Asset loss early warning system and method based on property right data fusion, and terminal equipment

Abstract

The invention provides an asset loss early warning system and method based on property right data fusion and terminal equipment, and relates to the technical field of data processing, and the system comprises an image collection module which is configured to extract first enterprise information from a target image; the first data acquisition module is configured to acquire second enterprise information of each enterprise and construct a property right tree model; the second data acquisition module is configured to acquire target investment project information and determine third enterprise information of the target enterprise; and an early warning module configured to match the third enterprise information with the property right tree model, determine a first weight, obtain fourth enterprise information of the target enterprise, determine a second weight based on the first weight and the fourth enterprise information, determine a third weight according to the first enterprise information of the target enterprise, determine a fourth weight according to the target investment project information, and perform early warning according to the target investment project information. And calling the asset loss early-warning model to output an asset loss early-warning index. According to the invention, the asset loss monitoring efficiency and accuracy can be effectively improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and specifically to an asset loss early warning system based on property rights data fusion, an asset loss early warning method based on property rights data fusion, a computer-readable storage medium and a terminal device. Background Art

[0002] The existing corporate equity structure is complex, and asset loss monitoring is usually done by manually searching for corporate information and equity information of each group and its subsidiaries through corporate information query software, and then conducting asset loss monitoring by means of manual screening. However, the above method requires a lot of time and manpower costs, and therefore cannot achieve high-frequency large-scale screening, making it difficult to timely and effectively predict the risk of asset loss in advance, and greatly increasing the complexity and workload of property rights work. Summary of the invention

[0003] The purpose of the embodiments of the present application is to provide an asset loss warning system based on property rights data fusion, an asset loss warning method based on property rights data fusion, a computer-readable storage medium and a terminal device to solve the above problems.

[0004] In order to achieve the above-mentioned purpose, the first aspect of the present application provides an asset loss early warning system based on property rights data fusion, which is applied to a server, and the system includes:

[0005] An image acquisition module is configured to acquire a target image and extract the first enterprise information from the acquired target image;

[0006] A first data acquisition module is configured to acquire second enterprise information of each enterprise, determine the investment relationship of each enterprise based on the second enterprise information of each enterprise, and construct a property tree model including the investment relationship of each enterprise;

[0007] The second data acquisition module is configured to acquire a target property code, call a specified interface, acquire target investment project information associated with the target property code through the target property code, and determine third enterprise information of the target enterprise associated with the target investment project information based on the target investment project information;

[0008] The early warning module is configured to call the property tree model, match the third enterprise information with the property tree model, determine a first weight based on the matching result, obtain the fourth enterprise information of the target enterprise, determine a second weight based on the first weight and the fourth enterprise information, determine a third weight based on the first enterprise information of the target enterprise, determine a fourth weight based on the target investment project information, call the asset loss early warning model, output an asset loss early warning index based on the first weight, the second weight, the third weight and the fourth weight, and generate an alarm message when the asset loss early warning index is abnormal.

[0009] Optionally, extracting the first enterprise information from the acquired target image includes:

[0010] Performing character recognition on a character region including the first enterprise information in the collected target image, and determining a first confidence level of each character;

[0011] Matching the first confidence of each character with a preset first confidence threshold, and if there is a first target character whose first confidence is less than the first confidence threshold in the current character area, generating a first sub-image including only all the first target characters and a second sub-image not including any first target character;

[0012] Perform character recognition on the first sub-image to determine the second confidence of each character. If the second confidence of all first target characters in the first sub-image reaches the first confidence threshold, extract the first enterprise information in the current target image based on the character recognition results of the first sub-image and the second sub-image.

[0013] Optionally, extracting the first enterprise information from the collected target image further includes:

[0014] If there are first target characters whose second confidence is less than the first confidence threshold in the first sub-image, determining whether the second confidence of each first target character is greater than a second confidence threshold, wherein the second confidence threshold is less than the first confidence threshold;

[0015] If the second confidences of all first target characters in the first sub-image reach the second confidence threshold, extracting the first enterprise information in the current target image based on the character recognition results of the first sub-image and the second sub-image;

[0016] If the second confidence of the current first target character is less than the second confidence threshold, determining that the current first target character is the second target character, and generating a third sub-image including only all the second target characters and a fourth sub-image not including any second target character based on the first sub-image;

[0017] After performing at least one of scaling, brightness adjustment and rotation on the third sub-image, character recognition is performed on the third sub-image to determine a third confidence level of each character, and if the third confidence levels of all second target characters in the third sub-image reach the second confidence level threshold, first enterprise information in the current target image is extracted based on the character recognition results of the second sub-image, the third sub-image and the fourth sub-image;

[0018] If there is a second target character in the third sub-image whose third confidence is less than the second confidence threshold, the second target character is determined to be the third target character, and a fifth sub-image including only all third target characters and a sixth sub-image that does not include any third target character are generated based on the third sub-image, and character recognition is performed several times on the fifth sub-image. The recognition result corresponding to the highest confidence of each third target character in the several character recognitions is used as the recognition result of the fifth sub-image, and the first enterprise information in the current target image is extracted based on the character recognition results of the second sub-image, the fourth sub-image, the fifth sub-image and the sixth sub-image.

[0019] Optionally, the first enterprise information includes information on board members of the enterprise, and the second enterprise information includes enterprise names, enterprise names of invested enterprises associated with the current enterprise, and investment proportions of the current enterprise to each invested enterprise, wherein the current enterprise is the investing enterprise of the invested enterprise associated with it; determining the investment relationship of each enterprise based on the second enterprise information of each enterprise includes:

[0020] Query the subordinate investment enterprises associated with each enterprise, and determine the subordinate investment enterprises whose investment proportion reaches the preset investment proportion threshold as the target investment enterprises of the current enterprise; or

[0021] If the investment proportion of the subordinate investment enterprise of the current enterprise does not reach the investment proportion threshold, obtain the board member information of the subordinate investment enterprise, match the obtained board member information with the preset human resources information table, if the board member information matches the human resources information table, determine that the board member corresponding to the board member information is a first category board member, if the proportion of the first category board members in the subordinate investment enterprise to all board members reaches the board member proportion threshold, determine that the subordinate investment enterprise is the target investment enterprise of the current enterprise, wherein the human resources information table includes different members and their corresponding member information;

[0022] If there are multiple subordinate investment enterprises of the current enterprise, if the sum of the investment proportions of all subordinate investment enterprises of the current enterprise reaches the investment proportion threshold, all subordinate investment enterprises of the current enterprise are determined as target investment enterprises;

[0023] Repeat the above process until the target investment enterprises of all enterprises are determined and the investment relationships of each enterprise are obtained.

[0024] Optionally, the third enterprise information includes the enterprise name of the target enterprise and the enterprise name of the invested target enterprise, and matching the third enterprise information with the property tree model and determining the first weight based on the matching result includes:

[0025] If the enterprise name of the target enterprise matches the enterprise name in the property right tree model, the shareholder category of the target enterprise is determined to be the first type of shareholder; otherwise, the shareholder category of the target enterprise is determined to be the second type of shareholder;

[0026] A corresponding first weight is determined based on the shareholder category of the target enterprise, wherein the first weights corresponding to the first category of shareholders and the second category of shareholders are different.

[0027] Optionally, the fourth enterprise information includes the paid-in amount and subscribed amount of the target enterprise, and the target investment project information includes the preset expenditure amount and the actual expenditure amount of the target investment project; determining the second weight based on the first weight and the fourth enterprise information includes:

[0028] Taking the ratio of the paid-in amount to the subscribed amount of the target enterprise as the first adjustment coefficient, and determining the second weight based on the product of the first adjustment coefficient and the first weight;

[0029] Determining a third weight according to the first enterprise information of the target enterprise includes:

[0030] Acquire information on board members of the target enterprise according to the first enterprise information of the target enterprise, and determine the number of first category board members of the target enterprise based on the human resources information table;

[0031] The ratio of the number of first category board members of the target enterprise to the total number of board members of the invested target enterprise is used as the second adjustment coefficient, and the second weight is determined based on the sum of the second adjustment coefficient and a preset third adjustment coefficient;

[0032] Determining a fourth weight according to the target investment project information includes:

[0033] The fourth weight is determined based on the ratio of the actual expenditure amount of the target investment project to the preset expenditure amount.

[0034] Optionally, the asset loss early warning model includes:

[0035] Asset loss warning index = ((shareholding ratio*S) / number of first category board members of the target enterprise)*D*F;

[0036] Among them, the shareholding ratio is the shareholding ratio of the target enterprise to the invested enterprise corresponding to the target investment project, S is the second weight, D is the third weight, and F is the fourth weight.

[0037] In a second aspect of the present application, there is provided an asset loss early warning method based on property rights data fusion, using the above system, the method comprising:

[0038] Capturing a target image, and extracting first enterprise information from the captured target image;

[0039] Acquire the second enterprise information of each enterprise, determine the investment relationship of each enterprise based on the second enterprise information of each enterprise, and construct a property tree model including the investment relationship of each enterprise;

[0040] Obtain a target property code, call a specified interface, obtain target investment project information associated with the target property code through the target property code, and determine third enterprise information of the target enterprise associated with the target investment project information based on the target investment project information;

[0041] Call the property tree model, match the third enterprise information with the property tree model, determine the first weight based on the matching result, obtain the fourth enterprise information of the target enterprise, determine the second weight based on the first weight and the fourth enterprise information, determine the third weight based on the first enterprise information of the target enterprise, determine the fourth weight based on the target investment project information, call the asset loss warning model, output the asset loss warning index based on the first weight, the second weight, the third weight and the fourth weight, and generate an alarm message when the asset loss warning index is abnormal.

[0042] In a third aspect of the present application, a computer-readable storage medium is provided, which stores a computer program which, when executed by a processor, causes the processor to execute the asset loss warning method based on property rights data fusion as described above.

[0043] In a fourth aspect of the present application, a terminal device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned asset loss warning method based on property rights data fusion when executing the computer program.

[0044] The embodiments provided in this application have the following beneficial effects:

[0045] This application is based on big data analysis and can effectively improve the monitoring efficiency and accuracy of asset loss and reduce labor costs and workload.

[0046] Other features and advantages of the embodiments or implementations of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present application, but do not constitute a limitation on the embodiments of the present application. In the accompanying drawings:

[0048] Figure 1 A schematic block diagram of an asset loss early warning system based on property rights data fusion according to an embodiment of the present application is schematically shown;

[0049] Figure 2 A method flow chart schematically illustrates an asset loss early warning method based on property rights data fusion according to an embodiment of the present application;

[0050] Figure 3 A schematic diagram of the structure of a terminal device according to an embodiment of the present application is shown schematically.

[0051] Description of Reference Numerals

[0052] 10-terminal device, 100-processor, 101-memory, 102-computer program. DETAILED DESCRIPTION

[0053] The specific implementation of the embodiment of the present application is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present application, and is not used to limit the embodiment of the present application.

[0054] In order to solve the above problems, Figure 1 As shown, the first aspect of the present application provides an asset loss early warning system based on property rights data fusion, which is applied to a server, and the system includes:

[0055] An image acquisition module is configured to acquire a target image and extract the first enterprise information from the acquired target image;

[0056] A first data acquisition module is configured to acquire second enterprise information of each enterprise, determine the investment relationship of each enterprise based on the second enterprise information of each enterprise, and construct a property tree model including the investment relationship of each enterprise;

[0057] The second data acquisition module is configured to acquire a target property code, call a specified interface, acquire target investment project information associated with the target property code through the target property code, and determine third enterprise information of the target enterprise associated with the target investment project information based on the target investment project information;

[0058] The early warning module is configured to call the property tree model, match the third enterprise information with the property tree model, determine a first weight based on the matching result, obtain the fourth enterprise information of the target enterprise, determine a second weight based on the first weight and the fourth enterprise information, determine a third weight based on the first enterprise information of the target enterprise, determine a fourth weight based on the target investment project information, call the asset loss early warning model, output an asset loss early warning index based on the first weight, the second weight, the third weight and the fourth weight, and generate an alarm message when the asset loss early warning index is abnormal.

[0059] In this way, this application, based on big data analysis, can effectively improve the monitoring efficiency and accuracy of asset loss and reduce labor costs and workload.

[0060] It is understandable that the technical personnel in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0061] Specifically, before implementing asset loss monitoring, first of all, before the company's various subsidiaries prepare to conduct formal property rights registration, pre-registration of property rights is carried out within the company, so as to have a preliminary understanding of the industrial and commercial registration to be carried out by each subsidiary, and play a restraining role in the property rights registration work of each subsidiary after the industrial and commercial registration, so that the property rights management workers can use the property rights registration situation to understand the company's shareholders, the board of directors and the board of supervisors, the subscription and actual payment of each shareholder, and the actual payment ratio, etc., to prevent the possibility of insufficient capital contribution, false shareholding, affiliation, etc.

[0062] In the present application, the image acquisition module may be an OCR recognition device, wherein OCR (Optical Character Recognition) is optical character recognition. For example, various documents of an enterprise, such as a business license, articles of association, etc., may be scanned by OCR to obtain enterprise-related data, and the obtained data may be used as a basis to form enterprise basic data.

[0063] However, in the current OCR recognition method, the OCR engine usually only recognizes an image once. When the OCR engine recognizes an image, it first converts the image into a black and white binary image. The OCR engine determines the position of the text in the image and the position of each character in the text by calculating the number of black pixels and white pixels in the image. Based on the determined character area, the OCR engine can analyze each generated character area to identify the corresponding characters, font size, and coordinates of the characters on the image. However, when other large non-text elements such as logos appear side by side with text in a document, the black pixels created by the non-text elements may cause the OCR engine to miss adjacent text elements. For example, if the font size and / or line spacing between text columns on a single page are different, the OCR engine may fail to detect or incorrectly recognize characters.

[0064] In order to solve this problem, in this application, the first enterprise information is extracted from the collected target image, including:

[0065] S100. Perform character recognition on the character area of ​​the collected target image that includes the first corporate information, and determine the first confidence of each character. It is understandable that the detector for OCR text recognition can be a traditional machine learning model, a deep convolutional neural network, or an SVM classifier, etc., which is not limited here. First, the character area of ​​the business license, charter, etc. is recognized by the OCR engine, and the first confidence of each character is calculated. It is understandable that the confidence is an indicator to measure the accuracy of the OCR recognition result. The higher the confidence, the higher the accuracy of the recognition result.

[0066] S200, matching the first confidence of each character with the preset first confidence threshold, if there is a first target character with a first confidence less than the first confidence threshold in the current character area, generating a first sub-image including only all first target characters and a second sub-image not including any first target character. The first confidence threshold may be 90%, that is, if the recognition result confidence of the target character reaches 90%, the character recognition result is confirmed to be accurate, otherwise, the character determined to be an accurate character recognition result is deleted or covered from the original character area, and a new image is generated, the newly generated image only includes characters whose character recognition confidence does not reach 90%, so as to perform secondary recognition on these characters.

[0067] S300, perform character recognition on the first sub-image, determine the second confidence of each character, if the second confidence of all the first target characters in the first sub-image reaches the first confidence threshold, extract the first enterprise information in the current target image based on the character recognition results of the first sub-image and the second sub-image. Based on the newly generated image, perform secondary recognition on the characters whose character recognition confidence does not reach 90%, if the recognition confidence reaches 90%, extract the first enterprise information based on the character recognition results of the first sub-image and the second sub-image. For example, the original image includes characters a, b, c, d. When performing character recognition on the original image, the recognition confidence of characters a, c, d reaches 90%, and the recognition confidence of character c is lower than 90%. After secondary recognition of character c, its confidence reaches 90%, then information extraction can be performed based on the positions of characters a, b, c, d in the original image.

[0068] In the present application, extracting the first enterprise information from the collected target image also includes:

[0069] S400: If there are first target characters whose second confidence is less than the first confidence threshold in the first sub-image, determine whether the second confidence of each first target character is greater than the second confidence threshold, wherein the second confidence threshold is less than the first confidence threshold. If there are still characters whose confidence is less than 90% after the second character recognition in the first sub-image, further determine whether the confidence of these characters reaches the second confidence threshold, for example, the second confidence threshold is 85%.

[0070] S500: If the second confidences of all first target characters in the first sub-image reach the second confidence threshold, extract the first enterprise information in the current target image based on the character recognition results of the first sub-image and the second sub-image. If the confidences of all characters reach 85% after the second character recognition of the first sub-image, extract the first enterprise information based on the character recognition results of the first sub-image and the second sub-image.

[0071] S600, if the second confidence of the current first target character is less than the second confidence threshold, determine that the current first target character is the second target character, and generate a third sub-image including only all second target characters and a fourth sub-image that does not include any second target character based on the first sub-image. If after the second character recognition of the first sub-image, there are characters with a confidence of less than 85%, refer to step S200, extract the characters with a confidence of less than 85%, and generate a new image, namely the third sub-image.

[0072] S700, after scaling, adjusting the brightness, and rotating the third sub-image, character recognition is performed on the third sub-image to determine the third confidence of each character. If the third confidence of all second target characters in the third sub-image reaches the second confidence threshold, the first corporate information in the current target image is extracted based on the character recognition results of the second sub-image, the third sub-image, and the fourth sub-image. In order to improve the recognition accuracy of the characters in the third sub-image, before performing character recognition on the third sub-image, the third sub-image is first scaled, for example, the third sub-image is enlarged, or the brightness of the third sub-image is increased, or the third sub-image is rotated, cropped, etc. The adjusted third sub-image is then subjected to character recognition again. If the confidence of all characters in the third sub-image reaches 85% after character recognition is performed again, the first corporate information can be extracted.

[0073] S800. If a second target character whose third confidence is less than the second confidence threshold exists in the third sub-image, the second target character is determined to be the third target character, and a fifth sub-image including only all the third target characters and a sixth sub-image not including any third target character are generated based on the third sub-image, and character recognition is performed several times on the fifth sub-image, and the recognition result corresponding to the highest confidence of each third target character in the several character recognitions is used as the recognition result of the fifth sub-image, and the first enterprise information in the current target image is extracted based on the character recognition results of the second sub-image, the fourth sub-image, the fifth sub-image and the sixth sub-image. If after character recognition is performed again, there are still characters in the third sub-image with recognition confidence levels lower than 85%, then refer to step S200 to extract the characters with confidence levels lower than 85%, generate a new image, i.e., the fifth sub-image, and perform a preset number of character recognitions on the fifth sub-image. For example, character recognition is performed three times, and for each character in the fifth sub-image, the recognition result with the highest confidence level among the three character recognitions is used as the final recognition result of the character, and the first enterprise information is extracted based on the recognition result.

[0074] Through the above steps of the present application, the accuracy of OCR character recognition can be effectively improved.

[0075] In this application, the first enterprise information includes the information of the board members of the enterprise, and the second enterprise information includes the enterprise name, the enterprise name of the invested enterprise associated with the current enterprise, and the investment ratio of the current enterprise to each invested enterprise, wherein the current enterprise is the investing enterprise of the invested enterprise associated with it. It can be understood that the second enterprise information can be obtained from the pre-registration information of the property rights within the enterprise. For example, the data structure of the second enterprise information is: {target enterprise name, target enterprise unified social credit code, investment ratio, invested enterprise unified social credit code, invested company name}.

[0076] Determine the investment relationship of each enterprise based on the second enterprise information of each enterprise, including:

[0077] Query the subordinate investment enterprises associated with each enterprise, and determine the subordinate investment enterprises whose investment proportion reaches the preset investment proportion threshold as the target investment enterprises of the current enterprise. For example, the second enterprise information can be traversed through the existing recursive algorithm to query the invested enterprises associated with enterprise A, including enterprise B, enterprise C, and enterprise D. Among them, the investment proportion of enterprise A in enterprise B and enterprise C has reached the preset investment proportion threshold, such as 51%, and then determine that the target investment enterprises of enterprise A are enterprise B and enterprise C.

[0078] Alternatively, if the investment proportion of the subordinate investment enterprise of the current enterprise does not reach the investment proportion threshold, the board member information of the subordinate investment enterprise is obtained, and the obtained board member information is matched with the preset human resources information table. If the board member information matches the human resources information table, it is determined that the board member corresponding to the board member information is a first category board member. If the proportion of the first category board members in the subordinate investment enterprise to all board members reaches the board member proportion threshold, the subordinate investment enterprise is determined to be the target investment enterprise of the current enterprise, wherein the human resources information table includes different members and their corresponding member information. Although the shareholding ratio of some enterprises does not reach 51%, the number of shareholders exceeds half, and actual control can be achieved. Therefore, after finding the enterprise with a proportion of less than 51% through the recursive algorithm, the first enterprise information obtained by OCR scanning the relevant documents of the enterprise is used to obtain all the board member information of the investment enterprise, and then the board member is judged whether it is an internal employee through the internal enterprise human resources information comparison. If the number of internal employees among the board members exceeds half of the total number of board members, the investment enterprise is judged to be the actual controlling enterprise, and the subordinate investment enterprises of the investment enterprise are continued to be found through the recursive algorithm function.

[0079] Alternatively, if there are multiple subordinate investment enterprises of the current enterprise, if the sum of the investment proportions of all subordinate investment enterprises of the current enterprise reaches the investment proportion threshold, all subordinate investment enterprises of the current enterprise are determined to be target investment enterprises. If the investment proportions of shareholders of some enterprises fail to meet the controlling requirements, then these enterprises are participating enterprises. However, since multiple shareholders jointly invest in one enterprise, the total shareholding ratio exceeds the controlling requirement ratio. From the perspective of group property rights, this enterprise is a controlling enterprise. Therefore, if there are multiple subordinate investment enterprises of the current enterprise, and they are all participating enterprises, then the controlling ratio requirements are determined by summing up the shareholding ratios of these enterprises. For example, the sum of the investment proportions of all participating enterprises reaches 51%. Repeat the above process until the target investment enterprises of all enterprises are determined, the investment relations of each enterprise are obtained, and a property rights tree model including the investment relations of each enterprise is established.

[0080] In this application, after establishing the property tree model, the names of companies within the group that are easily counterfeited can be set as keywords, and relevant company information can be obtained through keyword queries. If companies within the group are excluded, the group to which they belong is searched based on a recursive algorithm. If they are not in other groups, the company is written into the list of counterfeit companies.

[0081] In the present application, the second data acquisition module uses a property right code to query the information of relevant equity investment projects and the financial investment information corresponding to the equity investment project information from the corresponding platform by calling a specified interface. Among them, the property right code is a unique identifier bound to the enterprise, and the enterprise's property right information, legal person information and other enterprise information as well as relevant financial information and investment project information involved in the enterprise are fused and cleaned through the data fusion platform. It can be understood that each investment project information associated with the target property right code corresponds to at least one invested enterprise associated with the enterprise corresponding to the target property right code. For example, the target property right code is property right code A, and its corresponding enterprise is enterprise A. It is found that property right code A is associated with 2 investment project information, namely project A and project B. The invested enterprise corresponding to project A is enterprise B, and the invested enterprise corresponding to project B is enterprise C.

[0082] Wherein, the third enterprise information includes the enterprise name of the target enterprise and the enterprise name of the invested target enterprise corresponding to the target investment project information associated with the target enterprise, wherein the third enterprise information is matched with the property tree model, and the first weight is determined based on the matching result, including: when the target enterprise is a target investment enterprise, if the enterprise name of the target enterprise matches the enterprise name in the property tree model, the shareholder category of the target enterprise is determined to be a first-class shareholder; otherwise, the shareholder category of the target enterprise is determined to be a second-class shareholder, and the corresponding first weight is determined based on the shareholder category of the target enterprise, wherein the first weights corresponding to the first-class shareholders and the second-class shareholders are different.

[0083] This application divides corporate shareholders into two categories: within the group and outside the group. By comparing with the property rights tree, it can be determined whether the shareholder is within the group or outside. For example, if the name of the target enterprise to be queried is in the property rights tree, the target enterprise is judged to be an enterprise within the group and is determined to be a first-category shareholder. If the name of the target enterprise to be queried is not in the property rights tree, the target enterprise is judged to be an enterprise outside the group and is determined to be a second-category shareholder. For example, the first weight of shareholders within the group can be G=1, and the first weight of shareholders outside the group can be G=2.

[0084] In the present application, the fourth enterprise information includes the actually paid-in amount and the subscribed amount of the target enterprise, and the target investment project information includes the preset expenditure amount and the actual expenditure amount of the target investment project; the second weight is determined based on the first weight and the fourth enterprise information, including: taking the ratio of the actually paid-in amount to the subscribed amount of the target enterprise as the first adjustment coefficient, and determining the second weight based on the product of the first adjustment coefficient and the first weight, that is, S = (actually paid-in amount / subscribed amount) * G, where S is the second weight.

[0085] Among them, determining the third weight based on the first enterprise information of the target enterprise includes: obtaining the board member information of the target enterprise based on the first enterprise information of the target enterprise, and determining the number of first-category board members of the target enterprise based on the human resources information table; taking the ratio of the number of first-category board members of the target enterprise to the total number of board members of the invested target enterprise as the second adjustment coefficient, and determining the second weight based on the sum of the second adjustment coefficient and the preset third adjustment coefficient, that is, D=1+(number of first-category board members / total number of board members of the invested target enterprise).

[0086] The fourth weight is determined based on the target investment project information, including: determining the fourth weight based on the ratio of the actual expenditure amount of the target investment project to the preset expenditure amount, that is, F = actual expenditure amount / preset expenditure amount, wherein the actual expenditure amount is the actual financial expenditure, and the preset expenditure amount is the project specified expenditure.

[0087] Among them, the asset loss warning model includes: asset loss warning index = ((shareholding ratio * S) / number of first category board members of the target enterprise) * D * F; wherein the shareholding ratio is the shareholding ratio of the target enterprise to the invested target enterprise corresponding to the target investment project, S is the second weight, D is the third weight, and F is the fourth weight. In this application, the asset loss warning index abnormality can be determined by the value of the asset loss warning index. For example, when the calculated value of the asset loss warning index is higher than the preset threshold, the asset loss warning index is determined to be abnormal. An example of the calculation of the warning index is shown in Table 1.

[0088] Table 1

[0089]

[0090] Among them, in Table 1, the number of company shareholders is the number of first-category board members, and the number of board members is the total number of board members of the target company being invested.

[0091] According to Table 1, when the target enterprise is a holding enterprise, if the actual payment and subscription are inconsistent, the early warning index will be abnormal; when the number of board members of the first category is a minority and there is a situation of holding but not controlling power, the early warning index will also be abnormal; when the target enterprise is a shareholder, if the actual payment is greater than the subscription, the early warning index will also be abnormal. When the actual payment is normal and the number of shareholders meets the requirements, the early warning index will be within a reasonable range. In this way, this application automatically obtains and processes the relevant data information of the target enterprise periodically or by designated instructions, calculates the asset loss early warning index of the target enterprise, and automatically alarms according to the calculation results, which can effectively improve the efficiency and accuracy of asset loss early warning and reduce labor costs.

[0092] like Figure 2 As shown, in a second aspect of the present application, a method for early warning of asset loss based on property rights data fusion is provided, and the above system is applied. The method includes:

[0093] Capturing a target image, and extracting first enterprise information from the captured target image;

[0094] Acquire the second enterprise information of each enterprise, determine the investment relationship of each enterprise based on the second enterprise information of each enterprise, and construct a property tree model including the investment relationship of each enterprise;

[0095] Obtain a target property code, call a specified interface, obtain target investment project information associated with the target property code through the target property code, and determine third enterprise information of the target enterprise associated with the target investment project information based on the target investment project information;

[0096] Call the property tree model, match the third enterprise information with the property tree model, determine the first weight based on the matching result, obtain the fourth enterprise information of the target enterprise, determine the second weight based on the first weight and the fourth enterprise information, determine the third weight based on the first enterprise information of the target enterprise, determine the fourth weight based on the target investment project information, call the asset loss warning model, output the asset loss warning index based on the first weight, the second weight, the third weight and the fourth weight, and generate an alarm message when the asset loss warning index is abnormal.

[0097] In a third aspect of the present application, a computer-readable storage medium is provided, which stores a computer program which, when executed by a processor, causes the processor to execute the asset loss warning method based on property rights data fusion as described above.

[0098] In a fourth aspect of the present application, a terminal device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned asset loss warning method based on property rights data fusion when executing the computer program.

[0099] like Figure 3 FIG. 1 is a schematic diagram of a terminal device provided in an embodiment of the present application. Figure 3 As shown, the terminal device 10 of this embodiment includes: a processor 100, a memory 101, and a computer program 102 stored in the memory 101 and executable on the processor 100. When the processor 100 executes the computer program 102, the steps in the above method embodiment are implemented. Alternatively, when the processor 100 executes the computer program 102, the functions of each module / unit in the above device embodiments are implemented.

[0100] Exemplarily, the computer program 102 may be divided into one or more modules / units, one or more modules / units are stored in the memory 101, and are executed by the processor 100 to complete the present application. One or more modules / units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program 102 in the terminal device 10.

[0101] The terminal device 10 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The terminal device 10 may include, but is not limited to, a processor 100 and a memory 101. Those skilled in the art will appreciate that Figure 3 It is only an example of the terminal device 10 and does not constitute a limitation of the terminal device 10. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal device may also include input and output devices, network access devices, buses, etc.

[0102] The processor 100 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0103] The memory 101 may be an internal storage unit of the terminal device 10, such as a hard disk or memory of the terminal device 10. The memory 101 may also be an external storage device of the terminal device 10, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal device 10. Further, the memory 101 may also include both an internal storage unit of the terminal device 10 and an external storage device. The memory 101 is used to store computer programs and other programs and data required by the terminal device 10. The memory 101 may also be used to temporarily store data that has been output or is to be output.

[0104] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0105] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0106] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. An asset loss early warning system based on property rights data fusion, applied to a server, characterized in that: The system comprises: An image acquisition module is configured to acquire a target image and extract the first enterprise information from the acquired target image; A first data acquisition module is configured to acquire second enterprise information of each enterprise, determine the investment relationship of each enterprise based on the second enterprise information of each enterprise, and construct a property tree model including the investment relationship of each enterprise; The second data acquisition module is configured to acquire a target property code, call a specified interface, acquire target investment project information associated with the target property code through the target property code, and determine third enterprise information of the target enterprise associated with the target investment project information based on the target investment project information; The early warning module is configured to call the property tree model, match the third enterprise information with the property tree model, determine a first weight based on the matching result, obtain the fourth enterprise information of the target enterprise, determine a second weight based on the first weight and the fourth enterprise information, determine a third weight based on the first enterprise information of the target enterprise, determine a fourth weight based on the target investment project information, call the asset loss early warning model, output an asset loss early warning index based on the first weight, the second weight, the third weight and the fourth weight, and generate an alarm message when the asset loss early warning index is abnormal.

2. The asset loss early warning system based on property rights data fusion according to claim 1 is characterized in that: Extracting first enterprise information from the acquired target image includes: Performing character recognition on a character region including the first enterprise information in the collected target image, and determining a first confidence level of each character; Matching the first confidence of each character with a preset first confidence threshold, and if there is a first target character whose first confidence is less than the first confidence threshold in the current character area, generating a first sub-image including only all the first target characters and a second sub-image not including any first target character; Perform character recognition on the first sub-image to determine the second confidence of each character. If the second confidence of all first target characters in the first sub-image reaches the first confidence threshold, extract the first enterprise information in the current target image based on the character recognition results of the first sub-image and the second sub-image.

3. The asset loss early warning system based on property rights data fusion according to claim 2 is characterized in that: Extracting the first enterprise information from the collected target image also includes: If there are first target characters whose second confidence is less than the first confidence threshold in the first sub-image, determining whether the second confidence of each first target character is greater than a second confidence threshold, wherein the second confidence threshold is less than the first confidence threshold; If the second confidences of all first target characters in the first sub-image reach the second confidence threshold, extracting the first enterprise information in the current target image based on the character recognition results of the first sub-image and the second sub-image; If the second confidence of the current first target character is less than the second confidence threshold, determining that the current first target character is the second target character, and generating a third sub-image including only all the second target characters and a fourth sub-image not including any second target character based on the first sub-image; After performing at least one of scaling, brightness adjustment and rotation on the third sub-image, character recognition is performed on the third sub-image to determine a third confidence level of each character, and if the third confidence levels of all second target characters in the third sub-image reach the second confidence level threshold, first enterprise information in the current target image is extracted based on the character recognition results of the second sub-image, the third sub-image and the fourth sub-image; If there is a second target character in the third sub-image whose third confidence is less than the second confidence threshold, the second target character is determined to be the third target character, and a fifth sub-image including only all third target characters and a sixth sub-image that does not include any third target character are generated based on the third sub-image, and character recognition is performed several times on the fifth sub-image. The recognition result corresponding to the highest confidence of each third target character in the several character recognitions is used as the recognition result of the fifth sub-image, and the first enterprise information in the current target image is extracted based on the character recognition results of the second sub-image, the fourth sub-image, the fifth sub-image and the sixth sub-image.

4. The asset loss early warning system based on property rights data fusion according to claim 1 is characterized in that: The first enterprise information includes information on the board members of the enterprise, and the second enterprise information includes the enterprise name, the enterprise name of the invested enterprise associated with the current enterprise, and the investment ratio of the current enterprise to each invested enterprise, wherein the current enterprise is the investing enterprise of the invested enterprise associated with it; the investment relationship of each enterprise is determined based on the second enterprise information of each enterprise, including: Query the subordinate investment enterprises associated with each enterprise, and determine the subordinate investment enterprises whose investment proportion reaches the preset investment proportion threshold as the target investment enterprises of the current enterprise; or If the investment proportion of the subordinate investment enterprise of the current enterprise does not reach the investment proportion threshold, obtain the board member information of the subordinate investment enterprise, match the obtained board member information with the preset human resources information table, if the board member information matches the human resources information table, determine that the board member corresponding to the board member information is a first category board member, if the proportion of the first category board members in the subordinate investment enterprise to all board members reaches the board member proportion threshold, determine that the subordinate investment enterprise is the target investment enterprise of the current enterprise, wherein the human resources information table includes different members and their corresponding member information; If there are multiple subordinate investment enterprises of the current enterprise, if the sum of the investment proportions of all subordinate investment enterprises of the current enterprise reaches the investment proportion threshold, all subordinate investment enterprises of the current enterprise are determined as target investment enterprises; Repeat the above process until the target investment enterprises of all enterprises are determined and the investment relationships of each enterprise are obtained.

5. The asset loss early warning system based on property rights data fusion according to claim 4 is characterized in that: The third enterprise information includes the enterprise name of the target enterprise and the enterprise name of the invested target enterprise, and the third enterprise information is matched with the property tree model, and a first weight is determined based on the matching result, including: If the enterprise name of the target enterprise matches the enterprise name in the property right tree model, the shareholder category of the target enterprise is determined to be the first type of shareholder; otherwise, the shareholder category of the target enterprise is determined to be the second type of shareholder; A corresponding first weight is determined based on the shareholder category of the target enterprise, wherein the first weights corresponding to the first category of shareholders and the second category of shareholders are different.

6. The asset loss early warning system based on property rights data fusion according to claim 5 is characterized in that: The fourth enterprise information includes the paid-in amount and subscribed amount of the target enterprise, and the target investment project information includes the preset expenditure amount and actual expenditure amount of the target investment project; Determining a second weight based on the first weight and the fourth enterprise information includes: Taking the ratio of the paid-in amount to the subscribed amount of the target enterprise as the first adjustment coefficient, and determining the second weight based on the product of the first adjustment coefficient and the first weight; Determining a third weight according to the first enterprise information of the target enterprise includes: Acquire information on board members of the target enterprise according to the first enterprise information of the target enterprise, and determine the number of first category board members of the target enterprise based on the human resources information table; The ratio of the number of first category board members of the target enterprise to the total number of board members of the invested target enterprise is used as the second adjustment coefficient, and the second weight is determined based on the sum of the second adjustment coefficient and a preset third adjustment coefficient; Determining a fourth weight according to the target investment project information includes: The fourth weight is determined based on the ratio of the actual expenditure amount of the target investment project to the preset expenditure amount.

7. The asset loss early warning system based on property rights data fusion according to claim 6 is characterized in that: The asset loss early warning model includes: Asset loss warning index = ((shareholding ratio*S) / number of first category board members of the target enterprise)*D*F; Among them, the shareholding ratio is the shareholding ratio of the target enterprise to the invested enterprise corresponding to the target investment project, S is the second weight, D is the third weight, and F is the fourth weight.

8. An asset loss early warning method based on property rights data fusion, characterized in that: Using the system described in any one of claims 1 to 7, the method comprises: Capturing a target image, and extracting first enterprise information from the captured target image; Acquire the second enterprise information of each enterprise, determine the investment relationship of each enterprise based on the second enterprise information of each enterprise, and construct a property tree model including the investment relationship of each enterprise; Obtain a target property code, call a specified interface, obtain target investment project information associated with the target property code through the target property code, and determine third enterprise information of the target enterprise associated with the target investment project information based on the target investment project information; Call the property tree model, match the third enterprise information with the property tree model, determine the first weight based on the matching result, obtain the fourth enterprise information of the target enterprise, determine the second weight based on the first weight and the fourth enterprise information, determine the third weight based on the first enterprise information of the target enterprise, determine the fourth weight based on the target investment project information, call the asset loss warning model, output the asset loss warning index based on the first weight, the second weight, the third weight and the fourth weight, and generate an alarm message when the asset loss warning index is abnormal.

9. A computer-readable storage medium, characterized in that: A computer program is stored which, when executed by a processor, causes the processor to execute the asset loss early warning method based on property rights data fusion as described in claim 8.

10. A terminal 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 asset loss warning method based on property rights data fusion described in claim 8 is implemented.

Images