CN114320316B - Shield tunneling machine construction early warning method and device - Google Patents
Shield tunneling machine construction early warning method and device Download PDFInfo
- Publication number
- CN114320316B CN114320316B CN202210047086.8A CN202210047086A CN114320316B CN 114320316 B CN114320316 B CN 114320316B CN 202210047086 A CN202210047086 A CN 202210047086A CN 114320316 B CN114320316 B CN 114320316B
- Authority
- CN
- China
- Prior art keywords
- network
- rock
- image
- surrounding rock
- slag
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 27
- 238000010276 construction Methods 0.000 title claims abstract description 23
- 230000005641 tunneling Effects 0.000 title claims abstract description 18
- 239000011435 rock Substances 0.000 claims abstract description 98
- 239000002893 slag Substances 0.000 claims abstract description 36
- 238000012544 monitoring process Methods 0.000 claims abstract description 18
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 17
- 238000013461 design Methods 0.000 claims abstract description 6
- 238000011176 pooling Methods 0.000 claims description 24
- 238000004364 calculation method Methods 0.000 claims description 19
- 230000015654 memory Effects 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 7
- 230000006870 function Effects 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 5
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 238000003709 image segmentation Methods 0.000 claims description 4
- 238000013139 quantization Methods 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 3
- 238000007635 classification algorithm Methods 0.000 claims description 3
- 230000000903 blocking effect Effects 0.000 claims description 2
- 238000007781 pre-processing Methods 0.000 claims description 2
- 230000011218 segmentation Effects 0.000 claims description 2
- 238000013526 transfer learning Methods 0.000 claims description 2
- 230000000875 corresponding effect Effects 0.000 claims 4
- 230000000007 visual effect Effects 0.000 claims 2
- 230000001276 controlling effect Effects 0.000 claims 1
- 230000002596 correlated effect Effects 0.000 claims 1
- 238000012545 processing Methods 0.000 claims 1
- 238000005096 rolling process Methods 0.000 claims 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims 1
- 230000001133 acceleration Effects 0.000 abstract description 6
- 238000001514 detection method Methods 0.000 abstract description 3
- 230000002159 abnormal effect Effects 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000005299 abrasion Methods 0.000 description 2
- 238000009412 basement excavation Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 239000012634 fragment Substances 0.000 description 2
- 229910000831 Steel Inorganic materials 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000003673 groundwater Substances 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000003014 reinforcing effect Effects 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to a shield tunneling machine construction early warning method and device, belongs to the technical field of tunneling, and provides a full-face rock tunneling machine (tunnel boring machine, TBM) which has the advantages of high construction speed, high safety, good economy and the like, and is widely applied to important projects such as expressways, railway transportation, urban subways and the like. Due to the severe working environment, a real-time construction early warning method and a real-time construction early warning device are needed. Aiming at the two fields of cutterhead detection and surrounding rock classification, the invention designs a method based on rock slag classification and surrounding rock integrity monitoring, and realizes tunnel construction monitoring and early warning. Both methods are based on convolutional neural networks, so the invention provides an OpenCL-based convolutional neural network acceleration architecture, and a deployment acceleration device for an early warning method is realized.
Description
Technical Field
The invention belongs to the technical field of tunnel boring, and particularly relates to a shield tunneling machine construction early warning method and device.
Background
The full face rock tunneling machine (tunnel boring machine, TBM) has the advantages of high construction speed, high safety, good economy and the like, and is widely applied to important projects such as expressways, railway transportation, urban subways and the like. The cutter head hob is extremely easy to be worn and damaged due to severe working environment, if the damage of the cutter head can not be found and processed in time in the construction process, the tunneling efficiency can be reduced, the cutter head can be abnormally worn to influence the engineering progress and the construction quality, the abrasion reasons of the shield cutter are researched, the abrasion conditions of the cutter head hob are monitored in real time, and reasonable selection, use, maintenance and replacement of the cutter head are particularly necessary. According to the rock breaking mechanism of the TBM cutterhead hob, when the cutterhead is worn or damaged, the size of the rock fragments generated by tunneling can be increased. Therefore, the state of the cutterhead can be indirectly monitored by monitoring the condition of the rock slag fragments on the conveyor belt, so as to guide on-site constructors to check and replace cutters in time. The camera collects images above the conveyor belt, avoids the severe production environment, and has the advantages of simple equipment, low loss, low cost, long-time monitoring and the like.
Meanwhile, the classification of the tunnel surrounding rock grade plays a vital role in tunnel construction, and is an important factor for judging the stability after tunnel excavation and selecting corresponding supporting measures. The traditional surrounding rock classification is comprehensively determined according to the conditions of rock strength, rock integrity, crack filling, groundwater, ground stress and the like, and the surrounding rock classification in the early investigation is rough due to the fact that the workload is large and the surrounding rock classification is difficult to obtain. The tunnel construction environment and the process are complicated, and the TBM in the construction can influence the shooting sight, simultaneously, can adopt certain supporting measure to ensure the stability of country rock after the country rock excavation, for example stock, reinforcing bar net, steel bow member etc. can interfere the discernment of image. The surrounding rock integrity detection network based on the full convolution neural network can effectively avoid the influence of interferents, and meanwhile, the image segmentation network can segment cracks from the background, so that the current surrounding rock integrity is displayed more intuitively, and constructors are guided to select proper support design and construction according to the types of the surrounding rocks
Based on the background, the invention provides a shield tunneling machine construction early warning method and device based on a convolutional neural network. The method is characterized in that rock slag classification and surrounding rock integrity monitoring are realized through a convolutional neural network, and image acquisition and network calculation are realized through the proposed equipment.
Disclosure of Invention
The invention provides a shield tunneling machine early warning method and device based on computer vision, the system comprises a plurality of image sensors for collecting image information, heterogeneous computing devices and a display for displaying the results. In order to achieve the above purpose, the invention adopts the following technical scheme:
Step 1: and acquiring rock slag images and surrounding rock images in the tunneling process by image acquisition devices distributed at different positions of the shield tunneling machine.
Step 2: transmitting the images to a heterogeneous computing terminal, preprocessing the images, inputting the preprocessed images to different convolutional neural networks, and respectively performing network computation on the input rock slag images and surrounding rock images.
Step 3: and for the feature extraction, a convolutional layer is adopted to carry out feature extraction branches, for a rock slag classification network, the convolutional layer is connected with a full-connection layer to classify, and for a surrounding rock integrity detection network, the convolutional layer is connected with a decoding layer to classify and abnormally divide.
Step 4: feature classification, namely classifying rock slag images into normal and abnormal two types through a Softmax function, and dividing surrounding rock images into complete and incomplete areas through a decoder.
Step 5: and outputting the classification result to a display device for display, so as to help constructors to determine the current construction condition.
Drawings
FIG. 1 is a flow chart of a method for classifying rock and slag based on convolutional neural network
FIG. 2 is a schematic illustration of a rock slag dataset
FIG. 3 is a diagram of a method for monitoring the integrity of a surrounding rock based on a full convolutional network
FIG. 4 is a schematic diagram of a network calculation result for detecting the integrity of surrounding rock
Fig. 5 convolutional neural network heterogeneous computing architecture based on cpu+fpga
FIG. 6 overall architecture of the system
Detailed Description
The invention provides a shield tunneling machine early warning method and device based on computer vision, and in order to make the technical scheme and effect of the shield tunneling machine early warning method and device clearer and more clear, the detailed description of the specific embodiments of the invention is given below with reference to the accompanying drawings.
As shown in FIG. 1, the invention relates to a rock slag classification algorithm flow chart based on a convolutional neural network. The rock slag classification network consists of three network layers, namely a convolution layer, a pooling layer and a full connection layer. The convolutional layer is used for extracting features and is the most main calculation layer of the convolutional neural network. High-level features in the image can be extracted. For each convolution calculation, each output characteristic value is obtained by adding offset after corresponding convolution kernel and input characteristic value multiply-accumulate operation. The operation between different eigenvalues has no data dependency, thus meaning that convolution calculations can be performed in parallel.
The pooling layer is usually located after the convolution layer, and plays roles of reducing the calculation amount and controlling the overfitting, and generally, pooling downsampling is performed after each feature map to reduce the dimension of the feature map, and generally, two pooling methods exist: average pooling and maximum pooling, maximum pooling is used in the algorithms herein, as maximum pooling can preserve more information about edges, textures, etc. of images, and extracted features can better implement classification functions.
The full-connection layer plays a role of a classifier in the whole network, and as the result obtained by the network convolution operation is the characteristic information of the image and cannot be screened and processed manually, the full-connection layer is required to integrate the characteristic information, the characteristic extracted by the convolution layer is classified by the softMax classifier, and the relation between the characteristic information and the sample marking space is established.
The network training process uses the rock slag images acquired in the production process, as shown in fig. 2, and is divided into three types, wherein two types are normal types, one type is abnormal type, the normal type represents the normal work of the hob, and the abnormal type indicates the abnormal working state of the hob.
In the actual production process, the rock residues may be randomly distributed in the direction, and the camera may be affected by the environment to collect images containing noise, so that aiming at the characteristics, a data set is amplified by adopting a random rotation and Gaussian noise adding mode, and the rock residue classification network is retrained by using transfer learning. And learning the weight of the pre-training network into the rock slag classification network, thereby realizing the rock slag classification network with high accuracy.
As shown in fig. 3, the invention relates to a surrounding rock integrity monitoring method based on a full convolution network. The full convolution network is a convolution neural network in deep learning and is specially used for semantic segmentation of images. For a full convolution network, the network input is one image and the output is one image, and the function of the network is to learn pixel-to-pixel mapping to achieve pixel-level recognition. The method comprises the steps of constructing a network structure for identifying the integrity of surrounding rock on the basis of an original network by using a classical full convolution network SegNet, inputting an original image and a label image in a surrounding rock image sample data set into the network, carrying out convolution operation and pooling operation on an input image for a plurality of times through an encoder part of the network, extracting high-level features of the input image, then carrying out convolution and pooling operation through a decoder part, recovering the high-level features to the same size as the input image, simultaneously calculating the probability of the high-level features belonging to an incomplete area for each pixel point, comparing with corresponding stool prospect, realizing one-time forward reasoning operation, then adopting a back propagation algorithm to update and learn network weights, and carrying out continuous iteration to realize the design and training of the surrounding rock network, thereby obtaining the available surrounding rock integrity monitoring network. As shown in fig. 4, after the surrounding rock image is calculated through a network, surrounding rock cracks and the background are represented by pixels with different colors.
For the discrimination of rock integrity, table 1 describes different surrounding rock categories in detail.
Table 1: surrounding rock integrity differentiation
According to the development degree of the cracks and the combination degree of main cracks, the types of surrounding rock can be divided into 5 types, namely complete, more broken, broken and extremely broken. And (3) dividing the cracks and the background of the input surrounding rock image through calculation of the surrounding rock integrity monitoring network, so as to guide constructors to classify the surrounding rock for subsequent support design and construction.
During production, surrounding rock images are acquired through an image sensor, the sensor inputs the images into heterogeneous terminals, and network operation is performed by using the terminals. In terms of algorithm deployment acceleration, most convolutional neural networks currently operate on a CPU/GPU platform, but the number of calculation units in the CPU is small, and the calculation power cannot meet the operation requirement; the GPU has great advantages in the aspect of parallel computing and is mostly used in the training stage of the neural network, but in the reasoning stage, the GPU cannot exert the characteristics of high bandwidth and multiple computing cores, and the GPU faces the challenges of hardware power consumption, size and the like in the practical edge application deployment. The FPGA integrates abundant storage hardware resources, flexible programmable logic resources and high-performance computing resources, can meet the requirements of parallel computing and low power consumption at the same time, has high flexibility and reconfigurability, and provides an effective solution for the edge deployment of the convolutional neural network.
Therefore, the terminal designed by the method is based on the heterogeneous design of CPU and FPGA, and an FPGA convolutional neural network acceleration architecture based on the OpenCL development flow is provided for accelerating the rock slag classification network, as shown in figure 5. According to the independence among all calculation layers of the convolutional neural network, a configurable OpenCL kernel is designed to respectively complete the convolution, pooling and full-connection layer operation of the network. The whole architecture is mainly divided into an on-chip part and an off-chip part, and an off-chip global memory is mainly used for storing characteristic values, model parameters and quantization parameter information, and an on-chip memory is used for storing and transmitting characteristic value data, weight parameters, quantization parameters, calculation results of a current calculation layer and the like.
In the architecture, the cores are connected with each other through a configurable pipeline, and the pipeline is an efficient non-blocking first-in first-out queue data transmission module constructed by using on-chip resources, so that data between the cores are directly transmitted on-chip, the access times to off-chip memories can be reduced, and the data transmission efficiency is improved. The control words are flexibly combined with the kernel connection, so that the complex network operation acceleration can be realized through hardware resources.
When the control word is 0, the convolution or full connection operation is completed, and the connection mode of the pipeline and the kernel is as follows: off-chip memory-on-chip cache-data pipeline-convolution kernel-data pipeline-on-chip cache-off-chip memory.
When the control word is 1, the convolution and pooling operation is completed, and the connection modes of the pipeline and the kernel are as follows: the method comprises the steps of off-chip memory, on-chip cache, data pipeline, volume, core, data pipeline, pooling core, data pipeline, on-chip cache and off-chip memory.
When the control word is 2, the convolution and anti-pooling operation is completed, and the connection modes of the pipeline and the kernel are as follows: off-chip memory-on-chip cache-data pipeline-convolution kernel-anti-pooling kernel-data pipeline-on-chip cache-off-chip memory.
When relevant connection mode data stream is carried out, each kernel is only executed once, and the contention of kernel resources can not occur, so that the smoothness of the whole pipeline is ensured. By combining the control words, the controllable framework calculates a slag classification network or a surrounding rock integrity monitoring network.
Thus, in connection with the method and apparatus, the complete system is shown in FIG. 6, and the system operates as follows:
1. and respectively acquiring a rock slag image and a surrounding rock image through an image sensor above the rock slag conveying belt and an image sensor outside the shield tunneling machine.
2. The image sensor transmits the acquired rock slag image and surrounding rock image to heterogeneous computing equipment, and the equipment CPU end cuts the image so as to be suitable for the input requirement of a network.
3. The rock slag image and the surrounding rock image are respectively input into the network acceleration architecture provided by the invention, and the flow direction of the feature map data is controlled according to the set control words through the scheduling of the CPU end and the intensive calculation of the FPGA end, so that the operation of the rock slag classification network and the surrounding rock integrity monitoring network is realized.
4. After the calculation is completed, the original image and the output result are output from the FPGA end to the CPU end, and are displayed after being processed. The rock slag classification network outputs and displays the current rock slag image and the type thereof; and outputting and displaying the current surrounding rock image and a binary image after image segmentation by the surrounding rock integrity network.
Claims (3)
1. The utility model provides a shield constructs quick-witted construction early warning device which characterized in that includes: the visual sensor, heterogeneous computing equipment and the operation room display are used for completing the identification of the rock slag state and the identification of the integrity of surrounding rock; wherein:
The visual sensor is arranged above the rock slag conveyor belt and in front of the shield machine and is respectively used for acquiring rock slag images on the rock slag conveyor belt and surrounding rock images on the periphery of the shield machine, and the image sensor also has the function of acquiring images with different resolutions for matching with subsequent network calculation;
the heterogeneous computing device is connected with the vision sensor and is used for receiving and processing images acquired by the vision sensor, the heterogeneous computing device comprises a CPU part and an FPGA part, the pre-warning method based on a convolutional neural network is realized through the cooperation of the CPU and the FPGA, specifically, in the OpenCL flow, the CPU is responsible for preprocessing and scheduling data, the FPGA is responsible for intensive computing of the data, the configurable OpenCL kernel is designed according to the independence among computing layers of CNN, the convolution, pooling, anti-pooling and full-connection layer computing of the network are respectively completed, the whole framework is divided into an on-chip part and an off-chip part, the off-chip global memory is used for storing characteristic value, model parameter and quantization parameter information, the on-chip memory is used for storing and transmitting characteristic value data, weight parameter and quantization parameter of the current computing layer, in the framework, the cores are connected with each other through a configurable pipeline, and the pipeline is an efficient non-blocking first-in-out queue (First Input First Output, data transmission module constructed by using on-chip resources, so that the data between the cores can be directly transmitted on-chip, the number of times of the on-chip memory can be increased by controlling the memory to access the hardware through the hardware combination;
the operation room display is used for displaying the collected image information and the corresponding calculation result;
for a rock slag image, a rock slag classification algorithm based on a convolutional neural network is designed and deployed; for surrounding rock images, a surrounding rock complete monitoring network based on image segmentation is designed and deployed, network calculation is realized through a heterogeneous calculation architecture, a rock and slag classification network is formed by modifying and training a VGG16 network, a pre-training network is subjected to transfer learning through rock and slag images acquired in engineering scenes, a rock and slag classification network is obtained, a rock and slag classification result is positively correlated with the state of a cutter head hob of a current shield machine, and the aim of indirectly monitoring the state of the hob can be realized through monitoring the state of the rock and slag; the surrounding rock integrity monitoring network realizes semantic segmentation of an input image by an image segmentation network SegNet network, classifies pixels of an input surrounding rock picture into an incomplete surrounding rock area and a background area, and obtains specific position and accurate geometric information of the incomplete area;
the construction method of the surrounding rock integrity monitoring network comprises the following steps: on the basis of a full convolution network SegNet, a network structure for identifying the integrity of surrounding rock is constructed, an original image and a label image in a surrounding rock image sample dataset are input into the network, convolution operation and pooling operation are carried out on the input image for a plurality of times through an encoder part of the network, high-level characteristics of the input image are extracted, then rolling and anti-pooling operation are carried out through a decoder part, the high-level characteristics are restored to the same size as the input image, meanwhile, the probability that each pixel point belongs to an incomplete area is calculated, the probability is compared with corresponding stool prospect, forward reasoning operation is realized once, then the network weight is updated and learned through a back propagation algorithm, and the design and training of the surrounding rock network are realized through continuous iteration, so that the available surrounding rock integrity monitoring network is obtained.
2. The shield tunneling machine construction early warning device according to claim 1, wherein: the rock slag classification algorithm based on the convolutional neural network comprises a convolutional layer, a pooling layer and a full-connection layer, the rock slag image features are extracted through the convolutional layer, the network calculation amount is reduced through the pooling layer, the overfitting of the network is controlled, the function of a classifier is realized through the full-connection layer, and the relation between feature information and a mark space is established.
3. The shield tunneling machine construction pre-warning device according to claim 2, wherein: the surrounding rock complete monitoring network comprises an encoder and a decoder, wherein the encoder comprises a convolution layer and a pooling layer, in the encoder, an input surrounding rock image is subjected to convolution operation and pooling operation for a plurality of times to generate a high-level characteristic image, then the high-level characteristic image is subjected to the decoder, abstract high-level characteristic is restored to the same size as the input image through convolution and pooling operation, and meanwhile, a predicted value is set for each pixel point in an output image through network operation and is used for indicating whether the surrounding rock at the corresponding position of the pixel point has cracks or water leakage or not, and a strategy is selected according to stability judgment and corresponding supporting measures in the construction process of the segmented image guidance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210047086.8A CN114320316B (en) | 2022-01-14 | 2022-01-14 | Shield tunneling machine construction early warning method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210047086.8A CN114320316B (en) | 2022-01-14 | 2022-01-14 | Shield tunneling machine construction early warning method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114320316A CN114320316A (en) | 2022-04-12 |
| CN114320316B true CN114320316B (en) | 2024-11-08 |
Family
ID=81027616
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210047086.8A Active CN114320316B (en) | 2022-01-14 | 2022-01-14 | Shield tunneling machine construction early warning method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114320316B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114926101B (en) * | 2022-07-18 | 2022-10-28 | 石家庄铁道大学 | Construction speed and safety-oriented TBM (Tunnel boring machine) adaptive surrounding rock grading method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109886329A (en) * | 2019-02-18 | 2019-06-14 | 中国铁建重工集团有限公司 | Rock crusher level detection method, detection system and heading equipment |
| CN110245695A (en) * | 2019-05-30 | 2019-09-17 | 华中科技大学 | A kind of TBM rock slag order of magnitude recognition methods based on convolutional neural networks |
| CN110991632A (en) * | 2019-11-29 | 2020-04-10 | 电子科技大学 | Method for designing heterogeneous neural network computing accelerator based on FPGA |
| CN112127896A (en) * | 2020-09-18 | 2020-12-25 | 武汉大学 | Automatic acquisition and analysis system and method for TBM excavation rock slag information |
Family Cites Families (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3731950B2 (en) * | 1996-09-20 | 2006-01-05 | 株式会社小松製作所 | Eccentric start method of child shield machine in parent and child shield machine |
| CN101105126B (en) * | 2007-08-03 | 2010-09-15 | 重庆邮电大学 | Logging-while-drilling orientation measurement error compensation method based on micro-quartz angular rate sensor |
| JP2015004194A (en) * | 2013-06-20 | 2015-01-08 | 株式会社マシノ | Geological survey method of mountain tunnel working face front natural ground |
| DE112015002700B4 (en) * | 2015-03-11 | 2021-12-23 | Shandong University | TBM 3D advance exploration system with integration of rock quarry epicenter and active epicenter |
| CN107577862B (en) * | 2017-08-30 | 2019-12-03 | 中铁工程装备集团有限公司 | A kind of TBM is in pick rock mass state real-time perception system and method |
| CN107632523B (en) * | 2017-09-30 | 2019-07-23 | 中铁工程装备集团有限公司 | A kind of hard rock TBM digging control parameter intelligent decision-making technique and system |
| CN108643930B (en) * | 2018-05-08 | 2020-02-07 | 中铁工程装备集团有限公司 | Real-time early warning method for TBM tunnel construction |
| CN110109895B (en) * | 2019-03-29 | 2021-05-28 | 山东大学 | Surrounding rock grading combined prediction method suitable for TBM tunneling tunnel and application |
| CN110043267B (en) * | 2019-04-04 | 2020-07-31 | 山东大学 | TBM (Tunnel boring machine) carrying type advanced geological prediction system and method based on lithology and unfavorable geological precursor characteristic identification |
| CN110331983A (en) * | 2019-07-11 | 2019-10-15 | 中国电建集团成都勘测设计研究院有限公司 | Double-shielded TBM tunneling construction method |
| CN110359923A (en) * | 2019-07-26 | 2019-10-22 | 盾构及掘进技术国家重点实验室 | A kind of bottom open type TBM cleaning dregs system |
| CN110516730B (en) * | 2019-08-20 | 2022-09-23 | 中铁工程装备集团有限公司 | Surrounding rock quality online grading method based on PSO-SVM algorithm and image recognition |
| CN110751676A (en) * | 2019-10-21 | 2020-02-04 | 中国科学院空间应用工程与技术中心 | Heterogeneous computing system and method based on target detection and readable storage medium |
| CN110686616A (en) * | 2019-10-23 | 2020-01-14 | 中南大学 | TBM tunnel surrounding rock morphology acquisition method and device |
| CN110593929B (en) * | 2019-10-28 | 2024-06-18 | 中冶沈勘秦皇岛工程设计研究总院有限公司 | Secondary multi-section anchoring anchor cable and application method thereof |
| CN111879610B (en) * | 2020-07-10 | 2024-04-02 | 武汉大学 | Real-time measurement system and method for rock slag mechanical parameters in tunneling process |
| CN112647965B (en) * | 2020-12-09 | 2022-03-04 | 山东大学 | Method and system suitable for real-time card-blocking prediction of TBM tunneling tunnel |
| CN112990227B (en) * | 2021-02-08 | 2022-12-27 | 中国铁建重工集团股份有限公司 | Face geology detection method |
| CN113310949B (en) * | 2021-05-24 | 2023-01-13 | 山东大学 | Hyperspectral imaging-based TBM tunnel-mounted rock slag mineral identification system and method |
| CN113338971B (en) * | 2021-07-21 | 2023-04-07 | 韩臻 | Mining TBM rock slag slurry water-carrying closed-circuit circulating system and process |
| CN113780476A (en) * | 2021-10-09 | 2021-12-10 | 中国铁建重工集团股份有限公司 | Rock slag characteristic detection model training method, device, equipment and medium |
-
2022
- 2022-01-14 CN CN202210047086.8A patent/CN114320316B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109886329A (en) * | 2019-02-18 | 2019-06-14 | 中国铁建重工集团有限公司 | Rock crusher level detection method, detection system and heading equipment |
| CN110245695A (en) * | 2019-05-30 | 2019-09-17 | 华中科技大学 | A kind of TBM rock slag order of magnitude recognition methods based on convolutional neural networks |
| CN110991632A (en) * | 2019-11-29 | 2020-04-10 | 电子科技大学 | Method for designing heterogeneous neural network computing accelerator based on FPGA |
| CN112127896A (en) * | 2020-09-18 | 2020-12-25 | 武汉大学 | Automatic acquisition and analysis system and method for TBM excavation rock slag information |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114320316A (en) | 2022-04-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111899288B (en) | Tunnel leakage water area detection and identification method based on infrared and visible light image fusion | |
| CN106290388A (en) | A kind of insulator breakdown automatic testing method | |
| CN108875596A (en) | A kind of railway scene image, semantic dividing method based on DSSNN neural network | |
| CN117522149B (en) | Tunnel security risk identification method and device and security management platform | |
| CN114320316B (en) | Shield tunneling machine construction early warning method and device | |
| CN103927553B (en) | Coal and rock recognition method based on multi-scale micro-lamination and contrast ratio joint distribution | |
| Tan et al. | A lightweight metro tunnel water leakage identification algorithm via machine vision | |
| CN111626170A (en) | Image identification method for railway slope rockfall invasion limit detection | |
| CN113128476A (en) | Low-power consumption real-time helmet detection method based on computer vision target detection | |
| CN115830533A (en) | Helmet wearing detection method based on K-means clustering improved YOLOv5 algorithm | |
| CN114724063B (en) | Road traffic incident detection method based on deep learning | |
| CN110751150A (en) | FPGA-based binary neural network license plate recognition method and system | |
| CN116682010B (en) | Surrounding rock classification real-time prediction method based on TBM rock slag image | |
| CN117133052A (en) | Method and device for detecting wearing of safety helmet, electronic equipment and storage medium | |
| CN115690643A (en) | Improved YOLOX-based safety helmet detection method, system, equipment and medium | |
| CN114212483B (en) | CV algorithm-based scraper conveyor component fault identification method and device | |
| CN117423157A (en) | Mine abnormal video action understanding method combining migration learning and regional invasion | |
| CN116912628A (en) | Method and device for training defect detection model and detecting defects | |
| CN113362330B (en) | Pantograph cavel real-time detection method, device, computer equipment and storage medium | |
| Hu et al. | Improved YOLOv8 Algorithm with C2f-DCNv3 and Shuffle Attention for Detection of Coal Shearer Drum Teeth | |
| CN114120109A (en) | Belt longitudinal tearing detection method based on neural network | |
| CN116206155A (en) | Waste steel classification and identification method based on YOLOv5 network | |
| Min et al. | Rail Surface Defect Detection Based on Improved UPerNet and Connected Component Analysis. | |
| CN115641469A (en) | Gyolov 5-based X-ray dangerous article detection and identification method | |
| Deena et al. | Prohibited Object Detection using YOLOv7 on X-Ray Images of Airport Luggages (OPIXray) |
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 |