CN113627499A - Smoke intensity grade estimation method and device based on inspection station diesel vehicle tail gas image - Google Patents
Smoke intensity grade estimation method and device based on inspection station diesel vehicle tail gas image Download PDFInfo
- Publication number
- CN113627499A CN113627499A CN202110859026.1A CN202110859026A CN113627499A CN 113627499 A CN113627499 A CN 113627499A CN 202110859026 A CN202110859026 A CN 202110859026A CN 113627499 A CN113627499 A CN 113627499A
- Authority
- CN
- China
- Prior art keywords
- image
- frame
- network model
- convolutional
- residual
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 239000000779 smoke Substances 0.000 title claims abstract description 33
- 238000007689 inspection Methods 0.000 title claims abstract description 19
- 238000012549 training Methods 0.000 claims abstract description 24
- 238000012795 verification Methods 0.000 claims abstract description 14
- 238000007781 pre-processing Methods 0.000 claims abstract description 13
- 238000005070 sampling Methods 0.000 claims abstract description 10
- 230000006870 function Effects 0.000 claims description 22
- 238000004590 computer program Methods 0.000 claims description 13
- 230000004927 fusion Effects 0.000 claims description 10
- 101100194362 Schizosaccharomyces pombe (strain 972 / ATCC 24843) res1 gene Proteins 0.000 claims description 9
- 101100194363 Schizosaccharomyces pombe (strain 972 / ATCC 24843) res2 gene Proteins 0.000 claims description 9
- 238000011176 pooling Methods 0.000 claims description 9
- 230000015654 memory Effects 0.000 claims description 8
- 238000012544 monitoring process Methods 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 7
- 208000037170 Delayed Emergence from Anesthesia Diseases 0.000 claims description 6
- 230000003044 adaptive effect Effects 0.000 claims description 6
- 229910052739 hydrogen Inorganic materials 0.000 claims description 6
- 238000011156 evaluation Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000010200 validation analysis Methods 0.000 claims description 2
- 230000007613 environmental effect Effects 0.000 abstract description 2
- 239000000284 extract Substances 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 9
- 238000001514 detection method Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 239000003344 environmental pollutant Substances 0.000 description 2
- 238000000691 measurement method Methods 0.000 description 2
- 231100000719 pollutant Toxicity 0.000 description 2
- BUCXEFZXWKUCCY-UHFFFAOYSA-N 4-methyl-3-(2-phenylethyl)-1,2,4-oxadiazol-5-one Chemical compound O1C(=O)N(C)C(CCC=2C=CC=CC=2)=N1 BUCXEFZXWKUCCY-UHFFFAOYSA-N 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000002229 photoelectron microspectroscopy Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06T7/0004—Industrial image inspection
-
- 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
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20084—Artificial neural networks [ANN]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Abstract
The invention discloses a smoke intensity grade estimation method and equipment based on a diesel vehicle tail gas image of an inspection station, which comprises the following steps of: s1: dividing videos collected by an inspection station into subblocks, sampling each subblock, and defining a residual frame according to an obtained RGB frame; s2: preprocessing the RGB image and the residual frame image; s3: constructing a space-time cross network model, training the model by utilizing a training data set and a verification data set, finely adjusting and optimizing the whole network based on a preset loss function and an optimizer until the network is converged, and storing the trained network model parameters; s4: and (4) carrying out grade estimation on the image to be detected by using the trained network model. The method firstly extracts the characteristics of the picture of the tail gas emitted by the diesel vehicle during the environmental annual inspection, effectively removes redundant information, calculates the amount directly related to the smoke concentration in the tail gas emitted by the vehicle through an algorithm, and has higher pertinence and higher feasibility.
Description
Technical Field
The invention relates to the technical field of environmental inspection, in particular to a smoke intensity level estimation method based on a diesel vehicle tail gas image of an inspection station.
Background
Along with the healthy and rapid development of economic development in China, the reserve of motor vehicles in China is increased. The diesel vehicle is one of the important components of motor vehicles in China, and the exhaust emission of the diesel vehicle causes serious atmospheric pollution. Therefore, in order to monitor and control the exhaust emission of various types of diesel vehicles more specifically, it is necessary to effectively classify the smoke emission level of the diesel vehicles.
At present, the motor vehicle exhaust monitoring in China is mainly divided into that arranged on a vehicle body: the system comprises a vehicle-mounted tail gas detection device (PEMS), a vehicle-mounted automatic diagnosis system (OBD) and a remote sensing monitoring instrument arranged at a fixed point; data monitored by the two methods are often subdivided into specific values of various pollutants, and integration of data of each part is needed to realize pollutant emission level estimation, which often brings heavy workload.
The grade estimation of the diesel vehicle exhaust emission is better realized through the exhaust, and the existing research has recently shown that the method for identifying the diesel vehicle exhaust emission by using a computer vision technology is a low-cost and convenient method. The space-time cross network provided by the invention constructs a double pyramid characteristic structure, and promotes mutual guidance of time domain characteristics and space domain characteristics to help the whole network learning, so that the smoke intensity level of the tail gas emission of the diesel vehicle can be accurately estimated, and the relevant departments can conveniently supervise and make policies.
Disclosure of Invention
The invention provides a smoke intensity grade estimation method and device based on a diesel vehicle tail gas image of an inspection station, which can solve the technical problem.
In order to achieve the purpose, the invention adopts the following technical scheme:
a smoke intensity grade estimation method based on a diesel vehicle tail gas image of an inspection station comprises the following steps:
s1: dividing videos collected by an inspection station into subblocks, sampling each subblock, and defining a residual frame according to an obtained RGB frame;
s2: preprocessing the RGB image and the residual frame image;
s3: constructing a space-time cross network model, training the model by utilizing a training data set and a verification data set, finely adjusting and optimizing the whole network based on a preset loss function and an optimizer until the network is converged, and storing the trained network model parameters;
s4: and (4) carrying out grade estimation on the image to be detected by using the trained network model.
Further, in S1, the operations of dividing the video collected by the checkpoint into subblocks, performing frame sampling on each subblock, and defining a residual frame according to the obtained RGB frame are as follows:
s11: dividing the monitoring video into N subblocks, extracting an RGB Frame from each subblock, and recording each Frame as a Framei,i<＝N,
C, T, H, W are image channel number, time length, width and height;
s12: according to the RGB frame extracted at S11, the residual frame ResFrame is divided according to the following formula:
where alpha is a constant coefficient, beta is a set pixel value size to prevent residual frame size overflow,
c, T, H, W, which is the number of image channels, time length, width and height;
s13: the resulting image was displayed as 7: 2: the scale of 1 is randomly divided into a training set, a validation set, and a test set.
Further, the details of preprocessing the RGB image and the residual frame image in S2
The operation is as follows:
s21: simple scaling, random cropping, horizontal flipping, perspective transformation, region erasure and color dithering are performed on the RGB image and the residual frame image.
Further, a space-time cross network model is constructed in S3, the model is trained using a training data set and a verification data set, the whole network is fine-tuned and optimized based on a preset loss function and an optimizer until the network converges, and the specific operations of storing the trained network model parameters are as follows:
s31: the time domain and the space domain of the whole network model are consistent, and the main body adopts SE-ResNeXt-50 to extract space-time characteristics; the network structure comprises 4 convolutional residual blocks { res1, res2, res3, res4}, and the outputs of the blocks are {4,8,16,32} relative to stride of the input frame; the outside of the middle-included number is the repetition number of each residual block, on the basis of ResNeXt-50, the output identity mapping of the last layer of res1, res2 and res3 is selected at the residual stage of a time domain and a space domain to construct a double pyramid feature network, so that the time domain and the space domain are mutually guided, and feature maps output by res4 are fused and used for final classification head prediction; the feature fusion is the addition of corresponding elements of the spatial domain feature and the time domain feature, and the specific operation description is as follows:
wherein i is more than or equal to 1 and less than or equal to H, j is more than or equal to 1 and less than or equal to W, C is more than or equal to 1 and less than or equal to C and SF,
s32: the whole network architecture specifically comprises a 7 × 7 convolutional layer (conv1), a 3 × 3 pooling layer (pool1), three convolutional blocks (res1) comprising three convolutional layers, four convolutional blocks (res2) comprising three convolutional layers, 6 convolutional blocks (res3) comprising three convolutional layers, and three convolutional blocks (res4) comprising three convolutional layers, wherein the convolutional kernel size of the three convolutional layers contained in each convolutional block is 1 × 1, 3 × 3 and 1 × 1, and after module stacking, a feature fusion layer, an adaptive average pooling layer and a full connection layer are added, wherein the convolutional kernel size of the feature fusion layer and the adaptive average pooling layer is 1 × 1;
s33: the time domain flow part of the whole network model takes a residual error frame as input, the space domain part takes an RGB frame as input, a training set and a verification set are input into the network, Relu is selected as an activation function, a loss function is a cross entropy loss function, the batch size is set to be 3, the weight attenuation is set to be 0.0005, the momentum is set to be 0.9, the learning rate is 0.001, the weight of the model is optimized by using random gradient descent, and the iteration is stopped after the network converges, and the whole network model is stored.
Further, in S4, the trained network model is used, and the specific operation of performing the grade estimation on the image to be detected is as follows:
s41: and inputting the test set into a trained smoke degree estimation model in S3 to obtain a corresponding smoke degree grade.
On the other hand, the invention also discloses a smoke degree grade estimation system based on the inspection station diesel vehicle tail gas image, which comprises the following units:
the sub-block dividing unit is used for dividing the video collected by the inspection station into sub-blocks, sampling each sub-block and dividing a residual frame according to the obtained RGB frame;
the preprocessing unit is used for preprocessing the RGB image and the residual frame image;
the model construction unit is used for constructing a space-time cross network model, training the model by utilizing a training data set and a verification data set, finely adjusting and optimizing the whole network based on a preset loss function and an optimizer until the network is converged, and storing the trained network model parameters;
and the grade evaluation unit is used for carrying out grade estimation on the image to be detected by utilizing the trained network model.
In a second aspect, the invention also discloses a computer-readable storage medium, in which a computer program is stored, which, when executed by a processor, causes the processor to perform the steps of the method as described above.
In a third aspect, the present invention also discloses a computer device comprising a memory and a processor, the memory storing a computer program, which, when executed by the processor, causes the processor to perform the steps of the method as described above.
According to the technical scheme, the invention provides the smoke intensity level estimation method based on the diesel vehicle tail gas image of the detection station aiming at the problem that the environment-friendly detection of the vehicle detection station lacks an efficient supervision means.
In conclusion, the smoke degree grading of the diesel vehicle is estimated based on the black smoke discharged by the diesel vehicle at the monitoring station, and the method can highlight the fine motion characteristics of the black smoke by means of mutual guidance of a time domain and a space domain, so that the interference of trees and buildings can be effectively reduced. Compared with the traditional instrument measurement method, the method does not need expensive equipment, can greatly reduce the cost, has good real-time performance and high implementability, and is beneficial to popularization.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a diagram of a smoke level estimation network according to the present invention;
FIG. 3 is a double pyramid signature network architecture according to the present invention;
fig. 4, 4a, 4b, and 4c are graphs showing the recognition results of the model of the present invention for 3 diesel vehicle samples.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention.
As shown in fig. 1, the embodiment of the invention provides a smoke intensity level estimation method based on a diesel vehicle tail gas image of a detection station, aiming at the problem that environment-friendly detection of a vehicle detection station lacks an efficient supervision means.
The method comprises the following specific steps:
a smoke intensity grade estimation method based on a diesel vehicle tail gas image of an inspection station comprises the following steps:
s1: dividing videos collected by an inspection station into subblocks, sampling each subblock, and defining a residual frame according to an obtained RGB frame;
s2: preprocessing the RGB image and the residual frame image;
s3: constructing a space-time cross network model, training the model by utilizing a training data set and a verification data set, finely adjusting and optimizing the whole network based on a preset loss function and an optimizer until the network is converged, and storing the trained network model parameters;
s4: and (4) carrying out grade estimation on the image to be detected by using the trained network model.
The following are described separately:
in S1, the operations of dividing the video collected by the checkpoint into subblocks, performing frame sampling on each subblock, and defining a residual frame according to the obtained RGB frame are as follows:
s11: dividing a monitoring video into N subblocks (N is 8 in the invention), extracting an RGB Frame in each subblock, and recording each Frame as a Framei,i<＝N,
C, T, H, W are the number of image channels, time length, width and height (in the present invention C is 3);
s12: according to the RGB frame extracted at S11, the residual frame ResFrame is divided according to the following formula:
where α is a constant coefficient, β is a set pixel value size to prevent the residual frame size from overflowing, (in the present invention β is 255),
c, T, H, W are the number of image channels, time length, width, and height (C is 3 in this embodiment).
S13: the resulting image was displayed as 7: 2: 1, randomly dividing the proportion into a training set, a verification set and a test set;
the specific operation of preprocessing the RGB image and the residual frame image in S2 is as follows:
s21: simple scaling, random cropping, horizontal flipping, perspective transformation, region erasure and color dithering are performed on the RGB image and the residual frame image.
In S3, a spatio-temporal crossing network model is constructed, a training dataset and a verification dataset are used to train the model, the whole network is fine-tuned and optimized based on a preset loss function and an optimizer until the network converges, and the specific operations of storing the trained network model parameters are as follows:
s31: the time domain and space domain structures of the whole network model are consistent, the main body adopts SE-ResNeXt-50 to extract space-time characteristics, and the basic structure of the network is shown in figure 2. The network contains 4 convolutional residual blocks { res1, res2, res3, res4} whose outputs are {4,8,16,32} relative to the stride of the input frame. The outside of the middle-included number is the repetition number of each residual block, on the basis of ResNeXt-50, the output identity mapping of the last layer of res1, res2 and res3 is selected at the residual stage of a time domain and a space domain to construct a double pyramid feature network, so that the time domain and the space domain are mutually guided, and feature maps output by res4 are fused and used for final classification head prediction; the feature fusion is the addition of corresponding elements of the spatial domain feature and the time domain feature, and the specific operation can be described as:
wherein i is more than or equal to 1 and less than or equal to H, j is more than or equal to 1 and less than or equal to W, C is more than or equal to 1 and less than or equal to C and SF,
s32: the whole network architecture specifically comprises a 7 × 7 convolutional layer (conv1), a 3 × 3 pooling layer (pool1), three convolutional blocks (res1) comprising three convolutional layers, four convolutional blocks (res2) comprising three convolutional layers, 6 convolutional blocks (res3) comprising three convolutional layers, and three convolutional blocks (res4) comprising three convolutional layers, wherein the convolutional kernel size of the three convolutional layers contained in each convolutional block is 1 × 1, 3 × 3, and 1 × 1, and after module stacking, a feature fusion layer, an adaptive average pooling layer, and a full connection layer are added, wherein the convolutional kernel size of the feature fusion layer and the adaptive average pooling layer is 1 × 1, and the whole structure is shown in fig. 3.
S33: the time domain flow part of the whole network model takes a residual error frame as input, the space domain part takes an RGB frame as input, a training set and a verification set are input into the network, Relu is selected as an activation function, a loss function is a cross entropy loss function, the batch size is set to be 3, the weight attenuation is set to be 0.0005, the momentum is set to be 0.9, the learning rate is 0.001, the weight of the model is optimized by using random gradient descent, and the iteration is stopped after the network converges, and the whole network model is stored.
In S4, using the trained network model, the specific operation of performing the grade estimation on the image to be detected is as follows:
s41: inputting the test set into the trained smoke intensity estimation model in S3 to obtain a corresponding smoke intensity grade, wherein the prediction effect is shown in FIG. 4, and the prediction result is consistent with the label.
S42: in order to further illustrate the performance advantage of the space-time cross fusion network provided by the invention, the STCNet of the method is compared with I3D-TSM, I3D-LSTM, I3D-NL and I3D-TC, namely RGB stream is taken as input, and a corresponding module is accessed after an I3D framework. Wherein TSM is a temporal shift module, LSTM is a long-term and short-term memory network module, NL is a non-local module, TC is a time sensing layer, an evaluation index is F Score (F-Score), and specific values of F scores of all models on a test set are as follows:
| Model | F-score |
| I3D-TSM | 81.3％ |
| I3D-LSTM | 81.3％ |
| I3D-NL | 81.7％ |
| I3D-TC | 82.3％ |
| STCNet | 88.5％ |
as can be seen from the data in the table, the F-score value of the STCNet of the method of the embodiment of the invention is 88.5%, which is higher than that of other models, and the method has better performance.
In conclusion, the smoke degree grading of the diesel vehicle is estimated based on the black smoke discharged by the diesel vehicle at the monitoring station, and the method can highlight the fine motion characteristics of the black smoke by means of mutual guidance of a time domain and a space domain, so that the interference of trees and buildings can be effectively reduced. Compared with the traditional instrument measurement method, the method does not need expensive equipment, can greatly reduce the cost, has good real-time performance and high implementability, and is beneficial to popularization.
On the other hand, the invention also discloses a smoke degree grade estimation system based on the inspection station diesel vehicle tail gas image, which comprises the following units:
the sub-block dividing unit is used for dividing the video collected by the inspection station into sub-blocks, sampling each sub-block and dividing a residual frame according to the obtained RGB frame;
the preprocessing unit is used for preprocessing the RGB image and the residual frame image;
the model construction unit is used for constructing a space-time cross network model, training the model by utilizing a training data set and a verification data set, finely adjusting and optimizing the whole network based on a preset loss function and an optimizer until the network is converged, and storing the trained network model parameters;
and the grade evaluation unit is used for carrying out grade estimation on the image to be detected by utilizing the trained network model.
In a second aspect, the invention also discloses a computer-readable storage medium, in which a computer program is stored, which, when executed by a processor, causes the processor to perform the steps of the method as described above.
In a third aspect, the present invention also discloses a computer device comprising a memory and a processor, the memory storing a computer program, which, when executed by the processor, causes the processor to perform the steps of the method as described above.
It is understood that the system provided by the embodiment of the present invention corresponds to the method provided by the embodiment of the present invention, and the explanation, the example and the beneficial effects of the related contents can refer to the corresponding parts in the method.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (8)
1. A smoke intensity grade estimation method based on a diesel vehicle tail gas image of an inspection station is characterized by comprising the following steps:
the method comprises the following steps:
s1: dividing videos collected by an inspection station into subblocks, sampling each subblock, and defining a residual frame according to an obtained RGB frame;
s2: preprocessing the RGB image and the residual frame image;
s3: constructing a space-time cross network model, training the model by utilizing a training data set and a verification data set, finely adjusting and optimizing the whole network based on a preset loss function and an optimizer until the network is converged, and storing the trained network model parameters;
s4: and (4) carrying out grade estimation on the image to be detected by using the trained network model.
2. The checkpoint diesel vehicle exhaust image-based smoke level estimation method as claimed in claim 1, wherein: in S1, the operations of dividing the video collected by the checkpoint into subblocks, performing frame sampling on each subblock, and defining a residual frame according to the obtained RGB frame are as follows:
s11: dividing the monitoring video into N subblocks, extracting an RGB Frame from each subblock, and recording each Frame as a Framei,i<＝N,
C, T, H, W are image channel number, time length, width and height;
s12: according to the RGB frame extracted at S11, the residual frame ResFrame is divided according to the following formula:
where alpha is a constant coefficient, beta is a set pixel value size to prevent residual frame size overflow,
c, T, H, W, which is the number of image channels, time length, width and height;
s13: the resulting image was displayed as 7: 2: the scale of 1 is randomly divided into a training set, a validation set, and a test set.
3. The checkpoint diesel vehicle exhaust image-based smoke level estimation method as claimed in claim 2, wherein: the specific operation of preprocessing the RGB image and the residual frame image in S2 is as follows:
s21: simple scaling, random cropping, horizontal flipping, perspective transformation, region erasure and color dithering are performed on the RGB image and the residual frame image.
4. The checkpoint diesel vehicle exhaust image-based smoke level estimation method as claimed in claim 3, wherein: and (8) constructing a space-time cross network model in the S3, training the model by using a training data set and a verification data set, finely adjusting and optimizing the whole network based on a preset loss function and an optimizer until the network is converged, and storing the parameters of the trained network model as follows:
s31: the time domain and the space domain of the whole network model are consistent, and the main body adopts SE-ResNeXt-50 to extract space-time characteristics; the network structure comprises 4 convolutional residual blocks { res1, res2, res3, res4}, and the outputs of the blocks are {4,8,16,32} relative to stride of the input frame; the outside of the middle-included number is the repetition number of each residual block, on the basis of ResNeXt-50, the output identity mapping of the last layer of res1, res2 and res3 is selected at the residual stage of a time domain and a space domain to construct a double pyramid feature network, so that the time domain and the space domain are mutually guided, and feature maps output by res4 are fused and used for final classification head prediction; the feature fusion is the addition of corresponding elements of the spatial domain feature and the time domain feature, and the specific operation description is as follows:
wherein i is more than or equal to 1 and less than or equal to H, j is more than or equal to 1 and less than or equal to W, C is more than or equal to 1 and less than or equal to C,
s32: the whole network architecture specifically comprises a 7 × 7 convolutional layer (conv1), a 3 × 3 pooling layer (pool1), three convolutional blocks (res1) comprising three convolutional layers, four convolutional blocks (res2) comprising three convolutional layers, 6 convolutional blocks (res3) comprising three convolutional layers, and three convolutional blocks (res4) comprising three convolutional layers, wherein the convolutional kernel size of the three convolutional layers contained in each convolutional block is 1 × 1, 3 × 3 and 1 × 1, and after module stacking, a feature fusion layer, an adaptive average pooling layer and a full connection layer are added, wherein the convolutional kernel size of the feature fusion layer and the adaptive average pooling layer is 1 × 1;
s33: the time domain flow part of the whole network model takes a residual error frame as input, the space domain part takes an RGB frame as input, a training set and a verification set are input into the network, Relu is selected as an activation function, a loss function is a cross entropy loss function, the batch size is set to be 3, the weight attenuation is set to be 0.0005, the momentum is set to be 0.9, the learning rate is 0.001, the weight of the model is optimized by using random gradient descent, and the iteration is stopped after the network converges, and the whole network model is stored.
5. The checkpoint diesel vehicle exhaust image-based smoke level estimation method as claimed in claim 4, wherein: in S4, using the trained network model, the specific operation of performing the grade estimation on the image to be detected is as follows:
s41: and inputting the test set into a trained smoke degree estimation model in S3 to obtain a corresponding smoke degree grade.
6. The utility model provides a smoke intensity level estimation system based on checkpoint diesel vehicle tail gas image which characterized in that:
the method comprises the following units:
the sub-block dividing unit is used for dividing the video collected by the inspection station into sub-blocks, sampling each sub-block and dividing a residual frame according to the obtained RGB frame;
the preprocessing unit is used for preprocessing the RGB image and the residual frame image;
the model construction unit is used for constructing a space-time cross network model, training the model by utilizing a training data set and a verification data set, finely adjusting and optimizing the whole network based on a preset loss function and an optimizer until the network is converged, and storing the trained network model parameters;
and the grade evaluation unit is used for carrying out grade estimation on the image to be detected by utilizing the trained network model.
7. A computer-readable storage medium, storing a computer program which, when executed by a processor, causes the processor to carry out the steps of the method according to any one of claims 1 to 5.
8. A computer device comprising a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the method according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110859026.1A CN113627499B (en) | 2021-07-28 | 2021-07-28 | Smoke level estimation method and equipment based on diesel vehicle tail gas image of inspection station |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110859026.1A CN113627499B (en) | 2021-07-28 | 2021-07-28 | Smoke level estimation method and equipment based on diesel vehicle tail gas image of inspection station |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113627499A true CN113627499A (en) | 2021-11-09 |
| CN113627499B CN113627499B (en) | 2024-04-02 |
Family
ID=78381346
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110859026.1A Active CN113627499B (en) | 2021-07-28 | 2021-07-28 | Smoke level estimation method and equipment based on diesel vehicle tail gas image of inspection station |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113627499B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115798624A (en) * | 2023-02-01 | 2023-03-14 | 中国汽车技术研究中心有限公司 | Intelligent modeling method for diesel vehicle tail gas SCR reaction process |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2000030074A1 (en) * | 1998-11-13 | 2000-05-25 | Qualcomm Incorporated | Low bit-rate coding of unvoiced segments of speech |
| CN106680281A (en) * | 2016-12-31 | 2017-05-17 | 中国科学技术大学 | Diesel vehicle tail gas smoke intensity detection method based on deep residual error learning network |
| WO2018171109A1 (en) * | 2017-03-23 | 2018-09-27 | 北京大学深圳研究生院 | Video action detection method based on convolutional neural network |
| CN110969577A (en) * | 2019-11-29 | 2020-04-07 | 北京交通大学 | Video super-resolution reconstruction method based on deep double attention network |
-
2021
- 2021-07-28 CN CN202110859026.1A patent/CN113627499B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2000030074A1 (en) * | 1998-11-13 | 2000-05-25 | Qualcomm Incorporated | Low bit-rate coding of unvoiced segments of speech |
| CN106680281A (en) * | 2016-12-31 | 2017-05-17 | 中国科学技术大学 | Diesel vehicle tail gas smoke intensity detection method based on deep residual error learning network |
| WO2018171109A1 (en) * | 2017-03-23 | 2018-09-27 | 北京大学深圳研究生院 | Video action detection method based on convolutional neural network |
| CN110969577A (en) * | 2019-11-29 | 2020-04-07 | 北京交通大学 | Video super-resolution reconstruction method based on deep double attention network |
Non-Patent Citations (1)
| Title |
|---|
| 袁帅;秦贵和;晏婕;: "应用残差生成对抗网络的路况视频帧预测模型", 西安交通大学学报, no. 10 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115798624A (en) * | 2023-02-01 | 2023-03-14 | 中国汽车技术研究中心有限公司 | Intelligent modeling method for diesel vehicle tail gas SCR reaction process |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113627499B (en) | 2024-04-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110348445B (en) | Instance segmentation method fusing void convolution and edge information | |
| CN110826684B (en) | Convolutional neural network compression method, convolutional neural network compression device, electronic device, and medium | |
| CN111046821B (en) | Video behavior recognition method and system and electronic equipment | |
| CN110991311A (en) | Target detection method based on dense connection deep network | |
| CN113486726A (en) | Rail transit obstacle detection method based on improved convolutional neural network | |
| CN110532925B (en) | Driver fatigue detection method based on space-time graph convolutional network | |
| CN111428558A (en) | Vehicle detection method based on improved YO L Ov3 method | |
| CN106228528A (en) | A kind of multi-focus image fusing method based on decision diagram Yu rarefaction representation | |
| CN111126278A (en) | Target detection model optimization and acceleration method for few-category scene | |
| CN113569756B (en) | Abnormal behavior detection and positioning method, system, terminal equipment and readable storage medium | |
| CN110599459A (en) | Underground pipe network risk assessment cloud system based on deep learning | |
| CN110717863A (en) | Single-image snow removing method based on generation countermeasure network | |
| CN115965827A (en) | Lightweight small target detection method and device integrating multi-scale features | |
| CN113963333B (en) | Traffic sign board detection method based on improved YOLOF model | |
| CN113627499A (en) | Smoke intensity grade estimation method and device based on inspection station diesel vehicle tail gas image | |
| Wu et al. | Single shot multibox detector for vehicles and pedestrians detection and classification | |
| CN113850135A (en) | Dynamic gesture recognition method and system based on time shift frame | |
| CN116740495A (en) | Training method and defect detection method for defect detection model of road and bridge tunnel | |
| CN115588150A (en) | Pet dog video target detection method and system based on improved YOLOv5-L | |
| Ni et al. | Enhanced knowledge distillation for face recognition | |
| CN114782983A (en) | Road scene pedestrian detection method based on improved feature pyramid and boundary loss | |
| CN115620051A (en) | Large-resolution image target detection method and system based on improved YOLOv5 model | |
| CN115239733A (en) | Crack detection method, crack detection device, terminal equipment and storage medium | |
| CN114463732A (en) | Scene text detection method and device based on knowledge distillation | |
| CN114943834A (en) | Full scene semantic segmentation method based on prototype queue learning under few-label samples |
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 | ||
| GR01 | Patent grant |