Dynamic monitoring method and system for catering source emissions based on video intelligent analysis

Abstract

The present invention belongs to the technical field of pollution source monitoring, and relates to a method and system for dynamic monitoring of catering source emissions based on video intelligent analysis, including: obtaining catering enterprise information and video images in a region, and inputting the enterprise information and video images into a deep learning model for feature extraction; selecting emission factors corresponding to an emission model according to the enterprise information; generating a feature variation curve graph according to the features; generating an initial emission model according to the emission factor and the feature variation curve graph; training the initial emission model using video images to obtain a final emission model; and inputting the enterprise information and video images to be tested into the final emission model to obtain emission results for the enterprise to be tested.

Description

Technical Field

[0001] The present invention relates to a method and system for dynamic monitoring of restaurant source emissions based on video intelligent analysis, belonging to the technical field of pollution source monitoring, and in particular to emission monitoring of pollutants in restaurant fumes. Background Art

[0002] With the acceleration of China's urbanization process and the increase in the number of urban population, catering sources that are more closely related to human life have become a prominent environmental problem that urban managers face. Due to the closeness of catering sources to human life and the closer distance to human breathing, the sensory stimulation and health impacts they produce are more sensitive. Conducting research on emissions and monitoring technologies from catering sources plays a very important role in strengthening the control and management of catering pollution.

[0003] The vast majority of particulate matter emitted from catering sources are fine particles, which are mainly produced by high-temperature decomposition and oxidation during the heating of cooking oil or food. A large number of simulated emission studies from indoor catering sources have observed that about 90% of the particulate matter size range is within the ultrafine particle size range. Particulate matter emitted from catering is an important source of fine organic particles in the urban atmosphere. The receptor model source analysis results show that the contribution of catering sources to the atmospheric fine organic particles is about 23%.

[0004] In the prior art, the emission inventory of catering sources is usually investigated through statistical data or door-to-door surveys. However, the calculation method for establishing a city catering source inventory is relatively rough, and the emission model usually treats catering sources in an average manner, with low spatiotemporal resolution. Although the door-to-door survey method can more accurately obtain the geographical location of catering companies and improve the spatial resolution of the catering source emission inventory, it fails to reflect the dynamic change characteristics of catering emission sources over time. The online fume monitoring sensor can monitor the concentration of fume pollutants in real time, but its detection probe is easy to dirty and data calibration problems affect the accuracy of its monitoring data, and cannot meet the technical requirements of low-cost and high-precision dynamic monitoring of catering sources. Summary of the invention

[0005] In response to the above problems, the purpose of the present invention is to provide a dynamic monitoring method, system and readable medium for catering source emissions. The pollutant emission monitoring method established through an algorithm is low-cost, high-precision and easy to maintain compared to traditional monitoring methods. It breaks through the regulatory blind spots of environmental managers and improves management efficiency and effectiveness.

[0006] To achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a method for dynamic monitoring of emissions from catering sources, comprising: obtaining catering enterprise information and video images in a region, inputting the video images of the enterprise's cooking activities into a video analysis algorithm for feature extraction; selecting an emission factor corresponding to the emission model based on the enterprise information; generating a graph of the number of stoves used over time based on the features extracted from the video images of the stoves during the cooking activities; generating an initial emission model based on the emission factors and the graph of the number of stoves used over time; training the initial emission model using video images to obtain a final emission model; inputting the enterprise information and video images to the final emission model to obtain the emission results of the enterprise to be tested.

[0007] Furthermore, the catering enterprise information includes stove data, working time data and auxiliary information.

[0008] Furthermore, the catering enterprise information specifically includes: enterprise scale, enterprise cuisine type, number of stoves, stove type, enterprise name, enterprise type, fuel information, cooking oil information, business hours and customer flow information.

[0009] Furthermore, the method of inputting the video image of the enterprise's cooking activity into the video analysis algorithm for feature extraction is: extract each frame of the video image of the enterprise's cooking activity; extract the static features of each frame of the image, determine the type of the image according to the static features, and encode and mark it; according to the type of the image, extract dynamic features of the video image of the cooking activity, collect continuous frames in the video sequence for 3D convolution, and obtain information on the chef's actions; determine the usage time of the stove, combine the stove usage time, static features and dynamic features, and obtain the final emission model.

[0010] Furthermore, the method for determining the stove usage time is: use a deep learning model to perform feature training on the video image, mark the time from pouring oil to the pot leaving the stove as the known stove usage time; input the known stove usage time obtained by the deep learning algorithm into the supervised learning algorithm model, train the supervised learning algorithm model to obtain the optimal supervised learning model; input the video to be tested into the optimal supervised learning model to obtain its corresponding stove usage time.

[0011] Furthermore, the emission model formula obtained by combining the stove usage time, static characteristics and dynamic characteristics is:

[0012] E i =EF i ×n×t

[0013] Among them, E i is the emission of a certain pollutant i from the catering source, EFi is the emission factor of the pollutant, n is the number of working stoves, and t is the working time of the stoves.

[0014] Furthermore, the pollutant emission concentration of the monitored catering enterprise obtained according to the emission of a certain pollutant i from the catering source is:

[0015]

[0016] Among them, Ci is the emission concentration of a certain pollutant i from the catering source, E i is the emission of a certain pollutant i from the catering source, t is the emission time, V is the flue gas flow rate, and η is the purification efficiency of the purifier for the pollutant.

[0017] Furthermore, the static features include: stove, pots, oil, spatula, chopping board and seasoning; the dynamic features include preparing food, pouring oil, stir-frying, serving and plating operations.

[0018] The present invention also discloses a dynamic monitoring system for catering source emissions, including: a feature extraction module, used to obtain catering enterprise information and video images in a region, and input the enterprise information and video images into a deep learning model for feature extraction; an emission factor selection module, used to select the emission factor corresponding to the emission model according to the enterprise information; a curve graph generation module, used to generate a curve graph of the feature changes over time according to the feature; an initial emission model generation module, used to generate an initial emission model according to the emission factor and the curve graph of the feature changes over time; a model training module, used to train the initial emission model according to the video image to obtain a final emission model; an output module, used to input the enterprise information and video images to be tested into the final emission model to obtain the emission results of the enterprise to be tested.

[0019] The present invention also discloses a computer-readable storage medium, on which a computer program is stored. The computer program is executed by a processor to implement the method for dynamic monitoring of emissions from a catering source as described in any one of the above items.

[0020] The present invention adopts the above technical solution, which has the following advantages:

[0021] The solution in the present invention realizes dynamic monitoring of catering pollutant emissions through video analysis technology. The pollutant emission monitoring method established by the algorithm is low-cost, high-precision and easy to maintain compared with traditional monitoring methods. It breaks through the blind spots of environmental managers and improves management efficiency and effectiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a flow chart of a method for dynamic monitoring of emissions from catering sources in one embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the technical direction of the present invention, the present invention is described in detail through specific implementation examples. However, it should be understood that the provision of specific implementation examples is only for a better understanding of the present invention, and they should not be understood as limitations of the present invention. In the description of the present invention, it should be understood that the terms used are only for the purpose of description and cannot be understood as indicating or implying relative importance.

[0024] Example

[0025] This embodiment discloses a method for dynamic monitoring of emissions from catering sources, such as Figure 1 As shown, including:

[0026] Obtain catering business information and video images in the area, and input the business information and video images into the video analysis algorithm for feature extraction.

[0027] Obtain catering enterprise information and video images in the area, input the video images of the enterprise's cooking activities into the video analysis algorithm for feature extraction; select the emission factor corresponding to the emission model based on the enterprise information; generate a graph of the number of stoves used over time based on the features extracted from the video images of the stoves during cooking activities; generate an initial emission model based on the emission factor and the graph of the number of stoves used over time; use video images to train the initial emission model to obtain the final emission model; input the enterprise information and video images to the final emission model to obtain the emission results of the enterprise to be tested.

[0028] Based on the video images of the cooking positions of the catering enterprises, the video analysis algorithm is used to extract static features from the video images of the cooking activities, and the identification objects are coded and marked. The extracted features include: stoves, pots, oil, spatulas, cutting boards, seasonings, etc. At the same time, dynamic features are extracted from the video images of the cooking activities to obtain information about the chef's actions. The extracted features are classified and marked (including operations such as preparing dishes, pouring oil, stir-frying, taking out of the pot, and plating), and the collected sample images are trained using a deep learning model. The time from pouring oil to the pot leaving the stove is marked as the cooking activity time, and the stove is in use. The use time of each stove is obtained by supervised learning. After a large number of sample training, it can reach the level of intelligent recognition of the stove use time. Collect basic information of catering enterprises; catering enterprise information includes stove data, working time data and auxiliary information. Catering enterprise information specifically includes: enterprise scale, enterprise cuisine type, number of stoves, stove type, enterprise name, enterprise type, fuel information, cooking oil information, business hours and customer flow information.

[0029] According to the enterprise information, select the emission factor corresponding to the emission model.

[0030] Based on the characteristics, a curve graph showing the change in stove usage over time is generated.

[0031] An initial emission model is generated based on the emission factor and characteristic change curve over time; the time variation pattern of catering emissions is obtained through the emission model.

[0032] The initial emission model is trained using video images to obtain the final emission model, calculate the real-time dynamic emission inventory of catering enterprises, simulate the actual pollutant concentration level of catering enterprises through the emission model, and compare and verify with the actual monitoring results. Specifically, the method of inputting the video image of the enterprise's cooking activities into the video analysis algorithm for feature extraction is: extract each frame of the video image of the enterprise's cooking activities; extract the static features of each frame of the image, determine the type of the image according to the static features, and encode and mark it; according to the type of image, extract dynamic features of the video image of the cooking activity, collect continuous frames in the video sequence for 3D convolution, and obtain information about the chef's actions; determine the use time of the stove, combine the stove use time, static features and dynamic features, and obtain the final emission model. In this embodiment, the static features include: stove, pot, oil, spatula, chopping board and seasoning; the dynamic features include preparing dishes, pouring oil, stir-frying, taking out of the pot and plating.

[0033] The method for determining the stove usage time is as follows: use a deep learning model to perform feature training on video images, mark the time from pouring oil to the pot leaving the stove as the known stove usage time; input the known stove usage time obtained by the deep learning algorithm into the supervised learning algorithm model, train the supervised learning algorithm model, and obtain the optimal supervised learning model; input the video to be tested into the optimal supervised learning model to obtain its corresponding stove usage time.

[0034] Combining the stove usage time, static characteristics and dynamic characteristics, the emission model formula is obtained as follows:

[0035] E i =EF i ×n×t

[0036] Among them, E i is the emission of a certain pollutant i from the catering source, EF i is the emission factor of the pollutant, n is the number of working stoves, and t is the working time of the stoves.

[0037] The pollutant emission concentration of the monitored catering enterprise obtained based on the emission of a certain pollutant i from the catering source is:

[0038]

[0039] Among them, Ci is the emission concentration of a certain pollutant i from the catering source, Ei is the emission of a certain pollutant i from the catering source, t is the emission time, V is the flue gas flow rate, and η is the purification efficiency of the purifier for the pollutant.

[0040] The information of the enterprise to be tested and the video images are input into the final emission model to obtain the emission results of the enterprise to be tested.

[0041] Embodiment 2

[0042] Based on the same inventive concept, this embodiment discloses a dynamic monitoring system for catering source emissions, including:

[0043] The feature extraction module is used to obtain the catering enterprise information and video images in the area, and input the enterprise information and video images into the deep learning model for feature extraction; the catering enterprise information specifically includes: enterprise scale, enterprise cuisine type, number of stoves, stove type, enterprise name, enterprise type, fuel information, cooking oil information, business hours and customer flow information.

[0044] Catering enterprise information includes stove data, working time data and auxiliary information.

[0045] The emission factor selection module is used to select the emission factor corresponding to the emission model based on enterprise information;

[0046] A curve graph generation module is used to generate a curve graph of feature changes over time based on the features;

[0047] An initial emission model generation module is used to generate an initial emission model based on emission factors and characteristic change curves over time;

[0048] A model training module is used to train the initial emission model according to the video images to obtain the final emission model;

[0049] The output module is used to input the information of the enterprise to be tested and the video image into the final emission model to obtain the emission results of the enterprise to be tested.

[0050] Embodiment 3

[0051] Based on the same inventive concept, the present embodiment discloses a computer-readable storage medium on which a computer program is stored. The computer program is executed by a processor to implement any of the above-mentioned methods for dynamic monitoring of emissions from catering sources.

[0052] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0053] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0054] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0055] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0056] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit them. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation of the present invention can still be modified or replaced by equivalents, and any modification or equivalent replacement that does not deviate from the spirit and scope of the present invention should be included in the protection scope of the claims of the present invention. The above content is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or replacements within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A dynamic monitoring method for catering source emissions based on video intelligent analysis, characterized in that: include: Acquire information and video images of catering enterprises in the area, and input the video images of the enterprises' cooking activities into a video analysis algorithm for feature extraction; Select the emission factor corresponding to the emission model based on the enterprise information; Generate a curve chart of the usage of the number of stoves over time based on the features extracted from the video images of the stoves during the cooking activity; Generate an initial emission model based on the emission factor and the curve of the number of stoves used over time; Using video images to train the initial emission model to obtain a final emission model; Input the enterprise information and video images to be tested into the final emission model to obtain the emission results of the enterprise to be tested; The method of inputting the video image of the enterprise cooking activity into the video analysis algorithm for feature extraction is: Extract each frame of the video image of the enterprise's cooking activities; Extracting static features of each frame of the image, determining the type of the image according to the static features, and coding and marking the image; According to the type of the image, dynamic features are extracted from the video image of the cooking activity, and continuous frames are collected in the video sequence for 3D convolution to obtain information about the chef's actions; The stove usage time is determined, and the stove usage time, static characteristics and dynamic characteristics are combined to obtain a final emission model.

2. The method for dynamic monitoring of catering source emissions based on video intelligent analysis according to claim 1, characterized in that: The catering enterprise information includes stove data, working time data and auxiliary information.

3. The method for dynamic monitoring of catering source emissions based on video intelligent analysis according to claim 2, characterized in that: The catering enterprise information specifically includes: Enterprise scale, enterprise cuisine type, number of stoves, stove type, enterprise name, enterprise type, fuel information, cooking oil information, business hours and customer flow information.

4. The method for dynamic monitoring of catering source emissions based on video intelligent analysis according to claim 1, characterized in that: The method for determining the stove usage time is: Perform feature training on the video image using a deep learning model, marking the time from pouring oil to the pot leaving the stove as the known stove usage time; The known stove usage time obtained by the deep learning algorithm is input into the supervised learning algorithm model, and the supervised learning algorithm model is trained to obtain the optimal supervised learning model.

5. The method for dynamic monitoring of catering source emissions based on video intelligent analysis according to claim 1, characterized in that: Combining the stove usage time, static characteristics and dynamic characteristics, the emission model formula is obtained as follows: Among them, E i is the emission of a certain pollutant i from the catering source, EF i is the emission factor of the pollutant, n is the number of working stoves, and t is the working time of the stoves.

6. The method for dynamic monitoring of catering source emissions based on video intelligent analysis according to claim 5, characterized in that: The pollutant emission concentration of the monitored catering enterprise obtained based on the emission of a certain pollutant i from the catering source is: Among them, Ci is the emission concentration of a certain pollutant i from the catering source, E i is the emission of a certain pollutant i from the catering source, t is the emission time, V is the flue gas flow rate, and η is the purification efficiency of the purifier for the pollutant.

7. The method for dynamic monitoring of catering source emissions based on video intelligent analysis according to claim 3, characterized in that: The static features include: stove, pots, oil, spatula, chopping board and seasoning; the dynamic features include preparing food, pouring oil, stir-frying, serving and plating operations.

8. A dynamic monitoring system for catering source emissions based on video intelligent analysis, characterized in that: include: A feature extraction module is used to obtain information and video images of catering enterprises in the area, and input the video images of the enterprise's cooking activities into the video analysis algorithm for feature extraction; The emission factor selection module is used to select the emission factor corresponding to the emission model based on enterprise information; A curve graph generating module, used for generating a curve graph of the feature changing over time according to the feature; An initial emission model generation module, used to generate an initial emission model according to the emission factor and the characteristic change curve over time; A model training module, used for training the initial emission model according to the video image to obtain a final emission model; The output module is used to input the enterprise information and video images to the final emission model to obtain the emission results of the enterprise to be tested; The method of inputting the video image of the enterprise cooking activity into the video analysis algorithm for feature extraction is: Extract each frame of the video image of the enterprise's cooking activities; Extracting static features of each frame of the image, determining the type of the image according to the static features, and coding and marking the image; According to the type of the image, dynamic features are extracted from the video image of the cooking activity, and continuous frames are collected in the video sequence for 3D convolution to obtain information about the chef's actions; The stove usage time is determined, and the stove usage time, static characteristics and dynamic characteristics are combined to obtain a final emission model.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method for dynamic monitoring of catering source emissions based on video intelligent analysis as described in any one of claims 1 to 7.

Images