Data real carbon emission detection method based on multi-modal sensing fusion

Abstract

The invention relates to the technical field of dual-carbon data original acquisition, discloses a data real carbon emission detection method based on multi-modal sensing fusion, and solves the technical problem that the authenticity and the accuracy of a carbon dioxide original data acquisition result in the prior art have defects, and the method comprises the following steps: initializing a system environment; acquiring environment data in real time; calculating a self-fluctuation coefficient according to data collected by adjacent environment monitoring points, and matching a corresponding sampling frequency according to the self-fluctuation coefficient to realize variable-frequency sampling; calculating the average carbon emission; compensating according to the dynamic carbon emission compensation factor; calculating the total amount of carbon emission, and calculating the total amount of carbon emission according to an integral algorithm. According to the technical scheme, the influence of environmental influence factor numerical values on the detection precision of the carbon dioxide sensor is considered, meanwhile, the analysis process of dynamic influences brought by the fluctuation process of the environmental influence factors is added, the comprehensive performance has great improvement significance, and a scientific basis is provided for achieving the dual-carbon target.

Description

Technical Field

[0001] The present invention relates to the technical field of dual-carbon data original collection, and more specifically, to a data real carbon emission detection method based on multi-modal sensor fusion. Background Art

[0002] As the core part of sustainable development, carbon has attracted much attention in terms of scientific measurement indicators and calculation methods.

[0003] With the development of science and technology, the past extensive economic growth model is no longer feasible, and domestic listed companies are actively practicing ESG. According to reports, as of the end of December 2022, a total of 1,472 A-share listed companies disclosed their 2021 ESG reports, an increase of 351 from the previous year, with a disclosure rate of 30.86%. And many companies are still in urgent need of understanding some concepts and issues. For example, is ESG a quantitative upgrade of the CSR (corporate social responsibility) report? What is the relationship between ESG and the concept of "dual carbon"? For domestic companies, can they have a set of ESG indicator systems with localized characteristics in the future? In the long run, whether the disclosure of ESG information by listed companies will become a mandatory requirement is actually a phased requirement of the regulatory authorities, but this requirement is constantly increasing, and may become a trend in the capital market. Therefore, Chinese companies have also begun to change their thinking, accelerate the embrace of ESG, and embark on the road of high-quality development.

[0004] In the process of achieving the dual carbon goals, the authenticity and accuracy of the original data collection of carbon dioxide emissions is particularly important, which will directly affect the authenticity of the measurement of carbon peak and carbon neutrality. The carbon emission detection process of the existing technology usually ignores the factors such as ambient temperature, humidity, and pressure during detection, which affect the accuracy of the concentration detection results of the carbon detection sensor, and thus affect the accuracy of the emission calculation results. For example: Temperature factors: The sensitivity, response speed, and drift characteristics of the sensor may change with temperature. Humidity factors: Water vapor may interfere with the spectral signal of gas detection or cause unstable response of electrochemical sensors. Pressure factors: Fluctuations in ambient pressure will affect the calculation formula of gas concentration, because gas concentration is usually expressed in volume ratio (ppm or ppb), which is closely related to pressure.

[0005] Therefore, the detection process of the prior art ignores the influence of relevant factors, resulting in defects in the authenticity and accuracy of the original data collection results. Summary of the invention

[0006] In response to the technical problem that the authenticity and accuracy of the carbon dioxide raw data collection results are flawed as raised in the background technology, the present invention not only considers the impact of the numerical values ​​of environmental influencing factors on the detection accuracy of the carbon dioxide sensor, but also adds an analysis process for the dynamic impact brought about by the fluctuation process of environmental influencing factors. It has significant improvement significance in comprehensive performance and provides a scientific basis for achieving the dual carbon goals.

[0007] To achieve the above object, the present invention provides the following technical solutions: A method for detecting real carbon emissions based on multimodal sensor fusion data, comprising the following steps: The first step is to initialize the system environment. The initialization content includes defining the state, defining the compensation, and defining the action. The defined state refers to the initial sampling frequency of the predefined CO2 sensor; Defining compensation means predefining a compensation scheme for the concentration sampling result of the carbon dioxide sensor; Defining an action means predefining an action plan for adjusting the sampling frequency of the carbon dioxide sensor based on the self-fluctuation coefficient; Step 2: Collect data in real time; Use multi-modal sensor fusion sampling system to detect environmental parameters in real time; The third step is frequency conversion sampling; According to the environmental factor data collected by the multimodal sensor fusion sampling system at adjacent monitoring sampling points, the self-fluctuation coefficient is calculated, and the predefined carbon dioxide sensor sampling frequency is automatically matched according to the self-fluctuation coefficient; Step 4: Calculate the average carbon emissions. According to the variable frequency sampling results of the carbon dioxide sensor, use the average algorithm or weighted algorithm to calculate the average carbon dioxide concentration in the sampling interval, and then calculate the average carbon emissions. Step 5: Compensation: Compensate each average carbon emission amount according to the dynamic carbon emission compensation factor in the predefined compensation scheme; Step 6: Calculate the total carbon emissions; calculate the total carbon emissions based on the integral algorithm.

[0008] Through the above technical scheme, the advantage of the detection method of the present invention over the prior art is that, in the process of the carbon dioxide sensor collecting data on carbon emission indicators, the influence of environmental factors is taken into account. The present invention considers the influence of the numerical value of environmental influencing factors on the detection accuracy of the carbon dioxide sensor, and at the same time adds the dynamic influence brought by the fluctuation process of environmental influencing factors, which greatly improves the detection accuracy of the carbon dioxide sensor. At the same time, the variable frequency sampling scheme also reduces the system function to the extreme on the basis of ensuring the detection accuracy, and has significant improvement significance in comprehensive performance, thereby improving the authenticity and accuracy of the original data collection results, and providing a scientific basis for achieving the dual carbon goals.

[0009] The present invention further: wherein the action scheme for adjusting the sampling frequency of the carbon dioxide sensor based on the self-fluctuation coefficient is: setting a preset self-fluctuation coefficient threshold, and comparing the self-fluctuation coefficient with the preset self-fluctuation coefficient threshold; When the self-fluctuation coefficient is less than or equal to a preset self-fluctuation coefficient threshold, the sampling frequency of the carbon dioxide sensor is controlled to decrease; When the self-fluctuation coefficient is greater than a preset self-fluctuation coefficient threshold, the sampling frequency of the carbon dioxide sensor is controlled to increase.

[0010] The present invention further provides that: the predefined compensation scheme is: optimizing the average carbon emission according to a preset data processing strategy.

[0011] The present invention further states: wherein the preset data processing strategy refers to first importing the detection values ​​of each influencing factor into the factor influence model after data cleaning, calculating and outputting the dynamic carbon emission compensation factor through the factor influence model, and then optimizing and compensating the average carbon emissions based on the dynamic carbon emission compensation factor.

[0012] The present invention further: wherein the factor influence model is: The dynamic carbon emission compensation factor is defined as , ,in, is a carbon emission prediction model. It is the benchmark carbon emissions under standard environmental factors.

[0013] In summary, the present invention has the following beneficial effects: (1) The advantage of the detection method of the present invention over the prior art is that the influence of environmental factors is taken into account during the process of the carbon dioxide sensor collecting data on carbon emission indicators. The present invention not only takes into account the influence of the numerical value of environmental factors on the detection accuracy of the carbon dioxide sensor, but also adds an analysis process of the dynamic influence brought about by the fluctuation process of environmental factors, which greatly improves the detection accuracy of the carbon dioxide sensor. At the same time, the variable frequency sampling scheme also greatly reduces the system function on the basis of ensuring the detection accuracy, and has a significant improvement in comprehensive performance, thereby improving the authenticity and accuracy of the original data collection results, and providing a scientific basis for achieving the dual carbon goals; (2) The self-fluctuation coefficient in the data processing process of the detection method of the present invention is used to characterize the fluctuation of environmental factors. For example, when the numerical value of the environmental factor index increases or decreases sharply, high-frequency sampling and weighted algorithm are used to calculate the average carbon emissions to improve the accuracy of the detection results. This dimension is independent of the numerical value of the environmental factor. In order to characterize the impact of the numerical value of the environmental factor on the accuracy of the detection result, the present invention uses a dynamic carbon emission compensation factor to dynamically compensate for the result. The adjustment of the two dimensions provides a scientific basis for the authenticity of the data to achieve the dual carbon goals. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A schematic diagram of a method for detecting real carbon emissions based on multimodal sensor fusion data according to the present invention; Figure 2 This is a flowchart of a method for detecting real carbon emissions based on multimodal sensor fusion data in the present invention. DETAILED DESCRIPTION

[0015] The present invention will be further described in detail below in conjunction with embodiments and drawings, but the embodiments of the present invention are not limited thereto.

[0016] Before introducing the data real carbon emission detection method based on multimodal sensor fusion of the present invention, the hardware configuration required to implement the method is first described. The hardware configuration includes a carbon dioxide sensor, a multimodal sensor fusion sampling system and a server, and the carbon dioxide sensor and the multimodal sensor fusion sampling system are connected to the server through a bus.

[0017] Among them, the carbon dioxide sensor is used to detect the carbon dioxide concentration in the detection space, and then the carbon dioxide emissions are obtained by converting and calculating the volume and time parameters. The multimodal sensor fusion sampling system is configured as a fusion of various types of precision sensors such as temperature sensors, humidity sensors, and pressure sensors. The multimodal sensor fusion sampling system aims to achieve all-round monitoring and sampling of environmental factors that affect the detection accuracy of carbon dioxide sensors.

[0018] The server is used to analyze and process the environmental factor data collected by the multimodal sensor fusion sampling system. The server's analysis and processing flow includes two dimensions. The first dimension is used to analyze the numerical value of the environmental factor data, so as to further compensate and adjust the carbon dioxide sensor concentration sampling results according to the numerical value of the environmental factor, and improve the accuracy of the concentration sampling results; the second dimension is used to analyze the self-fluctuation coefficient of the sampling time points before and after the environmental factor (referring to the numerical change of the environmental factor sampling by the multimodal sensor fusion sampling system at adjacent time points). When the self-fluctuation coefficient is large, it means that the degree of numerical change before and after the environmental factor is large, and the sampling frequency of the carbon dioxide sensor should be controlled to increase. When the self-fluctuation coefficient is small, it means that the degree of numerical change before and after the environmental factor is small, and the sampling frequency of the carbon dioxide sensor should be controlled to decrease. Finally, the final carbon dioxide emissions are calculated based on the variable frequency sampling results of the carbon dioxide sensor (for specific methods on how to calculate the final carbon dioxide emissions from the variable frequency results, please refer to the carbon emission detection method content later). Therefore, based on the analysis content of the second dimension, the server adds the factor consideration of the environmental factor self-fluctuation coefficient, which effectively further improves the authenticity and accuracy of the carbon dioxide emission results. In addition, the use of a variable frequency sampling scheme based on the self-fluctuation coefficient can improve the authenticity and accuracy of the data generation results while minimizing the energy consumption of the carbon emission detection system. The reason for reducing the system energy consumption is that in the variable frequency sampling scheme, the number of sampling times when the self-fluctuation coefficient is relatively small is effectively reduced, thereby reducing the system energy consumption.

[0019] Based on this system, the following content of the present invention discloses in detail a data real carbon emission detection method based on multimodal sensor fusion, combined with Figure 1 and Figure 2 , the detection method comprises the following steps: The first step is to initialize the system environment. The initialization content includes defining the state, defining the compensation, and defining the action. The defined state refers to the initial sampling frequency of the predefined CO2 sensor; Defining compensation means predefining a compensation scheme for the concentration sampling result of the carbon dioxide sensor; Defining an action means predefining an action plan for adjusting the sampling frequency of the carbon dioxide sensor based on the self-fluctuation coefficient; Step 2: Collect data in real time; Use a multi-modal sensor fusion sampling system to regularly detect environmental parameters in real time, such as: environmental temperature parameters, environmental humidity parameters, environmental pressure parameters and other types of environmental parameters; The third step is frequency conversion sampling; According to the environmental factor data collected by the multimodal sensor fusion sampling system at adjacent monitoring sampling points, the self-fluctuation coefficient is calculated, and the exploration-exploitation strategy (such as Σ-greedy) is used to automatically match the predefined carbon dioxide sensor sampling frequency according to the self-fluctuation coefficient; Step 4: Calculate average carbon emissions; According to the variable frequency sampling results of the carbon dioxide sensor, the average carbon dioxide concentration in the sampling interval is calculated using an average algorithm or a weighted algorithm, and then the average carbon emissions are calculated; Step 5: Compensation: Compensate each average carbon emission amount according to the dynamic carbon emission compensation factor in the predefined compensation scheme; Step 6: Calculate the total carbon emissions; calculate the total carbon emissions based on the integral algorithm.

[0020] The predefined action plan for adjusting the sampling frequency of the carbon dioxide sensor based on the self-fluctuation coefficient is specifically as follows: setting a preset self-fluctuation coefficient threshold; When the self-fluctuation coefficient is less than or equal to the preset self-fluctuation coefficient threshold, it means that the environmental factors in this time period fluctuate less than those in the previous time period, and the sampling frequency of the carbon dioxide sensor should be controlled to be reduced to reduce the energy consumption of the system operation; When the self-fluctuation coefficient is greater than the preset self-fluctuation coefficient threshold, it means that the environmental factors in this time period fluctuate greatly compared with the previous time period, and the sampling frequency of the carbon dioxide sensor should be controlled to increase.

[0021] The predefined compensation scheme is specifically as follows: the average carbon emissions are optimized according to the preset data processing strategy to improve authenticity; the preset data processing strategy refers to first importing the detection values ​​of each influencing factor into the factor influence model after data cleaning, and then calculating and outputting the dynamic carbon emission compensation factor through the factor influence model, and then optimizing and compensating the average carbon emissions according to the dynamic carbon emission compensation factor.

[0022] The dynamic carbon emission compensation factor is calculated based on the factor impact model. This factor can dynamically adjust the average carbon emissions according to the changes in factors such as temperature T, humidity H and pressure P. The calculation method of the dynamic carbon emission compensation factor (i.e., the factor impact model) is: based on the basic experiments in the prior art combined with regression analysis, neural networks, support vector (SVN) and other modeling methods to quantify the impact of these factors on the average carbon emissions. Taking the multivariate linear regression analysis model as an example, the carbon emission prediction model based on environmental factors is obtained in this embodiment. , then the dynamic carbon emission compensation factor It can be defined as (in is the benchmark carbon emission under standard environmental factors). Therefore, the dynamic carbon emission compensation factor is calculated through the factor impact model. .

[0023] The following is combined with Figure 1 As a specific example, Figure 1 The figure shows a flow chart of a real carbon emission detection method based on multimodal sensor fusion data within any carbon emission detection cycle. During the detection cycle, the multimodal sensor fusion sampling system evenly collects data in real time, so that the server can calculate the self-fluctuation coefficient based on the environmental factor data collected by the multimodal sensor fusion sampling system at adjacent monitoring sampling points. The multimodal sensor fusion sampling system evenly collects data in real time and divides the detection cycle into 13 variable frequency sampling intervals of the carbon dioxide sensor. In each sampling interval, the sampling times (frequency) of the carbon dioxide sensor are different, thereby realizing variable frequency sampling of the carbon dioxide sensor.

[0024] The advantage of this detection method over the existing technology is that, in the process of the carbon dioxide sensor collecting data on carbon emission indicators, the influence of environmental factors is taken into account. The present invention not only takes into account the influence of the numerical values ​​of environmental factors on the detection accuracy of the carbon dioxide sensor, but also adds an analysis process of the dynamic impact brought about by the fluctuation process of environmental factors, which greatly improves the detection accuracy of the carbon dioxide sensor. At the same time, the variable frequency sampling scheme also reduces the system function to the extreme on the basis of ensuring the detection accuracy, and has significant improvement significance in comprehensive performance, thereby improving the authenticity and accuracy of the original data collection results, and providing a scientific basis for achieving the dual carbon goals.

[0025] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A method for detecting real carbon emissions based on multimodal sensor fusion data, characterized in that: The following steps are involved: The first step is to initialize the system environment. The initialization content includes defining the state, defining the compensation, and defining the action. The defined state refers to the initial sampling frequency of the predefined CO2 sensor; Defining compensation means predefining a compensation scheme for the concentration sampling result of the carbon dioxide sensor; Defining an action means predefining an action plan for adjusting the sampling frequency of the carbon dioxide sensor based on the self-fluctuation coefficient; Step 2: Collect data in real time; Use multi-modal sensor fusion sampling system to detect environmental parameters in real time; The third step is frequency conversion sampling; According to the environmental factor data collected by the multimodal sensor fusion sampling system at adjacent monitoring sampling points, the self-fluctuation coefficient is calculated, and the predefined carbon dioxide sensor sampling frequency is automatically matched according to the self-fluctuation coefficient; Step 4: Calculate the average carbon emissions. According to the variable frequency sampling results of the carbon dioxide sensor, use the average algorithm or weighted algorithm to calculate the average carbon dioxide concentration in the sampling interval, and then calculate the average carbon emissions. Step 5: Compensation: Compensate each average carbon emission amount according to the dynamic carbon emission compensation factor in the predefined compensation scheme; Step 6: Calculate the total carbon emissions; calculate the total carbon emissions based on the integral algorithm.

2. According to claim 1, a method for detecting real carbon emissions based on multimodal sensor fusion data, characterized in that: The action scheme for adjusting the sampling frequency of the carbon dioxide sensor based on the self-fluctuation coefficient is as follows: setting a preset self-fluctuation coefficient threshold, and comparing the self-fluctuation coefficient with the preset self-fluctuation coefficient threshold; When the self-fluctuation coefficient is less than or equal to a preset self-fluctuation coefficient threshold, the sampling frequency of the carbon dioxide sensor is controlled to decrease; When the self-fluctuation coefficient is greater than a preset self-fluctuation coefficient threshold, the sampling frequency of the carbon dioxide sensor is controlled to increase.

3. The method for detecting real carbon emissions based on multimodal sensor fusion data according to claim 1 is characterized by: The predefined compensation scheme is: optimizing the average carbon emissions according to the preset data processing strategy.

4. The method for detecting real carbon emissions based on multimodal sensor fusion data according to claim 3 is characterized by: The preset data processing strategy refers to first importing the detection values ​​of each influencing factor into the factor influence model after data cleaning, calculating and outputting the dynamic carbon emission compensation factor through the factor influence model, and then optimizing and compensating the average carbon emissions based on the dynamic carbon emission compensation factor.

5. The method for detecting real carbon emissions based on multimodal sensor fusion data according to claim 4 is characterized by: The factor impact model is as follows: The dynamic carbon emission compensation factor is defined as: , ,in, is a carbon emission prediction model. It is the benchmark carbon emissions under standard environmental factors.

Images