CN111914765A - Service area environment comfort level detection method and device and readable storage medium - Google Patents
Service area environment comfort level detection method and device and readable storage medium Download PDFInfo
- Publication number
- CN111914765A CN111914765A CN202010779020.9A CN202010779020A CN111914765A CN 111914765 A CN111914765 A CN 111914765A CN 202010779020 A CN202010779020 A CN 202010779020A CN 111914765 A CN111914765 A CN 111914765A
- Authority
- CN
- China
- Prior art keywords
- layer
- service area
- network
- comfort level
- convolution
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- 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
-
- 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 method, equipment and a readable storage medium for detecting environmental comfort level of a service area, wherein the method comprises the following steps: acquiring a service area environment comfort level detection image data set, calibrating the data set of pedestrians in a detection area, and determining a used one-stage target detection algorithm model; constructing a feature extraction upper sampling layer based on a scaling coefficient; constructing an upper sampling layer based on partial/integral region feature extraction; model training: and (3) the trained model is used for detecting and identifying the number of pedestrians in the detection area of the expressway service area in real time in an actual system, and then the environmental comfort level of the service area is calculated. The method has the advantages that the overall appearance and partial details of the target object are obtained simultaneously, the characteristics of all target areas and partial detail areas can be extracted in the up-sampling process, and the target detection precision of the model is further improved.
Description
Technical Field
The invention relates to the technical field of image recognition and computer vision, in particular to a service area environment comfort level detection method, equipment and a readable storage medium.
Background
With the promotion of highway construction pace and the improvement of people living standard, the requirements of road users on highway customization, comfort and the like are higher and higher, so that the users have more and more appeal on a large number of construction service areas. In the process, the construction of the expressway service area is from endless to ever, and the development process from the past by referring to foreign experiences to the current construction of the characteristic service area in China is realized. The service area is an indispensable facility of the highway, so that the basic requirements of drivers can be guaranteed, the vehicle running safety is improved, the transportation efficiency is improved, and the economic benefit of the highway is maximally exerted; the service area can provide services for users on the highway, and is very important for guaranteeing driving safety, guaranteeing transportation efficiency, and relieving physiological over-fatigue of drivers and limit conditions of vehicle use. With the continuous extension of the highway driving mileage and the continuous increase of the passenger flow volume, the service area is inevitably required to increase the types, the number and the like of service items, so that a special economic area is formed, and the importance of the industrial value is increasingly paid attention. With the rapid increase of the passenger flow in the service area, especially during holidays, a large number of tourists rush into the service area, and huge examination is brought to each work in the service area. Therefore, the passenger flow and environmental comfort data of the service area are mastered in time, reasonable measures are taken in time according to data changes, the passenger flow and the traffic flow are controlled, the optimal and most convenient service is provided for people to go out, and meanwhile, the working pressure of the service area is greatly reduced.
The environmental comfort level is calculated based on the crowd density grade, and is based on the computer vision and pattern recognition technology, and the quantization grade of the crowd density is obtained by analyzing and calculating the monitoring image or video. The environmental comfort level is a powerful basis for large-scale crowd monitoring management, and can provide crowd density information distributed according to different time periods in a market or a retail outlet to assist management layer in distributing service and managing resources. The system can be widely applied to monitoring of access passages of facilities such as bus stations, passenger transport stations, railway stations, airports and the like and people in important areas, accurate data of the number and distribution of passengers can be obtained in real time, potential safety hazards caused by crowding of customers can be prevented, and a basis is provided for scientific scheduling and safety guarantee. When crowd density is quantified as a density class, crowd density estimation is converted into a multi-class classification problem: the characteristic quantity related to the crowd is input, and the density level is output, namely the environmental comfort level. Common methods for solving the multi-class classification problem include polynomial fitting methods, recent clustering partitioning methods, and the like. Polynomial fitting methods often require certain assumptions about the functional relationship between input and output, and performance suffers when the assumptions do not match reality. The nearest clustering division method is that training samples with the same density grade are clustered in a feature space to form a plurality of representative points, and for a sample of an unknown class, the class of the sample is determined according to the nearest neighbor representative point or the class of the nearest neighbor representative point; the method is simple and easy to implement, but has high dependence on parameter selection, and performance is reduced or even overfitting is generated when the parameters are not properly selected.
The application of the current convolutional neural network algorithm based on video images in the field of image processing is gradually increased, and the convolutional neural network algorithm has the advantages of automatic image feature extraction, strong robustness and the like; however, the accuracy and real-time performance of the identification algorithm are high when the video stream of the camera is used for detecting and identifying the crowd density in the service area. Therefore, the target detection algorithm based on deep learning is reasonable. The target detection algorithm based on deep learning is divided into a two-stage model and a one-stage model. Although the two-stage convolutional neural network model has better detection accuracy, the forward reasoning speed is slow, and the real-time requirement of a service scene cannot be met. In a traditional one-stage target detection algorithm model, the real-time performance of the algorithm is good, but the detection precision of the two-stage convolutional neural network model cannot be achieved due to the fact that people are generally shielded mutually. The service area environment comfort level blockage detection method based on optimization of mutual crowd shading is beneficial to improving the detection precision of the system and ensuring that the real-time performance of the system meets the requirements of application scenes.
Disclosure of Invention
The method optimizes the upsampling process in a one-stage target detection algorithm model; the target characteristics are extracted in the deep learning up-sampling process, so that the training result of the convolutional neural network is directly influenced; the training result of the convolutional neural network directly affects the crowd density accuracy of the service area, so the design of the up-sampling process of the deep neural network with comfortable environment in the service area is particularly important.
The situation that a large number of people are shielded from each other exists in the service area environment comfort level detection training data set, for example, the left side of one person is shielded by other people, and then the target characteristic information brought by the right side of the person is more credible; it is because people are partly sheltered from and lead to the inaccuracy that target detected, and then influence the accuracy of environment comfort level index. Therefore, the invention provides a method for decomposing a target area into a plurality of small areas aiming at the problem of detection precision degradation caused by low target feature resolution due to shielding, so that the whole appearance and partial details of a target object can be simultaneously obtained, the features of all the target area and partial detail area can be extracted in the up-sampling process, and the target detection precision of the model is further improved.
In order to solve the technical problem, an up-sampling optimization strategy algorithm based on partial scaling factors and partial region features is adopted, and the specific technical scheme is as follows:
a service area environmental comfort level detection method, the method comprising the steps of:
1) constructing a service area environment comfort degree data set N, a training data set T, a verification data set V, a training data batch size batch, training batch numbers batches, a learning rate learninglate, a proportionality coefficient alpha between the training data set T and the verification data set V,
wherein: v T N, C is N+，α∈(0,1)，batches∈N+,learningrate∈N+，batch∈N+，
Representing the height and width of the image, r representing the imageThe number of channels of (a);
2) determining a stage target detection model to be trained, setting the depth of a convolutional neural network as L, setting a network convolutional layer convolutional kernel set G, setting a network output layer in a full-connection mode, setting a convolutional kernel set A and a network characteristic diagram set U,
representing the kth characteristic diagram in the l-th network
The corresponding grid number and anchor point set M are specifically defined as follows:
wherein:
respectively representing the height, width and dimension of a convolution kernel, a characteristic diagram and an anchor point corresponding to the layer I network,
indicating the fill size of the layer l network convolution kernel,
representing convolution step length of the layer I network, f representing excitation function of convolution neuron, theta representing selected input characteristic, and lambda epsilon N+Denotes the total number of anchor points xi E N in the layer I network+Represents the total number of nodes of the output layer, and belongs to N+Indicates the total number of the layer I network characteristic diagrams, and is Delta epsilon N+Represents the total number of the l layer convolution kernels;
3) designing a partial/whole region feature and target feature extraction model:
the region extraction adopts a scaling coefficient strategy beta to segment and extract the whole region to form a partial region, P is a region based on a scaling parameter beta, P is (x, y, w beta, h beta), meanwhile, a target object partial region Q is divided into an upper part, a lower part, a left part, a right part, Q epsilon { U, D, L, R },
wherein l belongs to { convolution layer 1, convolution layer 2}, r belongs to { scaling region P, target object partial region Q },
is a deviation factor; f (-) is the ReLU modified linear function unit, is the convolution operation,
is the convolution kernel of the l-th layer.
4) and outputting the number of pedestrians in the video detection area of the current service area through the trained network model, and calculating the ratio of the number of pedestrians to the area of the detection area to obtain the current environment comfort level.
An electronic device, comprising: the service area environment comfort level detection method comprises a memory and a processor, wherein the memory and the processor are mutually connected in a communication mode, computer instructions are stored in the memory, and the processor executes the computer instructions so as to execute the service area environment comfort level detection method.
A computer-readable storage medium having stored thereon computer instructions for causing the computer to execute the service area environmental comfort level detection method as described above.
The invention has the beneficial effects that: the method has the advantages that the overall appearance and partial details of the target object are obtained simultaneously, the characteristics of all target areas and partial detail areas can be extracted in the up-sampling process, and the target detection precision of the model is further improved.
Drawings
FIG. 1 is a diagram of a network architecture;
FIG. 2 is a schematic diagram of a partial/whole region feature and target feature extraction model;
fig. 3 is a flow chart of detection method deployment.
Detailed Description
The technical solution in the embodiments of the present invention is clearly and completely described below with reference to the drawings in the embodiments of the present invention.
According to the method for detecting the environmental comfort level of the service area, the sampling optimization strategy algorithm based on partial scaling factors and partial area features comprises the following steps:
step 1: the method comprises the steps of constructing 20000 data sets N by using pictures shot by a service area monitoring camera, wherein the number of training data sets T is 16000, the number of verification data sets V is 4000, the learning rate learningate value of 2 GPU video cards in a training hardware environment is 0.005, the value of a proportionality coefficient alpha between the training data sets T and the verification data sets V is 0.25, and the height h of the pictures is 0.25k＝608，wkR 608, r is 3 and satisfies that the height, width, and number of channels of all images are set to be consistent.
Step 2: determining the one-stage target detection model as YoloV3, setting the depth L of the convolutional neural network as 139, wherein the height, width and dimension settings of the convolutional kernel are specifically shown in FIG. 1, and the filling size of the convolutional kernel
Default to 1, convolution step size
Defaults to 1; anchor points are shared in each layer network, and the anchor point set M is set to be { (10,13), (30,61), (156,198) }, and Λ is 3; the network output layer adopts a full-connection mode, and the value of a convolution kernel set A is { (1,1,30), (1,1,30), (1,1,30) } and xi ═ 3.
And step 3: as shown in fig. 2, the scaling parameters are β ═ 0.5,0.7,1.2,1.5, model training is performed on the upsampling layer 1 and the upsampling layer 2 of the extraction model based on the partial/whole region features and the target features, and the model is trained by the gradient descent method using the training set until the model converges.
And 4, step 4: as shown in fig. 3, real-time detection is performed according to the video stream of the expressway service area, the number of target detection pedestrians in the current video monitoring area is output, and the environmental comfort level parameter is calculated.
The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (3)
1. A service area environmental comfort level detection method, characterized in that the method comprises the following steps:
1) constructing a service area environment comfort degree data set N, a training data set T, a verification data set V, a training data batch size batch, training batch numbers batches, a learning rate learninglate, a proportionality coefficient alpha between the training data set T and the verification data set V,
wherein: v T N, C is N+，α∈(0,1)，batches∈N+,learningrate∈N+，batch∈N+，
Representing the height and width of the image, and r represents the number of channels of the image;
2) determining a stage target detection model to be trained, setting the depth of a convolutional neural network as L, setting a network convolutional layer convolutional kernel set G, setting a network output layer in a full-connection mode, setting a convolutional kernel set A and a network characteristic diagram set U,
representing the kth characteristic diagram in the l-th network
The corresponding grid number and anchor point set M are specifically defined as follows:
wherein:
respectively representing the height, width and dimension of a convolution kernel, a characteristic diagram and an anchor point corresponding to the layer I network,
indicating the fill size of the layer l network convolution kernel,
representing convolution step length of the layer I network, f representing excitation function of convolution neuron, theta representing selected input characteristic, and lambda epsilon N+Denotes the total number of anchor points xi E N in the layer I network+Represents the total number of nodes of the output layer, and belongs to N+Indicates the total number of the layer I network characteristic diagrams, and is Delta epsilon N+Represents the total number of the l layer convolution kernels;
3) designing a partial/whole region feature and target feature extraction model:
the region extraction adopts a scaling coefficient strategy beta to segment and extract the whole region to form a partial region, P is a region based on a scaling parameter beta, P is (x, y, w beta, h beta), meanwhile, a target object partial region Q is divided into an upper part, a lower part, a left part, a right part, Q epsilon { U, D, L, R },
wherein l belongs to { convolution layer 1, convolution layer 2}, r belongs to { scaling region P, target object partial region Q },
is a deviation factor; f (-) is the ReLU modified linear function unit, is the convolution operation,
is the convolution kernel of the l-th layer.
4) and outputting the number of pedestrians in the video detection area of the current service area through the trained network model, and calculating the ratio of the number of pedestrians to the area of the detection area to obtain the current environment comfort level.
2. An electronic device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory having stored therein computer instructions, and the processor executing the computer instructions to perform the service area environment comfort level detection method of claim 1.
3. A computer-readable storage medium storing computer instructions for causing a computer to perform the service area environmental comfort level detection method of claim 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010779020.9A CN111914765B (en) | 2020-08-05 | 2020-08-05 | Service area environment comfort level detection method and device and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010779020.9A CN111914765B (en) | 2020-08-05 | 2020-08-05 | Service area environment comfort level detection method and device and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111914765A true CN111914765A (en) | 2020-11-10 |
| CN111914765B CN111914765B (en) | 2022-07-12 |
Family
ID=73287863
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010779020.9A Active CN111914765B (en) | 2020-08-05 | 2020-08-05 | Service area environment comfort level detection method and device and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111914765B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106960195A (en) * | 2017-03-27 | 2017-07-18 | 深圳市丰巨泰科电子有限公司 | A kind of people counting method and device based on deep learning |
| US20170344879A1 (en) * | 2016-05-31 | 2017-11-30 | Linkedln Corporation | Training a neural network using another neural network |
| CN107992841A (en) * | 2017-12-13 | 2018-05-04 | 北京小米移动软件有限公司 | The method and device of identification objects in images, electronic equipment, readable storage medium storing program for executing |
| CN108921822A (en) * | 2018-06-04 | 2018-11-30 | 中国科学技术大学 | Image object method of counting based on convolutional neural networks |
-
2020
- 2020-08-05 CN CN202010779020.9A patent/CN111914765B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170344879A1 (en) * | 2016-05-31 | 2017-11-30 | Linkedln Corporation | Training a neural network using another neural network |
| CN106960195A (en) * | 2017-03-27 | 2017-07-18 | 深圳市丰巨泰科电子有限公司 | A kind of people counting method and device based on deep learning |
| CN107992841A (en) * | 2017-12-13 | 2018-05-04 | 北京小米移动软件有限公司 | The method and device of identification objects in images, electronic equipment, readable storage medium storing program for executing |
| CN108921822A (en) * | 2018-06-04 | 2018-11-30 | 中国科学技术大学 | Image object method of counting based on convolutional neural networks |
Non-Patent Citations (2)
| Title |
|---|
| RUNZHOU GE ET AL；: "《AFDet: Anchor Free One Stage 3D Object Detection》", 《ARXIV:2006.12671V2》 * |
| 邵奇可 等；: "《基于深度学习的高速服务区车位检测算法》", 《计算机系统应用》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111914765B (en) | 2022-07-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110428428B (en) | Image semantic segmentation method, electronic equipment and readable storage medium | |
| CN110619369B (en) | Fine-grained image classification method based on feature pyramid and global average pooling | |
| CN111563902B (en) | Lung lobe segmentation method and system based on three-dimensional convolutional neural network | |
| CN110969626B (en) | Method for extracting hippocampus of human brain nuclear magnetic resonance image based on 3D neural network | |
| CN114120102A (en) | Boundary-optimized remote sensing image semantic segmentation method, device, equipment and medium | |
| CN111612754B (en) | MRI tumor optimization segmentation method and system based on multi-modal image fusion | |
| CN109101914A (en) | It is a kind of based on multiple dimensioned pedestrian detection method and device | |
| CN111597920B (en) | Full convolution single-stage human body example segmentation method in natural scene | |
| CN109034024B (en) | Logistics vehicle type classification and identification method based on image target detection | |
| CN114092917B (en) | MR-SSD-based shielded traffic sign detection method and system | |
| CN109447963A (en) | A kind of method and device of brain phantom identification | |
| CN112528961A (en) | Video analysis method based on Jetson Nano | |
| CN110659601A (en) | Depth full convolution network remote sensing image dense vehicle detection method based on central point | |
| CN115909006B (en) | Mammary tissue image classification method and system based on convolution transducer | |
| CN114445356A (en) | Multi-resolution-based full-field pathological section image tumor rapid positioning method | |
| CN113763371A (en) | Pathological image cell nucleus segmentation method and device | |
| CN114639101A (en) | Emulsion droplet identification system, method, computer equipment and storage medium | |
| CN111914765B (en) | Service area environment comfort level detection method and device and readable storage medium | |
| CN112132207A (en) | Target detection neural network construction method based on multi-branch feature mapping | |
| CN116884597A (en) | Pathological image breast cancer molecular typing method and system based on self-supervision pre-training and multi-example learning | |
| CN111914766B (en) | Method for detecting business trip behavior of city management service | |
| CN112598663B (en) | Grain pest detection method and device based on visual saliency | |
| CN112598664B (en) | Visual saliency-based insect grain grade determination method and device | |
| CN114565631A (en) | Plant leaf fine segmentation method based on double-layer convolution network and mask refinement | |
| CN114565626A (en) | Lung CT image segmentation algorithm based on PSPNet improvement |
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 |