Non-intrusive load monitoring method based on color-coded VI trajectory images

Abstract

The present invention provides a non-invasive load monitoring method based on color-coded V-I trajectory images. Based on the V-I trajectory images constructed by real-time voltage and current value data, the present invention calculates the size of the real-time current value, the number of corresponding current data points contained in the unit pixel grid in the V-I trajectory image and its current amplitude (maximum value), and maps the three types of values ​​obtained by calculation to the three primary colors of "R", "G", and "B" respectively, constructs an RGB three-channel color space, and performs color coding on the V-I trajectory image, which significantly enhances the expression ability of load characteristics. The feature differentiation of different load types is significantly improved. The present invention has achieved high accuracy and good generalization ability on multiple public non-invasive power load monitoring data sets (PLAID and WHITED), which is superior to the existing most advanced methods and has broad application prospects.

Description

Technical Field

[0001] The present invention relates to the technical field of non-intrusive load monitoring (NILM), and in particular to a non-intrusive load monitoring method based on RGB color-coded voltage-current (VI) trajectory images, which is used to improve the accuracy and generalization ability of power load classification. Background Art

[0002] Energy efficiency improvement and effective energy management are crucial. Non-intrusive load monitoring technology analyzes the voltage and current signals in the bus circuit and uses feature extraction and machine learning algorithms to identify and monitor individual appliances. However, the power load characteristics of different households vary significantly, which poses a challenge to building a classification model with strong generalization capabilities.

[0003] Although the existing VI trajectory images can capture the electrical characteristics of voltage and current and show geometric features similar to the original VI trajectory in a more aggregated form, this representation is still insufficient for subsequent classification tasks. In order to improve feature representation and enhance classification performance, a new method is urgently needed to enhance the feature expression ability of VI trajectory images to improve non-invasive load monitoring capabilities. Summary of the invention

[0004] The purpose of the present invention is to address the deficiencies of the prior art, to construct a VI trajectory image based on real-time voltage and current value data, to calculate the size of the real-time current value, the number of corresponding current data points contained in a unit pixel grid in the VI trajectory image and its current amplitude (maximum value), to respectively map the three types of values ​​obtained by calculation to the three primary colors of "R", "G" and "B", to construct an RGB three-channel color space, to color-code the VI trajectory image, to enhance the expression of load characteristics, and thereby to improve the accuracy of load classification.

[0005] The technical solution adopted by the present invention is specifically as follows:

[0006] A non-intrusive load monitoring method based on color-coded VI trajectory images, comprising:

[0007] Real-time acquisition of steady-state voltage and current signals of the target power system;

[0008] Normalize the collected steady-state voltage and current data to construct a VI trajectory image;

[0009] The constructed VI trace image is color-coded to generate an RGB color-coded VI trace image; wherein:

[0010] For each pixel grid unit in the VI trajectory image, the average value of the data points in the pixel grid unit is calculated and normalized as the red channel color coding value of the corresponding pixel grid unit;

[0011] For each pixel grid unit in the VI trajectory image, the number of data points in the pixel grid unit is calculated and normalized as the green channel color coding value of the corresponding pixel grid unit;

[0012] For each pixel grid unit in the VI trajectory image, the current amplitude interval and the blue channel color intensity interval are determined according to the maximum current value of the pixel grid unit, and then mapped to the corresponding blue channel color intensity interval as the blue channel color coding value of the corresponding pixel grid unit according to the ratio of the pixel grid unit; wherein the ratio of the pixel grid unit is the average value of the data points in the pixel grid unit and the normalized value; the current amplitude interval is determined according to the maximum steady-state current values ​​of several load categories in ascending order; the blue channel color intensity interval is obtained by dividing the blue channel color intensity range according to the number of load categories; the current amplitude interval and the blue channel color intensity interval correspond one to one;

[0013] Load classification is performed based on the generated RGB color-coded VI trajectory images to complete non-intrusive load monitoring.

[0014] Furthermore, the constructing of the VI trajectory image specifically includes:

[0015] The steady-state voltage v and current i signal time series data of the target power system collected in real time are scaled to their respective maximum absolute values, normalized, and a grid with a uniform unit size is created on a two-dimensional plane based on the normalized values;

[0016] Initialize a 2N×2N all-zero matrix;

[0017] For each data point , calculate the cell coordinates as follows:

[0018]

[0019] If satisfied and , then set the pixel value of the corresponding grid unit , forming the corresponding binary VI trajectory image.

[0020] Furthermore, each current amplitude interval is expressed as follows:

[0021] ;

[0022] In the formula, Im represents the mth current amplitude interval, M represents the number of load categories, It represents the mth maximum steady-state current value after several load categories are sorted from small to large according to their maximum steady-state current values.

[0023] Furthermore, each blue channel color intensity interval is expressed as follows:

[0024] ;

[0025] Where M represents the number of load categories, B m Represents the mth blue channel color intensity interval.

[0026] Furthermore, load classification is performed based on the generated RGB color-coded VI trajectory image as follows:

[0027] The generated RGB color-coded VI trajectory image is input into a pre-trained load classification model, and the classification result is obtained by output;

[0028] Wherein, the pre-trained load classification model is obtained based on data set pre-training.

[0029] Furthermore, each sample of the data set includes an RGB color-coded VI trajectory image and a corresponding load label.

[0030] Furthermore, the structure of the load classification model is ResNet18, ResNet34, ResNet50, ResNet101, ResNet152, AlexNet, GoogleNet, VGG16 or VGG19.

[0031] A non-intrusive load monitoring system based on a color-coded VI trajectory image, used to implement the non-intrusive load monitoring method based on a color-coded VI trajectory image, comprising:

[0032] A data acquisition module, used to collect steady-state voltage and current signals of the target power system;

[0033] The data preprocessing module is used to normalize the collected steady-state voltage and current data and construct a VI trajectory image;

[0034] A color coding module, used for color coding the constructed VI trajectory image to generate an RGB color-coded VI trajectory image;

[0035] The classification module is used to classify the load based on the generated RGB color-coded VI trajectory image to complete non-intrusive load monitoring.

[0036] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, a non-intrusive load monitoring method based on a color-coded VI trajectory image is implemented.

[0037] A storage medium containing computer executable instructions, which, when executed by a computer processor, implement a non-intrusive load monitoring method based on a color-coded VI trajectory image.

[0038] The method of the present invention integrates numerical features, statistical features and current amplitude information while maintaining the characteristics of the original data, significantly improving the feature distinction of different load types and enhancing the accuracy and generalization ability of load classification in non-invasive load monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of a VI trajectory image processing method based on RGB color coding of the present invention;

[0040] Figure 2 It is a VI trajectory image constructed based on steady-state voltage and current, where the horizontal axis is current i and the vertical axis is voltage v;

[0041] Figure 3 It is a schematic diagram of the red intensity matrix after the red channel is encoded, where the abscissa is the current i and the ordinate is the voltage v;

[0042] Figure 4 It is a schematic diagram of the green intensity matrix after green channel encoding, where the horizontal axis is the current i and the vertical axis is the voltage v;

[0043] Figure 5 It is a schematic diagram of the blue intensity matrix after blue channel encoding, where the horizontal axis is the current i and the vertical axis is the voltage v;

[0044] Figure 6 is a schematic diagram of the generated RGB color-coded VI trajectory image, where the abscissa is the current i and the ordinate is the voltage v;

[0045] Figure 7 It is a non-intrusive load monitoring system structure diagram based on color-coded VI trajectory images. DETAILED DESCRIPTION

[0046] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0047] The terminology used in the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application.

[0048] As used in this application and the appended claims, the singular forms "a", "an", "said", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0049] It should be understood that although the terms first, second, third, etc. may be used in the present application to describe various information, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0050] The present invention provides a non-intrusive load monitoring method based on color-coded VI trajectory images, such as Figure 1 As shown, the following steps are included:

[0051] Step 1: Collect steady-state voltage and current signals of the target power system in real time;

[0052] Step 2: Normalize the collected steady-state voltage and current data to construct a VI trajectory image;

[0053] The VI trace image captures the electrical characteristics of the currently acquired steady-state voltage and current data, maintains the consistency of feature scales, and displays similar geometric features to the original VI trace in a more aggregated form; in a specific implementation, the construction method is as follows:

[0054] 1) The steady-state voltage v and current i signal time series data of the target power system collected in real time are scaled to their respective maximum absolute values, normalized, and a grid with uniform cell size is created on a two-dimensional plane based on the normalized values. The length of each cell is and width The calculation is as follows:

[0055]

[0056] in, represents the maximum value operation, represents the absolute value, Indicates the number of cells in the mesh. The result is a elements of a matrix.

[0057] 2) Initialize a size of The all-zero matrix .

[0058] 3) For each data point , calculate the cell coordinates as follows:

[0059]

[0060] Indicates rounding down; if and , then set the coordinates to the cell pixel value .

[0061] This method converts the normalized steady-state voltage and current data into discrete coordinates through grid mapping and forms the corresponding binary trajectory image. The construction result is as follows: Figure 2 shown.

[0062] Step 3: Color-code the constructed VI trajectory image to generate an RGB color-coded VI trajectory image; wherein:

[0063] (1) For each pixel grid unit in the VI trajectory image, the average value of the data points in the pixel grid unit is calculated and normalized as the red channel color coding value of the corresponding pixel grid unit; in a specific embodiment, the steps are as follows:

[0064] (1.1) For each pixel grid cell (i, j) in the VI trajectory image matrix X, calculate the average value of all data points in the cell :

[0065]

[0066] Among them, x k is the steady-state current data value of the kth data point in the unit, and n is the number of data points in the unit. The average values ​​calculated by all units constitute the average value matrix .

[0067] (1.2) Normalization processing:

[0068] The mean matrix Normalize to the range [0,1] to get the matrix ,matrix Each unit of is represented as follows:

[0069]

[0070] , Represents the mean value matrix The minimum and maximum values ​​of .

[0071] (1.3) Mapping to color intensity:

[0072] The normalized matrix The red channel intensity matrix R is mapped to [0,255] to achieve color representation from light red to dark red, and to intuitively distinguish low-value and high-value data points, such as Figure 3 As shown; each unit of the red channel intensity matrix R is expressed as follows:

[0073]

[0074] (2) For each pixel grid unit in the VI trajectory image, the number of data points in the pixel grid unit is calculated and normalized as the green channel color coding value of the corresponding pixel grid unit; in a specific implementation scheme, the steps are as follows:

[0075] (2.1) Statistical frequency:

[0076] Calculate the number of data points (frequency of occurrence) C in each pixel grid cell (i, j) in the VI trajectory image ij , forming a frequency matrix C. Specifically, according to the formula for calculating the cell coordinates in step 2, each data point is set based on the corresponding coordinates of the ratio of its own current and voltage to the length and width of the cell, which is rounded down. Then, close data points will be assigned to the same cell coordinates, that is, a cell will contain multiple data points. The coordinates of each cell calculated by each data point when constructing the VI trajectory image can be reversed to obtain the number of data points in each pixel grid unit.

[0077] (2.2) Normalization processing:

[0078] Normalize the frequency matrix C to the range [0,1] to get the matrix ,matrix Each unit of is represented as follows:

[0079]

[0080] , Respectively represent the minimum and maximum values ​​of the frequency matrix C.

[0081] (2.3) Mapped to color intensity:

[0082] The normalized matrix The green channel intensity matrix G mapped to [0,255] uses a color gradient from light green to dark green to represent the density distribution of data points in the grid, such as Figure 4 As shown; each unit of the green channel intensity matrix G is expressed as follows:

[0083]

[0084] (3) For each pixel grid unit in the VI trajectory image, determine the current amplitude interval and the blue channel color intensity interval according to the maximum current value of the pixel grid unit, and then map the current amplitude interval and the blue channel color intensity interval to the corresponding blue channel color intensity interval as the blue channel color coding value of the corresponding pixel grid unit according to the ratio of the pixel grid unit; wherein the ratio of the pixel grid unit is the average value of the data points in the pixel grid unit and the normalized value; the current amplitude interval is determined according to the maximum steady-state current values ​​of several load categories in ascending order; the blue channel color intensity interval is obtained by dividing the blue channel color intensity range according to the number of load categories; the current amplitude interval and the blue channel color intensity interval correspond to each other one by one; in a specific implementation scheme, the steps are as follows:

[0085] (3.1) Calculate the maximum steady-state current of several collected load categories and sort them; the number of load categories is determined based on the actual collection situation, and the more load categories, the better:

[0086]

[0087] (3.2) Divide the current amplitude interval:

[0088] According to the sorting results, the current amplitude range is divided into M intervals , each current amplitude interval I m Defined as:

[0089]

[0090] in, .

[0091] (3.3) Divide the blue channel color intensity interval:

[0092] Divide the blue channel color intensity range [0,255] into M discrete intervals , each B m The interval corresponds to a current amplitude interval I m , defined as:

[0093]

[0094] (3.4) Mapping current amplitude to color intensity:

[0095] For each pixel grid unit in the VI trajectory image, according to the maximum current value of the pixel grid unit The current amplitude range I m , and find the corresponding blue channel color intensity interval B according to the corresponding relationship m Specifically, as mentioned above, each pixel grid unit (i, j) contains multiple data points. The data points in the pixel grid unit are sorted according to the current value, and the current value corresponding to the data point with the largest current value is taken as the maximum current value of the pixel grid unit; according to the size of the maximum current value, the current amplitude interval I to which it belongs is determined. m For example, the maximum current value is 5, and the third current amplitude interval is I 3 The range of is 4-6, then the current amplitude interval to which the maximum current value of the current pixel grid unit belongs is I 3 Finally, the corresponding blue channel color intensity interval B is found according to the corresponding relationship. 3 .

[0096] (3.5) Calculate the blue channel intensity matrix:

[0097] The average value of each pixel grid cell (i, j) in the VI trajectory image matrix X is normalized to As a ratio, the blue channel intensity matrix B is calculated, and the color gradient from light blue to dark blue is used to represent the current amplitude characteristic distribution of the data points in the grid, such as Figure 5 As shown, each unit of the blue channel intensity matrix B is expressed as follows:

[0098]

[0099] in, and are the lower and upper limits of the color intensity range of the mth blue channel, respectively.

[0100] Combine the matrices R, G, and B according to the RGB channels to form a color image matrix I, and obtain the RGB color-coded VI trajectory image, such as Figure 6 As shown; each unit of the color image matrix I is expressed as follows:

[0101]

[0102] In the color coding of the present invention, the red channel coding reflects the data value characteristics, the green channel coding represents the data frequency characteristics, and the blue channel coding reflects the current amplitude characteristics. Among them, the red channel calculates the average value of the data points of each pixel unit and maps it to the red intensity after normalization, so as to realize the intuitive distinction of the data value; the green channel calculates the number of statistical points and maps it to the green intensity after normalization to show the density distribution of the data points; the blue channel calculates the maximum steady-state current value of each load category and sorts it, divides the current amplitude interval and maps it to the corresponding blue intensity interval, which can significantly enhance the expression of the current amplitude characteristics. This color coding method not only retains the numerical and statistical characteristics of the original data, but also integrates the current amplitude information, significantly improves the feature distinction of different load types, and thus can improve the non-invasive load monitoring effect.

[0103] Step 4: Perform load classification based on the generated RGB color-coded VI trajectory image to complete non-intrusive load monitoring. In a specific implementation scheme, the details are as follows:

[0104] The generated RGB color-coded VI trajectory image is input into a pre-trained load classification model, and the classification result is obtained by output;

[0105] The pre-trained load classification model is obtained by pre-training based on the data set with the goal of minimizing the error between the classification result output by the model and the true value; each sample of the data set includes an RGB color-coded VI trajectory image constructed based on the steady-state voltage and current signals of actual load operation and a corresponding load label;

[0106] The structures of the load classification models are ResNet18, ResNet34, ResNet50, ResNet101, ResNet152, AlexNet, GoogleNet, VGG16, VGG19, etc.

[0107] In order to make the purpose, technical solution and advantages of the present invention clearer, the effects of the present invention are described in detail below in combination with PLAID, WHITED datasets and specific implementation cases.

[0108] The PLAID and WHITED datasets were used respectively, and the training set and the test set were divided by 10-fold verification. The division ratios of the training set and the test set in each fold were 0.9 and 0.1, respectively. The ResNet18 deep model was combined as the load classification model, and the color-coded VI trajectory image constructed by the present invention was used as input, and the classification result was used as output. The pre-training was performed by minimizing the loss function (cross entropy loss function), and the pre-trained ResNet18 deep model was further tested by processing the test set using the color-coded VI trajectory image method constructed by the present invention. The results are shown in Table 1. The generated RGB color-coded VI trajectory image can more comprehensively express the electrical characteristics of the load, including the size of the real-time current value, the number of corresponding current data points contained in the unit pixel grid in the VI trajectory diagram, and its current amplitude information (maximum value). Thereby, the accuracy of load classification can be significantly improved and the generalization ability of the model can be enhanced. At the same time, the method does not require an additional complex model structure or an additional loss function, and has the advantages of high computational efficiency and wide applicability. Experimental results show that this method has achieved excellent classification performance on multiple public non-intrusive power load monitoring datasets (PLAID and WHITED), and has high practical value.

[0109] Table 1

[0110]

[0111] In Table 1, precision reflects the accuracy of the model in predicting positive examples, recall reflects the model's ability to find all positive examples, and the F score is a comprehensive indicator used to evaluate the performance of the classification model. It is calculated based on precision and recall, as shown in the following formula. The higher the score, the better the performance of the model in the load classification task.

[0112]

[0113] Corresponding to the aforementioned embodiment of a non-intrusive load monitoring method based on a color-coded VI trajectory image, the present invention also provides an embodiment of a non-intrusive load monitoring system based on a color-coded VI trajectory image.

[0114] See also Figure 7 , a non-intrusive load monitoring system based on a color-coded VI trajectory image is provided in an embodiment of the present invention, comprising:

[0115] The data acquisition module is used to collect steady-state voltage and current signals of the target power system; it generally includes a voltage transformer and a current transformer.

[0116] A data preprocessing module is connected to the data acquisition module and is used to normalize the collected steady-state voltage and current data and construct a VI trajectory image;

[0117] A color coding module, connected to the data preprocessing module, includes three channel coding submodules of R, G, and B, which are respectively used to color code the constructed VI trajectory image to generate an RGB color-coded VI trajectory image;

[0118] The classification module is connected with the color coding module and is used to classify the load based on the generated RGB color-coded VI trajectory image to complete non-intrusive load monitoring. Figure 7 It is shown in the figure that the classification module uses a pre-trained load classification model to perform load classification.

[0119] The embodiment of the non-intrusive load monitoring system based on color-coded VI trajectory images of the present invention can be applied to any device with data processing capability, and the device with data processing capability can be a device or apparatus such as a computer.

[0120] The present invention also provides an electronic device, comprising one or more processors, for implementing a non-intrusive load monitoring method based on a color-coded VI trajectory image in the above embodiment.

[0121] The device embodiment can be implemented by software, hardware or a combination of software and hardware. Taking software implementation as an example, as a device in a logical sense, it is formed by the processor of any device with data processing capability in which it is located to read the corresponding computer program instructions in the non-volatile memory into the memory for execution. From the hardware level, the electronic device includes a processor, memory, network interface, and non-volatile memory. In addition, any device with data processing capability in which the device in the embodiment is located can also include other hardware according to the actual function of the device with data processing capability, which will not be described in detail.

[0122] The implementation process of the functions and effects of each unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0123] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of the present invention. Ordinary technicians in this field can understand and implement it without paying creative work.

[0124] An embodiment of the present invention further provides a computer-readable storage medium having a program stored thereon. When the program is executed by a processor, a non-invasive load monitoring method based on a color-coded VI trajectory image in the above embodiment is implemented.

[0125] The computer-readable storage medium may be an internal storage unit of any device with data processing capability described in any of the aforementioned embodiments, such as a hard disk or a memory. The computer-readable storage medium may also be any device with data processing capability, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), an SD card, a flash card (Flash Card), etc. equipped on the device. Furthermore, the computer-readable storage medium may also include both an internal storage unit of any device with data processing capability and an external storage device. The computer-readable storage medium is used to store the computer program and other programs and data required by any device with data processing capability, and may also be used to temporarily store data that has been output or is to be output.

[0126] Obviously, the above embodiments are merely examples for the purpose of clear explanation, and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived therefrom are still within the scope of protection of the present invention.

Claims

1. A non-intrusive load monitoring method based on color-coded VI trajectory images, characterized in that: include: Real-time acquisition of steady-state voltage and current signals of the target power system; Normalize the collected steady-state voltage and current data to construct a VI trajectory image; The constructed VI trace image is color-coded to generate an RGB color-coded VI trace image; wherein: For each pixel grid unit in the VI trajectory image, the average value of the data points in the pixel grid unit is calculated and normalized as the red channel color coding value of the corresponding pixel grid unit; For each pixel grid unit in the VI trajectory image, the number of data points in the pixel grid unit is calculated and normalized as the green channel color coding value of the corresponding pixel grid unit; For each pixel grid unit in the VI trajectory image, the current amplitude interval and the blue channel color intensity interval are determined according to the maximum current value of the pixel grid unit, and then mapped to the corresponding blue channel color intensity interval as the blue channel color coding value of the corresponding pixel grid unit according to the ratio of the pixel grid unit; wherein the ratio of the pixel grid unit is the average value of the data points in the pixel grid unit and the normalized value; the current amplitude interval is determined according to the maximum steady-state current values ​​of several load categories in ascending order; the blue channel color intensity interval is obtained by dividing the blue channel color intensity range according to the number of load categories; the current amplitude interval and the blue channel color intensity interval correspond one to one; Load classification is performed based on the generated RGB color-coded VI trajectory images to complete non-intrusive load monitoring.

2. The method according to claim 1, characterized in that The constructing of the VI trajectory image specifically includes: The steady-state voltage v and current i signal time series data of the target power system collected in real time are scaled to their respective maximum absolute values, normalized, and a grid with a uniform unit size is created on a two-dimensional plane based on the normalized values; Initialize a 2N×2N all-zero matrix; For each data point , calculate the cell coordinates as follows: ； If satisfied and , then set the pixel value of the corresponding grid unit , forming the corresponding binary VI trajectory image.

3. The method according to claim 1, characterized in that Each current amplitude interval is expressed as follows: ； In the formula, I m represents the mth current amplitude interval, M represents the number of load categories, It represents the mth maximum steady-state current value after several load categories are sorted from small to large according to their maximum steady-state current values.

4. The method according to claim 1, characterized in that: Each blue channel color intensity interval is represented as follows: ； Where M represents the number of load categories, B m Represents the mth blue channel color intensity interval.

5. The method according to claim 1, characterized in that: Load classification is performed based on the generated RGB color-coded VI trajectory image as follows: The generated RGB color-coded VI trajectory image is input into a pre-trained load classification model, and the classification result is obtained by output; Wherein, the pre-trained load classification model is obtained based on data set pre-training.

6. The method according to claim 5, characterized in that Each sample of the dataset includes an RGB color-coded VI trajectory image and a corresponding load label.

7. The method according to claim 5, characterized in that The structure of the load classification model is ResNet18, ResNet34, ResNet50, ResNet101, ResNet152, AlexNet, GoogleNet, VGG16 or VGG19.

8. A non-intrusive load monitoring system based on color-coded VI trajectory images, used to implement the non-intrusive load monitoring method based on color-coded VI trajectory images as claimed in any one of claims 1 to 7, characterized in that: include: A data acquisition module, used to collect steady-state voltage and current signals of the target power system; The data preprocessing module is used to normalize the collected steady-state voltage and current data and construct a VI trajectory image; A color coding module, used for color coding the constructed VI trajectory image to generate an RGB color-coded VI trajectory image; The classification module is used to classify the load based on the generated RGB color-coded VI trajectory image to complete non-intrusive load monitoring.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the non-invasive load monitoring method based on the color-coded VI trajectory image as described in any one of claims 1 to 7 is implemented.

10. A storage medium comprising computer executable instructions, wherein the computer executable instructions, when executed by a computer processor, implement a non-intrusive load monitoring method based on a color-coded VI trajectory image as claimed in any one of claims 1 to 7.

Images