherwise, C is updated by the following formula:
Ave Back =αAve Back +(1-α)Ave current
And step S340, returning to step S320 until all image data processing is finished.
acquiring a two-dimensional gray image of a road surface, eliminating the image containing abnormal gray information based on a set threshold value, and extracting N images which can be used for calculating a preprocessing coefficient C 0 Calculating the gray average of N images as the initial overall average
The step S4 includes the following steps:
s410, processing the relative elevation data of the road surface obtained through data calibration by adopting a pixel left-right adjacent interpolation method to obtain a three-dimensional image elevation value point set after filtering;
step S420, processing the elevation value point set of the original three-dimensional image and the filtered elevation value point set of the three-dimensional image by using a three-dimensional image enhancement function to obtain a three-dimensional image enhancement output result.
Wherein, the interpolation method for the left and right adjacent pixels comprises
Wherein, X = { X 1 ,X 2 ,...,X w Is a matrix of cross-sectional pixels per row, X w =0 (W =1,2.., W) is the set of outliers, X w Not equal to 0 (W =1,2.,. W) is the normal set of points, X '= { X' 1 ,X′ 2 ,...,X′ w And W is the number of elevation points of the cross section.
The calculation formula of the three-dimensional image enhancement function is as follows:
A′(x,y)=A+MFF k |A|+MFF k |A(x-I,y-j)|
wherein, A represents a matrix of an original three-dimensional image elevation value point set, A 'represents a matrix of a filtered three-dimensional image elevation value point set, MFF is an adaptive mean value filtering technology, K is a filtering coefficient, and (2K + 1) × (2K + 1) is taken as a filtering template, A' min Represents the minimum value of the A 'set, A' max Represents the maximum value of the A 'set, A' represents the three-dimensional image enhancement output result, i, j E [ -K, K [ ]]。
The three-dimensional image enhancement function is used for eliminating unevenness of elevation.
The data calibration establishes the conversion relation between the pixels and the space by means of the structured light principle, has high measurement precision, and can assist the vehicle-mounted line structured light system to acquire two-dimensional gray scale images and three-dimensional high-precision data of the road surface.
The image preprocessing algorithm is respectively researched and developed for the two-dimensional gray scale image and the three-dimensional elevation data in the step S3 and the step S4, so that the problems of image data loss, complex data characteristics and non-uniform data distribution caused by vehicle jolting, mechanical vibration, laser distribution nonuniformity and the like can be effectively solved.
The step S5 includes the steps of:
step S510, marking the damage of a high-quality two-dimensional and three-dimensional image containing the damage of the pavement;
s520, superposing the marked two-dimensional and three-dimensional images to obtain a two-channel training sample database;
and S530, training a deep learning network model by using a double-channel training sample database to obtain a road surface damage and disease identification model.
The damage diseases of the pavement comprise diseases types such as cracks, pits, tracks, crack pouring, block repairing, loosening and the like.
Specifically, high quality two-dimensional and three-dimensional images containing diseases are screened manually.
By using the two-channel training sample database to train the deep learning network model, the pavement damage disease identification model is obtained, and various pavement damage diseases such as cracks, pits, tracks, crack pouring, block repairing and loosening can be quickly, accurately and widely identified.
After the deep learning network model is trained by using the dual-channel training sample database, a semantic segmentation network model is constructed by using a pyramid scene analysis network (PSPNet) and U network (U-Net) model fusion strategy, and a road surface damage and disease recognition model with high precision is obtained by considering a fusion Condition Random Field (CRF) module.
The above method, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The embodiments described above are described to facilitate an understanding and use of the invention by those skilled in the art. It will be readily apparent to those skilled in the art that various modifications to these embodiments may be made, and the generic principles described herein may be applied to other embodiments without the use of the inventive faculty. Therefore, the present invention is not limited to the above embodiments, and those skilled in the art should make improvements and modifications within the scope of the present invention based on the disclosure of the present invention.
Claims (10)
1. A multidimensional pavement damage data processing method is characterized by comprising the following steps:
processing the acquired two-dimensional gray level image by using a two-dimensional image preprocessing algorithm to obtain a processed two-dimensional image;
processing the acquired three-dimensional elevation image by using a three-dimensional image preprocessing algorithm to obtain a three-dimensional image enhancement output result;
and taking the processed two-dimensional image and the three-dimensional image enhanced output result as the input of a pavement damage disease identification model, and positioning and automatically and quickly detecting a pavement damage area.
2. The method of claim 1, wherein the two-dimensional image preprocessing algorithm comprises:
step S310, calculating an initial preprocessing coefficient C 0 ,
Wherein, the first and the second end of the pipe are connected with each other,is an initial overall mean value, and the Benchmark is a set reference gray value Benchmark;
step S320, acquiring a two-dimensional pavement gray level image I, judging whether the two-dimensional pavement gray level image contains abnormal gray level information or not based on a set threshold, if so, rejecting the image, and if not, adding the image into an image row to obtain a gray level average value Ave of the image row current ;
Step S330, based on the preprocessing coefficient and the gray average value Ave current Acquiring a processed two-dimensional image I of a two-dimensional gray level image I of the road surface new ,
I new =I×C
Wherein C is a preprocessing coefficient, and if I is the first image to be processed, C is C 0 Otherwise, C is updated by the following formula:
Ave Back =αAve Back +(1-α)Ave current
And step S340, returning to step S320 until all image data processing is finished.
3. The method of claim 2, wherein the initial global mean is a global mean of all the road surfacesThe obtaining method comprises the following steps:
acquiring a two-dimensional gray image of a road surface, eliminating the image containing abnormal gray information based on a set threshold value, and extracting N images which can be used for calculating a preprocessing coefficient C 0 Calculating the gray level mean value of the N images as the initial overall mean value
4. The method of claim 1, wherein the three-dimensional data preprocessing algorithm comprises:
s410, processing the relative elevation data of the road surface obtained through data calibration by adopting a pixel left-right adjacent interpolation method to obtain a three-dimensional image elevation value point set after filtering;
step S420, processing the elevation value point set of the original three-dimensional image and the filtered elevation value point set of the three-dimensional image by using a three-dimensional image enhancement function to obtain a three-dimensional image enhancement output result.
5. The method according to claim 4, wherein the data calibration is performed by means of a line structured light principle, and a conversion relation between pixels and space coordinates is established between the two-dimensional gray scale image and the three-dimensional elevation image, so as to generate a calibration parameter file for obtaining relative elevation data of the road surface.
6. The multi-dimensional road surface damage data processing method according to claim 4, wherein the interpolation method for left and right adjacent pixels is,
wherein, X = { X 1 ,X 2 ,…,X w Is the per-row cross-sectional pixel matrix, X w =0 (W =1,2, \8230;, W) is a set of outliers, X w Not equal to 0 (W =1,2, \8230;, W) is the normal set of points, X '= { X' 1 ,X′ 2 ,…,X′ w And W is the number of elevation points of the cross section.
7. The method of claim 4, wherein the three-dimensional image enhancement function is calculated by the formula:
A′(x,y)=A+MFF k ∣A∣+MFF k ∣A(x-I,y-j)∣
wherein, A represents a matrix of an original three-dimensional image elevation value point set, A 'represents a matrix of a filtered three-dimensional image elevation value point set, MFF is an adaptive mean value filtering technology, K is a filtering coefficient, and (2K + 1) × (2K + 1) is taken as a filtering template, A' min Represents the minimum value of the set A ', A' max Represents the maximum value of the A 'set, A' represents the three-dimensional image enhancement output result, i, j E [ -K, K [ ]]。
8. The multidimensional road damage data processing method according to claim 1, wherein the method for constructing the road damage disease identification model comprises:
step S510, marking the damage of the high-quality two-dimensional and three-dimensional images containing the damage of the pavement;
s520, superposing the marked two-dimensional and three-dimensional images to obtain a two-channel training sample database;
and S530, training a deep learning network model by using a dual-channel training sample database to obtain a road surface damage and disease identification model.
9. The method for processing the multi-dimensional road surface damage data according to claim 8, wherein the road surface damage disease recognition model is superimposed with a conditional random field module for improving the accuracy of the road surface damage disease recognition model.
10. The method of claim 8, wherein the damage condition comprises cracks, pits, ruts, fissures, block repairs, and loosening.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211193198.0A CN115512322A (en) | 2022-09-28 | 2022-09-28 | Multidimensional pavement damage data processing method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211193198.0A CN115512322A (en) | 2022-09-28 | 2022-09-28 | Multidimensional pavement damage data processing method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115512322A true CN115512322A (en) | 2022-12-23 |
Family
ID=84508656
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211193198.0A Pending CN115512322A (en) | 2022-09-28 | 2022-09-28 | Multidimensional pavement damage data processing method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115512322A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116051542A (en) * | 2023-03-06 | 2023-05-02 | 深圳市深视智能科技有限公司 | Defect detection method and defect detection device |
-
2022
- 2022-09-28 CN CN202211193198.0A patent/CN115512322A/en active Pending
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116051542A (en) * | 2023-03-06 | 2023-05-02 | 深圳市深视智能科技有限公司 | Defect detection method and defect detection device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110569704A (en) | Multi-strategy self-adaptive lane line detection method based on stereoscopic vision | |
CN111179152B (en) | Road identification recognition method and device, medium and terminal | |
CN110031829B (en) | Target accurate distance measurement method based on monocular vision | |
CN110516633B (en) | Lane line detection method and system based on deep learning | |
CN109447939B (en) | Weld width prediction method before molten pool forming based on mixed spectrum vision | |
CN115546768B (en) | Pavement marking identification method and system based on multi-scale mechanism and attention mechanism | |
CN115512322A (en) | Multidimensional pavement damage data processing method | |
CN113313031B (en) | Deep learning-based lane line detection and vehicle transverse positioning method | |
Salari et al. | Automated pavement distress inspection based on 2D and 3D information | |
CN111178193A (en) | Lane line detection method, lane line detection device and computer-readable storage medium | |
CN112819066A (en) | Res-UNet single tree species classification technology | |
CN112184725B (en) | Method for extracting center of structured light bar of asphalt pavement image | |
CN116883439A (en) | Method and device for detecting slab contour with water stain on surface | |
CN115128628A (en) | Road grid map construction method based on laser SLAM and monocular vision | |
CN113593035A (en) | Motion control decision generation method and device, electronic equipment and storage medium | |
CN115170479A (en) | Automatic extraction method for asphalt pavement repairing diseases | |
CN115082377A (en) | Building surface crack geometric parameter measuring method and system based on unmanned aerial vehicle | |
CN114842430A (en) | Vehicle information identification method and system for road monitoring | |
CN112801021B (en) | Method and system for detecting lane line based on multi-level semantic information | |
CN111402593B (en) | Video traffic parameter acquisition method based on polynomial fitting | |
CN111696147B (en) | Depth estimation method based on improved YOLOv3 model | |
CN116895060A (en) | Three-dimensional lane marking method and device based on scene reconstruction | |
Guan et al. | Multi-scale asphalt pavement deformation detection and measurement based on machine learning of full field-of-view digital surface data | |
CN111336991A (en) | Tunnel ellipticity analysis method, device and system based on laser point cloud | |
CN115035193A (en) | Bulk grain random sampling method based on binocular vision and image segmentation technology |
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 |