Method and system for classifying disclosure content of large enterprise environment information images

Abstract

The application provides a large-scale enterprise environment information image disclosure content classification method and system, and relates to the technical field of image disclosure content classification. First, the environment information image to be classified is acquired and pretreated; then, the environment information image classification index is acquired based on the image classification method of the content analysis method and the grounded theory, the standard environment information image under each classification index is acquired as a data set based on the index, and the computer image classification model based on machine learning which is constructed in advance is trained by using the data set; finally, the pretreated environment information image to be classified is classified by using the trained computer image classification model based on machine learning. The application can automatically classify the large-scale environment information image set, and has the characteristics of high efficiency and high accuracy compared with artificial classification.

Description

Technical Field

[0001] This invention relates to the field of image disclosure content classification technology, specifically to a method and system for classifying the disclosure content of large-scale corporate environmental information images. Background Technology

[0002] Environmental information disclosure is an essential part of corporate social responsibility (CSR) information disclosure. It includes not only quantitative data but also subjective information and behavioral descriptions in the form of text and images. Compared to other types of data, image data, due to its strong readability, comprehensibility, and appeal, can effectively accelerate human information processing and influence perception through visual features such as color, objects, and emotions. Therefore, if we can quickly extract and measure the content of environmental images disclosed by a large number of companies, we can understand the disclosure characteristics of corporate environmental information images and thus quickly assess the company's environmental information disclosure tendencies.

[0003] Currently, content analysis of environmental information images still relies on manual image content recognition. This method is extremely time-consuming and labor-intensive, resulting in high costs when dealing with the recognition and classification of large-scale environmental information image content. While machine learning-based computer image classification models have become relatively mature, there are few methods for using these models to recognize and classify large-scale environmental image content, and the classification accuracy cannot be guaranteed.

[0004] In summary, there is currently no effective method in the technology for accurate identification and classification of large-scale environmental image content using computer image classification models. Summary of the Invention

[0005] (a) Technical problems to be solved

[0006] To address the shortcomings of existing technologies, this invention provides a method and system for classifying the content of large-scale enterprise environmental information images, solving the problem of low accuracy in existing computer image classification models for large-scale environmental image content recognition and classification.

[0007] (II) Technical Solution

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] Firstly, this invention proposes a method for classifying the content disclosed in large-scale enterprise environmental information images, the method comprising:

[0010] Acquire an image of the environment information to be classified, and preprocess the image of the environment information to be classified;

[0011] Image classification methods based on content analysis and grounded theory are used to obtain environmental information image classification indicators;

[0012] Based on the environmental information image classification index, standard environmental information images under each classification index are obtained as a dataset, and the dataset is used to train a pre-built computer image classification model based on machine learning.

[0013] The preprocessed environmental information image to be classified is classified based on a trained machine learning-based computer image classification model.

[0014] Preferably, the step of acquiring the environmental information image to be classified and preprocessing the environmental information image to be classified includes:

[0015] S11. Use web crawler programs to obtain large-scale environmental information images;

[0016] S12. Perform preprocessing on the environmental information image, including image data cleaning.

[0017] Preferably, the environmental information image classification indicators include environmental management awareness indicators, renewable energy indicators, environmental protection facility indicators, environmental certification indicators, and natural environment image indicators.

[0018] Preferably, the step of obtaining standard environmental information images under each classification index as a dataset based on the environmental information image classification index, and using the dataset to train a pre-built machine learning-based computer image classification model includes:

[0019] S31. Based on the environmental information image classification index, obtain standard environmental information images under each classification index as a dataset, and divide the dataset into a training dataset and a validation dataset according to a certain index.

[0020] S32. Train the machine learning-based computer image classification model using the training set, and use the validation set to test the image classification accuracy of the machine learning-based computer image classification model.

[0021] Preferably, the machine learning-based computer image classification model includes a residual network.

[0022] Secondly, this invention also proposes a large-scale enterprise environmental information image disclosure content classification system, the system comprising:

[0023] The image acquisition and processing module is used to acquire an image of the environment information to be classified and to preprocess the image of the environment information to be classified.

[0024] The image classification index acquisition module is used to acquire environmental information image classification indices based on content analysis and grounded theory image classification methods.

[0025] The image classification model training and optimization module is used to obtain standard environmental information images under each classification index as a dataset based on the environmental information image classification index, and to use the dataset to train a pre-built computer image classification model based on machine learning.

[0026] The environmental information image classification module is used to classify the preprocessed environmental information image to be classified based on a trained machine learning-based computer image classification model.

[0027] Preferably, the image acquisition and processing module acquires an image of the environment information to be classified, and preprocesses the image of the environment information to be classified, including:

[0028] S11. Use web crawler programs to obtain large-scale environmental information images;

[0029] S12. Perform preprocessing on the environmental information image, including image data cleaning.

[0030] Preferably, the environmental information image classification indicators include environmental management awareness indicators, renewable energy indicators, environmental protection facility indicators, environmental certification indicators, and natural environment image indicators.

[0031] Preferably, the image classification model training and optimization module obtains standard environmental information images under each classification index as a dataset based on the environmental information image classification index, and uses the dataset to train a pre-built machine learning-based computer image classification model, including:

[0032] S31. Based on the environmental information image classification index, obtain standard environmental information images under each classification index as a dataset, and divide the dataset into a training dataset and a validation dataset according to a certain index.

[0033] S32. Train the machine learning-based computer image classification model using the training set, and use the validation set to test the image classification accuracy of the machine learning-based computer image classification model.

[0034] Preferably, the machine learning-based computer image classification model includes a residual network.

[0035] (III) Beneficial Effects

[0036] This invention provides a method and system for classifying the content disclosed in large-scale enterprise environmental information images. Compared with existing technologies, it has the following advantages:

[0037] 1. This invention first acquires environmental information images to be classified and preprocesses them; then, it obtains classification indicators for environmental information images based on content analysis and grounded theory image classification methods; and based on these indicators, it acquires standard environmental information images under each classification indicator as a dataset, and uses this dataset to train a pre-constructed machine learning-based computer image classification model; finally, it uses the trained machine learning-based computer image classification model to classify the preprocessed environmental information images to be classified. This invention uses the acquired environmental information image classification indicators as the classification standard for environmental information images, and trains the machine learning-based computer image classification model to the expected effect, finally performing automated classification of large-scale environmental information image sets. Compared with manual classification, it has the characteristics of high efficiency and high accuracy. Furthermore, based on the accurate classification results of environmental information images obtained by this invention, it allows for the study of corporate environmental information disclosure from an image perspective, quickly understanding the visual disclosure tendencies of enterprises in a short time.

[0038] 2. This invention establishes a large-scale environmental information image classification standard based on content analysis and grounded theory, which is closely aligned with the content presented by the image itself and is applicable to environmental information image classification. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a flowchart of a method for classifying the content of large-scale enterprise environmental information images disclosed according to the present invention;

[0041] Figure 2 This is a system block diagram of a large-scale enterprise environmental information image disclosure content classification system according to the present invention. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention are described clearly and completely. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] This application provides a method and system for classifying the content of large-scale corporate environmental information images, which solves the problem of low accuracy in existing computer image classification models for large-scale environmental image content recognition and classification. It aims to accurately grasp the disclosure tendencies of corporate environmental information by precisely understanding the disclosure characteristics of corporate environmental information images.

[0044] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0045] Example 1:

[0046] Firstly, this invention proposes a method for classifying the content disclosed in large-scale enterprise environmental information images, see [link to relevant documentation]. Figure 1 The method includes:

[0047] S1. Obtain the environmental information image to be classified, and preprocess the environmental information image to be classified;

[0048] S2. Obtaining environmental information image classification indicators based on content analysis and grounded theory;

[0049] S3. Based on the environmental information image classification index, obtain standard environmental information images under each classification index as a dataset, and use the dataset to train a pre-built computer image classification model based on machine learning.

[0050] S4. Classify the preprocessed environmental information image to be classified based on the trained machine learning-based computer image classification model.

[0051] As can be seen, this embodiment first acquires the environmental information images to be classified and preprocesses them; then, it obtains environmental information image classification indicators based on content analysis and grounded theory image classification methods; based on these indicators, it acquires standard environmental information images under each classification indicator as a dataset, and uses this dataset to train a pre-built machine learning-based computer image classification model; finally, it uses the trained machine learning-based computer image classification model to classify the preprocessed environmental information images to be classified. This embodiment uses the acquired environmental information image classification indicators as the classification standard for environmental information images, trains the machine learning-based computer image classification model to the expected effect, and finally performs automated classification of a large-scale environmental information image set. Compared with manual classification, it has the characteristics of high efficiency and high accuracy.

[0052] The following is in conjunction with the appendix Figure 1 The following details the implementation process of an embodiment of the present invention, including explanations of the specific steps S1-S4.

[0053] S1. Obtain the environmental information image to be classified and preprocess the environmental information image to be classified.

[0054] S11. Use web crawler programs to obtain large-scale environmental information images.

[0055] Corporate social responsibility reports and corporate social media accounts often disclose environmental information images. However, due to space limitations in social responsibility reports, the number of images included is relatively small. Therefore, this embodiment selects the official accounts of corporate social media platforms as the source of environmental information images. Taking WeChat official accounts as an example, after a company creates a WeChat official account, it can publish individual information through official account articles. An official account article generally includes the following parts: title, article content, and images in the content.

[0056] Then, commercial web crawlers are used to collect large-scale image data. Their working principle is to send HTTP requests to the target server, which then returns a response. The crawler client receives the response, extracts the data, and stores it. Based on whether the titles of WeChat public account articles contain keywords related to environmental management, such as environment, green, environmental protection, and pollution control, articles related to environmental management are selected, and images from these articles are extracted.

[0057] S12. Perform preprocessing on the environmental information image, including image data cleaning.

[0058] After acquiring large-scale environmental information images, image data cleaning was performed on these images; images that could not convey environmental information, such as QR codes and decorative icons, were removed from the articles; and a small number of images that appeared repeatedly in different articles were retained.

[0059] S2. Image classification methods based on content analysis and grounded theory to obtain environmental information image classification indicators.

[0060] Content analysis is a research method that objectively, systematically, and quantitatively describes the content of communication. Its essence is the analysis of the amount of information contained in the content of communication and its changes, that is, the process of inferring the accurate meaning from the meaningful words and phrases. The process of content analysis is a process of layer-by-layer reasoning.

[0061] This embodiment collects research papers related to environmental information disclosure from 2010 to 2020, and uses content analysis to explore the indicators used in previous research projects to measure the degree of environmental information disclosure.

[0062] Simultaneously, combining grounded theory's image classification method, five main image themes were identified for social and environmental images in sustainable development reports: (a) corporate efforts to protect the environment and care for others, (b) corporate activities that contribute to social welfare, (c) innovative and creative solutions, (d) collaboration and external recognition, and (e) pristine nature protected by corporations. This embodiment focuses solely on environmental information images, using GRI (Global Reporting Initiative) disclosure indicators to develop preliminary standards for environmental information image classification.

[0063] Ultimately, we selected and summarized the five most widely used categories of environmental information image classification indicators, including environmental management awareness indicators, renewable energy indicators, environmental protection facility indicators, environmental certification indicators, and natural environment image indicators. Specifically,

[0064] (a) Environmental Management Awareness: Environmental awareness refers to the level and degree of people's understanding of the environment and environmental protection. It is the behavior of people constantly adjusting their own activities and consciously coordinating the relationship between people and the environment, and between people and nature, in order to protect the environment. The environmental management awareness imagery showcases the active environmental management practices of corporate management or employees, conveying their moral understanding and commitment to environmental protection. This includes employee environmental awareness training, environmental publicity, tree planting activities, environmental slogans, etc.

[0065] (b) Renewable Energy: Renewable energy refers to non-fossil energy sources such as wind, solar, and hydropower, which are clean energy sources. Images of renewable energy mainly include wind turbines and solar panels.

[0066] (c) Environmental protection facilities: Environmental protection facilities are the equipment and devices required to treat substances generated during industrial production and operation that have an impact on the environment, so as to meet legal requirements, as well as environmental monitoring equipment.

[0067] (d) Environmental Certification: Environmental certification refers to various environmental certificates and awards obtained by a company. Environmental certifications come from independent external organizations, such as environmental protection organizations, newspapers, and certification bodies. Environmental certification images are considered evidence that a company's environmental management efforts or achievements are recognized externally.

[0068] (e) Natural Environment: Images of the natural environment include original natural images that are protected by the company and have not been damaged, as well as images of the company's production and operation sites and surrounding greenery.

[0069] S3. Based on the environmental information image classification index, obtain standard environmental information images under each classification index as a dataset, and use the dataset to train a pre-built computer image classification model based on machine learning.

[0070] To address the time-consuming and labor-intensive nature of manual classification of large-scale environmental information images in existing technologies, this embodiment selects a machine learning-based computer image classification model to classify the environmental information images to be classified.

[0071] There are many machine learning-based computer image classification models. In this embodiment, we choose convolutional neural networks (CNNs) as the machine learning-based computer image classification model. CNNs are widely used in image content recognition and classification tasks. A CNN is a type of feedforward neural network that includes convolutional computations and has a deep structure. Depending on the network structure, specific examples include LeNet, AlexNet, VGG, GoogLeNet, and ResNet. Residual networks (ResNets) are widely used in object classification and other fields, and are a typical convolutional network, forming part of the classic neural network backbone for computer vision tasks. In this embodiment, we use ResNe-50 (a residual neural network with 50 convolutional layers) to classify environmental information images.

[0072] S31. Based on the environmental information image classification index, obtain standard environmental information images under each classification index as a dataset, and divide the dataset into a training dataset and a validation dataset according to certain indicators.

[0073] In this embodiment, we employ supervised machine learning to train and optimize a machine learning-based computer image classification model. Supervised learning is a training method in machine learning that involves adjusting the parameters of a classifier using a set of samples of known categories to achieve the desired performance. Therefore, given a defined image classification standard, we use supervised machine learning to train the image classification model.

[0074] We collected training and validation datasets and labeled them with corresponding categories. We searched and downloaded 500 additional images each from the Baidu Images website for environmental management awareness, renewable energy, environmental protection facilities, environmental certification, and natural environment, and labeled them with environmental information categories. We used 85.6% of the images in each category as the labeled dataset in the machine learning process, and used the remaining 14.4% of the images in each category as the validation set to measure the accuracy of machine learning.

[0075] S32. Train the machine learning-based computer image classification model using the training set, and use the validation set to test the image classification accuracy of the machine learning-based computer image classification model.

[0076] Select, train, and validate an image classification network model. After training the model with the labeled dataset described above, use a validation set to validate its image classification accuracy.

[0077] S4. Classify the preprocessed environmental information image to be classified based on the trained machine learning-based computer image classification model.

[0078] Once the image classification accuracy of the aforementioned machine learning-based computer image classification model reaches the expected target after training, the model can be used for image classification.

[0079] By using the trained ResNe-50 image classification model to classify the environmental information images collected in S1, the number of images related to environmental management awareness, renewable energy, environmental protection facilities, environmental certification, and natural environment can be obtained. Researchers can then measure a company's tendency to disclose environmental information by the number of various types of environmental information images disclosed by the company.

[0080] This completes the entire process of the present invention, a method for classifying the content of large-scale enterprise environmental information images.

[0081] Example 2:

[0082] Secondly, this invention also provides a large-scale enterprise environmental information image disclosure content classification system, see [link to relevant documentation]. Figure 2 The system includes:

[0083] The image acquisition and processing module is used to acquire an image of the environment information to be classified and to preprocess the image of the environment information to be classified.

[0084] The image classification index acquisition module is used to acquire environmental information image classification indices based on content analysis and grounded theory image classification methods.

[0085] The image classification model training and optimization module is used to obtain standard environmental information images under each classification index as a dataset based on the environmental information image classification index, and to use the dataset to train a pre-built computer image classification model based on machine learning.

[0086] The environmental information image classification module is used to classify the preprocessed environmental information image to be classified based on a trained machine learning-based computer image classification model.

[0087] Optionally, the image acquisition and processing module acquires an image of the environment information to be classified, and preprocesses the image of the environment information to be classified, including:

[0088] S11. Use web crawler programs to obtain large-scale environmental information images;

[0089] S12. Perform preprocessing on the environmental information image, including image data cleaning.

[0090] Optionally, the environmental information image classification indicators include environmental management awareness indicators, renewable energy indicators, environmental protection facility indicators, environmental certification indicators, and natural environment image indicators.

[0091] Optionally, the image classification model training and optimization module obtains standard environmental information images under each classification index as a dataset based on the environmental information image classification index, and uses the dataset to train a pre-built machine learning-based computer image classification model, including:

[0092] S31. Based on the environmental information image classification index, obtain standard environmental information images under each classification index as a dataset, and divide the dataset into a training dataset and a validation dataset according to a certain index.

[0093] S32. Train the machine learning-based computer image classification model using the training set, and use the validation set to test the image classification accuracy of the machine learning-based computer image classification model.

[0094] Optionally, the machine learning-based computer image classification model includes a residual network.

[0095] It is understood that the classification system for large-scale enterprise environmental information image disclosure provided in this embodiment of the invention corresponds to the above-mentioned classification method for large-scale enterprise environmental information image disclosure. The explanations, examples, and beneficial effects of its relevant content can be referred to the corresponding content in the classification method for large-scale enterprise environmental information image disclosure, and will not be repeated here.

[0096] In summary, compared with existing technologies, it has the following beneficial effects:

[0097] 1. This invention first acquires environmental information images to be classified and preprocesses them; then, it obtains classification indicators for environmental information images based on content analysis and grounded theory image classification methods; and based on these indicators, it acquires standard environmental information images under each classification indicator as a dataset, and uses this dataset to train a pre-constructed machine learning-based computer image classification model; finally, it uses the trained machine learning-based computer image classification model to classify the preprocessed environmental information images to be classified. This invention uses the acquired environmental information image classification indicators as the classification standard for environmental information images, and trains the machine learning-based computer image classification model to the expected effect, finally performing automated classification of large-scale environmental information image sets. Compared with manual classification, it has the characteristics of high efficiency and high accuracy. Furthermore, based on the accurate classification results of environmental information images obtained by this invention, it allows for the study of corporate environmental information disclosure from an image perspective, quickly understanding the visual disclosure tendencies of enterprises in a short time.

[0098] 2. This invention establishes a large-scale environmental information image classification standard based on content analysis and grounded theory, which closely matches the content presented by the image itself and is applicable to environmental information image classification.

[0099] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0100] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for classifying the content of large-scale enterprise environmental information image disclosures, characterized in that, The method includes: Acquire an image of the environment information to be classified, and preprocess the image of the environment information to be classified; Image classification methods based on content analysis and grounded theory are used to obtain environmental information image classification indicators; Based on the environmental information image classification index, standard environmental information images under each classification index are obtained as a dataset, and the dataset is used to train a pre-built computer image classification model based on machine learning. The preprocessed environmental information image to be classified is classified based on a trained machine learning-based computer image classification model. The environmental information image classification indicators include environmental management awareness indicators, renewable energy indicators, environmental protection facility indicators, environmental certification indicators, and natural environment image indicators.

2. The method as described in claim 1, characterized in that, The process of acquiring the environmental information image to be classified and preprocessing the environmental information image to be classified includes: S11. Use web crawler programs to obtain large-scale environmental information images; S12. Perform preprocessing on the environmental information image, including image data cleaning.

3. The method as described in claim 1, characterized in that, The process of obtaining standard environmental information images under each classification index as a dataset based on the environmental information image classification index, and using the dataset to train a pre-built machine learning-based computer image classification model includes: S31. Based on the environmental information image classification index, obtain standard environmental information images under each classification index as a dataset, and divide the dataset into a training dataset and a validation dataset according to a certain index. S32. Train the machine learning-based computer image classification model using the training dataset, and use the validation dataset to test the image classification accuracy of the machine learning-based computer image classification model.

4. The method as described in claim 3, characterized in that, The machine learning-based computer image classification model includes a residual network.

5. A large-scale enterprise environmental information image disclosure content classification system, characterized in that, The system includes: The image acquisition and processing module is used to acquire an image of the environment information to be classified and to preprocess the image of the environment information to be classified. The image classification index acquisition module is used to acquire environmental information image classification indices based on content analysis and grounded theory image classification methods. The image classification model training and optimization module is used to obtain standard environmental information images under each classification index as a dataset based on the environmental information image classification index, and to use the dataset to train a pre-built computer image classification model based on machine learning. An environmental information image classification module is used to classify the preprocessed environmental information image to be classified based on a trained machine learning-based computer image classification model. The environmental information image classification indicators include environmental management awareness indicators, renewable energy indicators, environmental protection facility indicators, environmental certification indicators, and natural environment image indicators.

6. The system as described in claim 5, characterized in that, The image acquisition and processing module acquires an image of the environment information to be classified, and performs preprocessing on the image of the environment information to be classified, including: S11. Use web crawler programs to obtain large-scale environmental information images; S12. Perform preprocessing on the environmental information image, including image data cleaning.

7. The system as described in claim 5, characterized in that, The image classification model training and optimization module obtains standard environmental information images under each classification index as a dataset based on the environmental information image classification index, and uses the dataset to train a pre-built machine learning-based computer image classification model, including: S31. Based on the environmental information image classification index, obtain standard environmental information images under each classification index as a dataset, and divide the dataset into a training dataset and a validation dataset according to a certain index. S32. Train the machine learning-based computer image classification model using the training dataset, and use the validation dataset to test the image classification accuracy of the machine learning-based computer image classification model.

8. The system as described in claim 7, characterized in that, The machine learning-based computer image classification model includes a residual network.

Images