CN113936261B - Processing method, device and equipment applied to highway monitoring data in plateau area - Google Patents
Processing method, device and equipment applied to highway monitoring data in plateau area Download PDFInfo
- Publication number
- CN113936261B CN113936261B CN202111549360.3A CN202111549360A CN113936261B CN 113936261 B CN113936261 B CN 113936261B CN 202111549360 A CN202111549360 A CN 202111549360A CN 113936261 B CN113936261 B CN 113936261B
- Authority
- CN
- China
- Prior art keywords
- adaptive
- neural network
- convolutional neural
- self
- network model
- 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
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
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a processing method, a processing device and processing equipment applied to highway monitoring data in a plateau area. The processing method comprises the following steps: constructing a self-adaptive hybrid convolutional neural network model based on a convolutional neural network; and extracting images from the highway monitoring video of the plateau area collected in real time, and inputting the extracted images into the trained adaptive hybrid convolutional neural network model for processing to obtain highway visibility data. According to the processing method applied to the highway monitoring data in the plateau area, the image in the highway monitoring video is processed through the self-adaptive hybrid convolutional neural network to identify the visibility on the highway, the identification speed is high, the identification accuracy is high, the visibility identification result can provide reference for workers, and the requirement of practical application can be well met.
Description
Technical Field
The application relates to the technical field of highway monitoring, in particular to a processing method, a processing device and processing equipment applied to highway monitoring data in a plateau area.
Background
With the advancement of science and technology and the development of economy, the road network has already reached extremely high coverage rate at present, and the road coverage is realized even in plateau areas, for example, highways such as the highways starting from song to pizza are constructed in plateau areas, which brings great convenience to the life of people. The visibility of the road has a direct influence on traffic safety, and low visibility brings potential safety hazards, so that traffic accidents are easily caused, and casualties and property loss are caused. Particularly in plateau areas, the fog days are more, and the low visibility of roads is easily caused. From the perspective of highway safety management, taking different management measures according to different situations of visibility is an important method for ensuring safe road traffic. Road monitoring data are processed to obtain accurate road visibility data, and corresponding measures can be taken in a targeted manner. Those skilled in the art are constantly working on optimizing the monitoring data processing function of highways in the highways to improve the visibility acquisition accuracy of highways in the highways.
Disclosure of Invention
The application aims to provide a processing method, a processing device and processing equipment applied to highway monitoring data in a plateau area. The following presents a simplified summary in order to provide a basic understanding of some aspects of the disclosed embodiments. This summary is not an extensive overview and is intended to neither identify key/critical elements nor delineate the scope of such embodiments. Its sole purpose is to present some concepts in a simplified form as a prelude to the more detailed description that is presented later.
According to an aspect of the embodiments of the present application, there is provided a processing method applied to highway monitoring data in a plateau area, including:
constructing a self-adaptive hybrid convolutional neural network model based on a convolutional neural network; the adaptive hybrid convolutional neural network model comprises a forward propagation part and a backward propagation part; the forward propagation part comprises a pre-training layer, a feature extraction layer, a connection layer and an output layer which are connected in sequence; the output layer is provided with a self-adaptive adjusting module; the self-adaptive adjusting module is used for carrying out self-adaptive updating on the weight and the bias of the pre-training layer; the backward propagation part is used for realizing the self-adaptive adjustment of the weight in the backward propagation process according to the analysis result of the self-adaptive adjusting module;
training the self-adaptive hybrid convolutional neural network model to obtain a trained self-adaptive hybrid convolutional neural network model;
and extracting images from the highway monitoring video of the plateau area collected in real time, and inputting the extracted images into the trained adaptive hybrid convolutional neural network model for processing to obtain highway visibility data.
In some embodiments of the present application, the building an adaptive hybrid convolutional neural network model based on a convolutional neural network comprises:
constructing the self-adaptive adjusting module;
constructing a user-defined network layer;
a support vector machine is used as the output layer.
In some embodiments of the present application, the constructing the adaptive adjustment module includes:
obtaining a classification result through a forward propagation part, and comparing the classification result with a classification true value to obtain a classification error;
determining an adaptive adjustment coefficient according to the iteration times and the classification result;
carrying out self-adaptive adjustment on an error value corresponding to the classification result to obtain an enhanced error;
calculating the adjusted residual error, feeding back the adjusted residual error to the hidden layer, and calculating to obtain the weight and the bias of the enhanced hidden layer;
and updating the weight and the bias of the enhanced hidden layer through the back propagation part.
In some embodiments of the present application, the pre-training layer is a model structure combining ResNet and VGG 19; the feature extraction layer comprises a plurality of convolutional layers and a plurality of pooling layers; the connection layer includes an SVM classifier.
The method comprises the steps of extracting images from highway monitoring videos of the plateau area collected in real time, inputting the extracted images into the trained adaptive hybrid convolutional neural network model for processing, and comprises the following steps:
extracting a plurality of images in the same preset time period from the highway monitoring video of the plateau area collected in real time according to a preset period;
comparing the plurality of images in the same preset time period, and selecting an image with the highest image quality according to a comparison result;
and inputting the image with the highest image quality into the trained self-adaptive hybrid convolutional neural network model for processing.
In some embodiments of the present application, the comparing the plurality of images in the same preset time period, and selecting an image with the highest image quality according to a comparison result includes:
and comparing the image noise amount contained in the plurality of images in the same preset time period, and selecting the image with the least image noise amount as the image with the highest image quality according to the comparison result.
According to another aspect of the embodiments of the present application, there is provided a processing device for highway monitoring data in a plateau area, including:
the building module is used for building a self-adaptive hybrid convolutional neural network model based on a convolutional neural network; the adaptive hybrid convolutional neural network model comprises a forward propagation part and a backward propagation part; the forward propagation part comprises a pre-training layer, a feature extraction layer, a connection layer and an output layer which are connected in sequence; the output layer is provided with a self-adaptive adjusting module; the self-adaptive adjusting module is used for carrying out self-adaptive updating on the weight and the bias of the pre-training layer; the backward propagation part is used for realizing the self-adaptive adjustment of the weight in the backward propagation process according to the analysis result of the self-adaptive adjusting module;
the training module is used for training the self-adaptive hybrid convolutional neural network model to obtain a trained self-adaptive hybrid convolutional neural network model;
and the extraction processing module is used for extracting images from the highway monitoring video of the plateau area acquired in real time, and inputting the extracted images into the trained adaptive hybrid convolutional neural network model for processing to obtain highway visibility data.
In some embodiments of the present application, the building module comprises:
the first construction unit is used for constructing the self-adaptive adjusting module;
the second construction unit is used for constructing a custom network layer;
and the third construction unit is used for adopting the support vector machine as an output layer.
According to another aspect of the embodiments of the present application, there is provided an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the program to implement any one of the processing methods described above applied to highway monitoring data in a plateau area.
According to another aspect of embodiments of the present application, there is provided a computer-readable storage medium having a computer program stored thereon, where the computer program is executed by a processor to implement any one of the processing methods applied to highway monitoring data in a highlands.
The technical scheme provided by one aspect of the embodiment of the application can have the following beneficial effects:
according to the processing method applied to the highway monitoring data in the plateau area, the image in the highway monitoring video is processed through the self-adaptive hybrid convolutional neural network to identify the visibility on the highway, the identification speed is high, the identification accuracy is high, the visibility identification result can provide reference for workers, and the requirement of practical application can be well met.
Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the embodiments of the application, or may be learned by the practice of the embodiments.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments described in the present application, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart illustrating a processing method applied to highway monitoring data in a plateau region according to an embodiment of the present application;
FIG. 2 illustrates an adaptive hybrid convolutional neural network model structure diagram of one embodiment of the present application;
fig. 3 shows a flowchart of step S10 in fig. 1;
fig. 4 shows a flowchart of step S101 in fig. 3;
fig. 5 is a block diagram illustrating a processing device applied to highway monitoring data in a plateau region according to another embodiment of the present application;
FIG. 6 shows a block diagram of the construction of the building block of FIG. 5;
FIG. 7 shows a block diagram of the first building element of FIG. 6;
fig. 8 shows a block diagram of an electronic device of another embodiment of the present application;
FIG. 9 shows a schematic diagram of a computer-readable storage medium of another embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is further described with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
It will be understood by those within the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
The plateau area has more foggy days and the visibility of the highway is lower in the foggy days. Road management personnel need to take corresponding measures according to visibility of each road section of the road, such as temporary closing of a highway section and the like. The manual judgment of the visibility of each road section of the highway usually requires a large amount of manual labor, the labor cost is high, and the accuracy of visibility identification results is low. The video acquisition equipment in the road monitoring system, such as cameras and other equipment, acquires the road video in real time. The video of the road is one of the road monitoring data. Images in the video may be used to construct a training set and a test set for training the adaptive hybrid convolutional neural network model. The road monitoring data can be processed by adopting a convolutional neural network. Convolutional Neural Networks (CNN) are feed-forward Neural Networks which comprise Convolutional calculation and have a deep structure, have characterization learning capacity, can perform translation invariant classification on input information according to a hierarchical structure of the Convolutional Neural Networks, are constructed by imitating a visual perception mechanism of a living being, can perform supervised learning and unsupervised learning, and can enable the Convolutional Neural Networks to perform lattice characterization with smaller calculation amount due to convolution kernel parameter sharing in a hidden layer and sparsity of interlayer connection.
As shown in fig. 1, an embodiment of the present application provides a processing method applied to highway monitoring data in a plateau area, including steps S10-S30.
And S10, constructing an adaptive hybrid convolutional neural network model based on the convolutional neural network.
The adaptive hybrid convolutional neural network model structure constructed in the embodiment of the application is shown in fig. 3 and comprises a forward propagation part and a backward propagation part.
The forward propagation part comprises a pre-training layer, a feature extraction layer, a connection layer and an output layer which are connected in sequence. The pre-training layer adopts a model structure combining ResNet and VGG 19; the feature extraction layer comprises a plurality of convolution layers and a plurality of pooling layers and is used for extracting image features; the connection layer comprises an SVM classifier for classifying images; and the output layer is provided with a self-adaptive adjusting module.
And the backward propagation part is used for realizing the self-adaptive adjustment of the weight in the backward propagation process according to the analysis result of the self-adaptive adjusting module. The input data is subjected to a forward propagation part to obtain classification results, wherein each classification value corresponds to a unique class, the class corresponding to the maximum value is identified as the class to which the input belongs, and the class to which the input data actually belongs is called a true value class; the classification true value is training supervision data, and real classification results corresponding to input data are stored, wherein the value corresponding to the true value category is 1, and the values corresponding to the other categories are 0; the classification error is generated by the classification result and the classification truth value.
As shown in fig. 2, in a specific embodiment, the forward propagation part normalizes the image, and after the normalization results in a size of 224 × 224, the normalized image is input into a pre-training layer composed of ResNet and VGG19, and a filtering operation is performed by using 6 convolution kernels of 5 × 5, so that a feature map is obtained as C1, and C1 is down-sampled by 6 maximum values of 24 × 24 to obtain S1 layers; at the level of S1, convolving the data with 32 convolution kernels with the size of 5 × 5 to obtain an output feature map C2, wherein the feature map C2 consists of 6 feature maps with the size of 12 × 12; and after an S2 layer, generating a 1 × 256 one-dimensional vector of a fully-connected layer, wherein the layer comprises a hidden layer formed by 84 neurons, the hidden layer is connected with an output layer through a linear relation and comprises five nodes, the value of each node is 0 or 1, and the node is a classification label of the visibility grade to be predicted. And after the classification label is compared and analyzed with the true value, the weight coefficient of the misclassification sample is enhanced, and then the reverse propagation is carried out so as to update the weight of the pre-training layer and adaptively adjust the weight of the network layer. The structure of the adaptive hybrid convolutional neural network model is shown in the following table.
Table network layer structure parameter
The structure of the adaptive hybrid convolutional neural network model can be adjusted according to actual needs, and is not limited herein.
And S20, training the self-adaptive hybrid convolutional neural network model to obtain the trained self-adaptive hybrid convolutional neural network model.
In the adaptive hybrid convolutional neural network model, an error function E (epsilon, b) is utilized to measure the training effect of each parameter in the hidden layer on the input image. By reducing the output of the error function during the iterative training process, the classification result k is obtainedωMaximally approaching classification true value k ̂ωAnd when the error of two adjacent times of learning is smaller than a preset threshold value, determining that the network model training is in a convergence state, and finishing the learning.
For a multi-class problem, the objective error function is defined as:
wherein the content of the first and second substances,inacc ω representing the error between the classification result of the ω -th class and the true value, and n is the number of classifications.
Based on the gradient descent method, the expression of the variable quantity of the weight and the bias is as follows:
where d is the residual, x is the value in the input feature map,f′as a function of activationfAnd (6) carrying out derivation.
And S30, extracting images from the highway monitoring video of the plateau area collected in real time, and inputting the extracted images into the trained adaptive hybrid convolutional neural network model for processing to obtain highway visibility data.
For example, a plurality of images within the same preset time period (the time period may be set according to actual needs, for example, may be set to 1 minute or 2 minutes, etc.) may be extracted from the highway monitoring video in the plateau area collected in real time according to a preset period (for example, the preset period may be set to 12 hours or 8 hours, etc., and may be specifically set according to actual needs), the plurality of images are compared, one image with the highest image quality is selected according to the comparison result, and the image with the highest image quality is input into the trained adaptive hybrid convolutional neural network model for processing, so as to obtain the highway visibility data. For example, the amounts of image noise included in the plurality of images may be compared, and an image having the smallest amount of image noise may be selected as the highest quality image.
For another example, the images may be randomly extracted from the highway monitoring videos in the plateau area collected in real time, for example, in the case of a foggy day, the images may be randomly extracted from the highway monitoring videos in the plateau area collected in real time (for example, the images are extracted once or multiple times every hour according to the severity of the fog), instead of being extracted according to a preset period, which may be specifically set according to actual needs.
As shown in fig. 3, in some embodiments, step S10 includes:
s101, constructing the self-adaptive adjusting module; the self-adaptive adjusting module is used for carrying out self-adaptive updating on the weight and the bias of the pre-training layer and adjusting the weight parameter so as to obtain an optimized self-adaptive hybrid convolutional neural network model;
s102, constructing a custom network layer; the user-defined network layer is used for further training the optimized self-adaptive hybrid convolutional neural network model;
s103, adopting a support vector machine as an output layer; the support vector machine is used for multi-classification.
The self-adaptive hybrid convolutional neural network model can solve the problem of self-adaptive enhancement of classification characteristics under different recognition results and iteration times, can realize accelerated convergence of the network model, and improves the recognition accuracy.
In some embodiments, the comparing the plurality of images in the same preset time period, and selecting an image with the highest image quality according to the comparison result includes:
and comparing the image noise amount contained in the plurality of images in the same preset time period, and selecting the image with the least image noise amount as the image with the highest image quality according to the comparison result.
In certain embodiments, as shown in fig. 4, step S101 comprises the steps of:
(1) a classification error is calculated. The self-adaptive mixed convolution neural network obtains a classification result through a forward propagation part, the classification result is compared with a classification true value to obtain a classification error, and the calculation mode is shown as a formula (1).
(2) And determining the self-adaptive adjusting coefficient according to the iteration times and the classification result of the forward propagation part. The basic computational expression for the adjustment coefficient μ is:
wherein Gaus = e ^ [ (x-b)2/(2c2)]K is the product coefficient, η is the number of current iterations, and ε is the correction term.
(3) And carrying out self-adaptive adjustment on the error value corresponding to the classification result to obtain an enhanced error. Wherein, the calculation expression of the ω -th classification enhancement error is:
inacc′ω = μω × inaccω（6）；
(4) the adjusted residual is calculated. The error calculated by equation (6) is taken into equation (4), and the adjusted residual error is calculated.
(5) And feeding back the adjusted residual error to the hidden layer, and calculating to obtain the weight and the bias of the enhanced hidden layer. The calculation method of the weight and the bias in the hidden layer is shown in the formulas (2) and (3).
(6) And updating the weight and the bias of the enhanced hidden layer through a back propagation part.
According to the self-adaptive hybrid convolutional neural network provided by the embodiment of the application, a self-adaptive adjusting module is additionally arranged between a forward propagation part and a backward propagation part of the convolutional neural network, and the weight and the offset amplitude of a hidden layer are self-adaptively adjusted according to the iteration times and the error analysis of a classification result; the enhanced residual error is fed back to the hidden layer parameters through the back propagation part, so that the self-adaptive updating of the weight and the bias in the convolution kernel and the full connection layer is realized, and the training classification effect of the model in the next period is improved.
According to the self-adaptive hybrid convolutional neural network model constructed in the embodiment of the application, a pretraining layer combining ResNet and VGG19 is used for transfer learning in an input layer of the convolutional neural network, so that the overfitting phenomenon of a training result can be avoided; a parameter adaptive adjusting module is constructed between a forward propagation part and a backward propagation part of the adaptive hybrid convolutional neural network model, and weight adaptive updating is realized according to training period number of the convolutional neural network and training error enhancement parameter coefficients, so that the convergence speed of the adaptive hybrid convolutional neural network model training and the accuracy of visibility identification are effectively improved.
According to the processing method applied to the highway monitoring data in the plateau area, the images in the highway monitoring video are processed through the self-adaptive hybrid convolutional neural network to identify the visibility of each road section on the highway, the identification speed is high, the identification accuracy is high, the visibility identification result can provide reference for workers, the manual workload is reduced, the labor cost is reduced, and the requirements of practical application can be well met. The identification result of each road section of the road is accurate, so that the identification is not needed to be carried out manually, and the staff can take corresponding measures according to the identification result, for example, the road section with low visibility implements road sealing measures and the like.
As shown in fig. 5, another embodiment of the present application provides a processing device applied to highway monitoring data in a plateau area, including:
the building module is used for building a self-adaptive hybrid convolutional neural network model based on a convolutional neural network; the adaptive hybrid convolutional neural network model comprises a forward propagation part and a backward propagation part; the forward propagation part comprises a pre-training layer, a feature extraction layer, a connection layer and an output layer which are connected in sequence; the output layer is provided with a self-adaptive adjusting module; the self-adaptive adjusting module is used for carrying out self-adaptive updating on the weight and the bias of the pre-training layer; the backward propagation part is used for realizing the self-adaptive adjustment of the weight in the backward propagation process according to the analysis result of the self-adaptive adjusting module;
the training module is used for training the self-adaptive hybrid convolutional neural network model to obtain a trained self-adaptive hybrid convolutional neural network model;
and the extraction processing module is used for extracting images from the highway monitoring video of the plateau area acquired in real time, and inputting the extracted images into the trained adaptive hybrid convolutional neural network model for processing to obtain highway visibility data.
In certain embodiments, as shown in fig. 6, the building block comprises:
the first construction unit is used for constructing the self-adaptive adjusting module;
the second construction unit is used for constructing a custom network layer;
and the third construction unit is used for adopting the support vector machine as an output layer.
In some embodiments, as shown in fig. 7, the first building unit includes:
the comparison subunit is used for obtaining a classification result through the forward propagation part and comparing the classification result with the classification true value to obtain a classification error;
the determining subunit is used for determining the self-adaptive adjusting coefficient according to the iteration times and the classification result;
the adjusting subunit is used for carrying out self-adaptive adjustment on the error value corresponding to the classification result to obtain an enhanced error;
the calculation subunit is used for calculating the adjusted residual error, feeding the adjusted residual error back to the hidden layer, and calculating to obtain the weight and the offset of the enhanced hidden layer;
and the updating subunit is used for updating the weight and the bias of the enhanced hidden layer through the back propagation part.
Another embodiment of the present application provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor executes the program to implement the processing method applied to highway monitoring data in a plateau area according to any one of the above embodiments.
As shown in fig. 8, the electronic device 10 may include: the system comprises a processor 100, a memory 101, a bus 102 and a communication interface 103, wherein the processor 100, the communication interface 103 and the memory 101 are connected through the bus 102; the memory 101 stores a computer program that can be executed on the processor 100, and the processor 100 executes the computer program to perform the method provided by any of the foregoing embodiments of the present application.
The Memory 101 may include a high-speed Random Access Memory (RAM), and may also include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. The communication connection between the network element of the system and at least one other network element is realized through at least one communication interface 103 (which may be wired or wireless), and the internet, a wide area network, a local network, a metropolitan area network, and the like can be used.
The bus 102 may be an ISA bus, PCI bus, EISA bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory 101 is used for storing a program, and the processor 100 executes the program after receiving an execution instruction, and the method disclosed in any of the foregoing embodiments of the present application may be applied to the processor 100, or implemented by the processor 100.
The electronic device provided by the embodiment of the application and the method provided by the embodiment of the application have the same inventive concept and have the same beneficial effects as the method adopted, operated or realized by the electronic device.
Another embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a processor to implement the processing method applied to highway monitoring data in a plateau area according to any one of the above embodiments.
Referring to fig. 9, the computer-readable storage medium is an optical disc 20, on which a computer program (i.e., a program product) is stored, and when the computer program is executed by a processor, the computer program performs the method of any of the foregoing embodiments.
It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, a phase change memory (PRAM), a Static Random Access Memory (SRAM), a Dynamic Random Access Memory (DRAM), other types of Random Access Memories (RAM), a Read Only Memory (ROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a flash memory, or other optical and magnetic storage media, which are not described in detail herein.
The computer-readable storage medium provided by the above-mentioned embodiments of the present application and the method provided by the embodiments of the present application have the same advantages as the method adopted, executed or implemented by the application program stored in the computer-readable storage medium.
It should be noted that:
the term "module" is not intended to be limited to a particular physical form. Depending on the particular application, a module may be implemented as hardware, firmware, software, and/or combinations thereof. Furthermore, different modules may share common components or even be implemented by the same component. There may or may not be clear boundaries between the various modules.
The algorithms and displays presented herein are not inherently related to any particular computer, virtual machine, or other apparatus. Various general purpose devices may also be used with the examples based on this disclosure. The required structure for constructing such a device will be apparent from the description above. In addition, this application is not directed to any particular programming language. It will be appreciated that a variety of programming languages may be used to implement the teachings of the present application as described herein, and any descriptions of specific languages are provided above to disclose the best modes of the present application.
It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in a strict order unless explicitly stated in the embodiments of the present application, and may be performed in other orders. Moreover, at least a portion of the steps in the flow chart of the figure may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed alternately or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
The above-mentioned embodiments only express the embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.
Claims (7)
1. A processing method applied to highway monitoring data in a plateau area is characterized by comprising the following steps:
constructing a self-adaptive hybrid convolutional neural network model based on a convolutional neural network; the adaptive hybrid convolutional neural network model comprises a forward propagation part and a backward propagation part; the forward propagation part comprises a pre-training layer, a feature extraction layer, a connection layer and an output layer which are connected in sequence; the output layer is provided with a self-adaptive adjusting module; the self-adaptive adjusting module is used for carrying out self-adaptive updating on the weight and the bias of the pre-training layer; the backward propagation part is used for realizing the self-adaptive adjustment of the weight in the backward propagation process according to the analysis result of the self-adaptive adjusting module;
training the self-adaptive hybrid convolutional neural network model to obtain a trained self-adaptive hybrid convolutional neural network model;
extracting images from highway monitoring videos of the plateau area collected in real time, and inputting the extracted images into the trained adaptive hybrid convolutional neural network model for processing to obtain highway visibility data;
training the adaptive hybrid convolutional neural network model to obtain a trained adaptive hybrid convolutional neural network model, including: measuring the training effect of each parameter in the hidden layer on the input image by using an error function E (epsilon, b); by reducing the output of the error function during the iterative training process, the classification result k is obtainedωMaximally approaching classification true value k ̂ωWhen the error of two adjacent times of learning is smaller than a preset threshold value, determining that the network model training is in a convergence state, and finishing the learning;
wherein the content of the first and second substances,inacc ω representing the error between the classification result of the ω -th class and the true value, n being the number of classifications;
the method for constructing the self-adaptive hybrid convolutional neural network model based on the convolutional neural network comprises the following steps:
constructing the self-adaptive adjusting module;
constructing a user-defined network layer;
a support vector machine is adopted as an output layer;
the constructing of the adaptive adjustment module includes:
obtaining a classification result through a forward propagation part, and comparing the classification result with a classification true value to obtain a classification error;
determining an adaptive adjustment coefficient according to the iteration times and the classification result;
carrying out self-adaptive adjustment on an error value corresponding to the classification result to obtain an enhanced error;
calculating the adjusted residual error, feeding back the adjusted residual error to the hidden layer, and calculating to obtain the weight and the bias of the enhanced hidden layer;
and updating the weight and the bias of the enhanced hidden layer through the back propagation part.
2. The process of claim 1, wherein the pre-training layer is a model structure combining ResNet and VGG 19; the feature extraction layer comprises a plurality of convolutional layers and a plurality of pooling layers; the connection layer includes an SVM classifier.
3. The processing method according to claim 1, wherein the extracting images from the real-time collected highroad monitoring video of the plateau area, and inputting the extracted images into the trained adaptive hybrid convolutional neural network model for processing comprises:
extracting a plurality of images in the same preset time period from the highway monitoring video of the plateau area collected in real time according to a preset period;
comparing the plurality of images in the same preset time period, and selecting an image with the highest image quality according to a comparison result;
and inputting the image with the highest image quality into the trained self-adaptive hybrid convolutional neural network model for processing.
4. The processing method according to claim 3, wherein the comparing the plurality of images within the same preset time period and selecting an image with the highest image quality according to the comparison result comprises:
and comparing the image noise amount contained in the plurality of images in the same preset time period, and selecting the image with the least image noise amount as the image with the highest image quality according to the comparison result.
5. A processing apparatus for being applied to plateau district highway monitoring data, characterized by includes:
the building module is used for building a self-adaptive hybrid convolutional neural network model based on a convolutional neural network; the adaptive hybrid convolutional neural network model comprises a forward propagation part and a backward propagation part; the forward propagation part comprises a pre-training layer, a feature extraction layer, a connection layer and an output layer which are connected in sequence; the output layer is provided with a self-adaptive adjusting module; the self-adaptive adjusting module is used for carrying out self-adaptive updating on the weight and the bias of the pre-training layer; the backward propagation part is used for realizing the self-adaptive adjustment of the weight in the backward propagation process according to the analysis result of the self-adaptive adjusting module;
the training module is used for training the self-adaptive hybrid convolutional neural network model to obtain a trained self-adaptive hybrid convolutional neural network model;
the extraction processing module is used for extracting images from highway monitoring videos of the plateau area collected in real time, inputting the extracted images into the trained adaptive hybrid convolutional neural network model for processing, and obtaining highway visibility data;
training the adaptive hybrid convolutional neural network model to obtain a trained adaptive hybrid convolutional neural network model, including: measuring the training effect of each parameter in the hidden layer on the input image by using an error function E (epsilon, b); by reducing the output of the error function during the iterative training process, the classification result k is obtainedωMaximally approaching classification true value k ̂ωWhen the error of two adjacent times of learning is smaller than a preset threshold value, determining that the training of the network model is in a convergence state, and learningThe learning is finished;
wherein the content of the first and second substances,inacc ω representing the error between the classification result of the ω -th class and the true value, n being the number of classifications;
the building module comprises:
the first construction unit is used for constructing the self-adaptive adjusting module;
the second construction unit is used for constructing a custom network layer;
the third construction unit is used for adopting a support vector machine as an output layer;
the constructing of the adaptive adjustment module includes:
obtaining a classification result through a forward propagation part, and comparing the classification result with a classification true value to obtain a classification error;
determining an adaptive adjustment coefficient according to the iteration times and the classification result;
carrying out self-adaptive adjustment on an error value corresponding to the classification result to obtain an enhanced error;
calculating the adjusted residual error, feeding back the adjusted residual error to the hidden layer, and calculating to obtain the weight and the bias of the enhanced hidden layer;
and updating the weight and the bias of the enhanced hidden layer through the back propagation part.
6. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor executing the program to implement the method of any one of claims 1-4.
7. A computer-readable storage medium, on which a computer program is stored, characterized in that the program is executed by a processor to implement the method according to any of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111549360.3A CN113936261B (en) | 2021-12-17 | 2021-12-17 | Processing method, device and equipment applied to highway monitoring data in plateau area |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111549360.3A CN113936261B (en) | 2021-12-17 | 2021-12-17 | Processing method, device and equipment applied to highway monitoring data in plateau area |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113936261A CN113936261A (en) | 2022-01-14 |
| CN113936261B true CN113936261B (en) | 2022-03-22 |
Family
ID=79289341
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111549360.3A Active CN113936261B (en) | 2021-12-17 | 2021-12-17 | Processing method, device and equipment applied to highway monitoring data in plateau area |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113936261B (en) |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109816625A (en) * | 2018-11-27 | 2019-05-28 | 广东电网有限责任公司 | A kind of video quality score implementation method |
| CN109740670B (en) * | 2019-01-02 | 2022-01-11 | 京东方科技集团股份有限公司 | Video classification method and device |
| CN112749654A (en) * | 2020-12-31 | 2021-05-04 | 南京恩瑞特实业有限公司 | Deep neural network model construction method, system and device for video fog monitoring |
-
2021
- 2021-12-17 CN CN202111549360.3A patent/CN113936261B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113936261A (en) | 2022-01-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111079639B (en) | Method, device, equipment and storage medium for constructing garbage image classification model | |
| US20220108546A1 (en) | Object detection method and apparatus, and computer storage medium | |
| CN111027493B (en) | Pedestrian detection method based on deep learning multi-network soft fusion | |
| US20220215227A1 (en) | Neural Architecture Search Method, Image Processing Method And Apparatus, And Storage Medium | |
| CN111126704B (en) | Multi-region precipitation prediction model construction method based on multi-graph convolution and memory network | |
| EP3792840A1 (en) | Neural network method and apparatus | |
| CN111563431A (en) | Plant leaf disease and insect pest identification method based on improved convolutional neural network | |
| CN109815979B (en) | Weak label semantic segmentation calibration data generation method and system | |
| CN111310604A (en) | Object detection method and device and storage medium | |
| CN104915972A (en) | Image processing apparatus, image processing method and program | |
| CN113592060A (en) | Neural network optimization method and device | |
| CN111783754A (en) | Human body attribute image classification method, system and device based on part context | |
| CN112699941A (en) | Plant disease severity image classification method and device, computer equipment and storage medium | |
| CN113051983B (en) | Method for training field crop disease recognition model and field crop disease recognition | |
| Hu et al. | Perception granular computing in visual haze-free task | |
| CN115439654A (en) | Method and system for finely dividing weakly supervised farmland plots under dynamic constraint | |
| Kiran et al. | Dimensionality reduction procedure for bigdata in machine learning techniques | |
| Sharifi et al. | A new deep learning approach for classification of hyperspectral images: Feature and decision level fusion of spectral and spatial features in multiscale CNN | |
| CN116861262B (en) | Perception model training method and device, electronic equipment and storage medium | |
| Vuković et al. | Thermal image degradation influence on R-CNN face detection performance | |
| Sari et al. | Daily rainfall prediction using one dimensional convolutional neural networks | |
| CN113936261B (en) | Processing method, device and equipment applied to highway monitoring data in plateau area | |
| WO2024011859A1 (en) | Neural network-based face detection method and device | |
| Jadhav et al. | Comprehensive review on machine learning for plant disease identification and classification with image processing | |
| Shamiul Islam et al. | Hybrid Convolution Neural Network with Transfer Learning Approach for Agro-Crop Leaf Disease Identification |
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 |