Retrieval system for multi-source heterogeneous data fusion

Abstract

The invention relates to a retrieval system for multi-source heterogeneous data fusion, which is applied to the field of electric digital data processing and is characterized in that various heterogeneous source data are analyzed and subjected to state calibration, and the various heterogeneous source data with similar calibration states in the same time period are grouped; compared with the prior art, the method has the advantages that various heterogeneous data can be obtained, the running state of the equipment can be quickly known and the safety of the equipment can be improved without independently retrieving the data of a single type, and meanwhile, the data can be accurately and quickly retrieved through the calibration vocabulary reflecting the data state. According to the method, a guiding function can be provided for the retrieval process, the problem that retrieval vocabularies are not accurate in the existing retrieval process is effectively solved, in addition, after the data state is calibrated, calibration vocabularies of multiple kinds of source data in the same time period are checked, when a certain same probability is achieved, equipment risk early warning is conducted, and the equipment safety is further improved.

Description

Technical Field

[0001] The present invention relates to a retrieval system, and in particular to a retrieval system for multi-source heterogeneous data fusion applied in the field of electronic digital data processing. Background Art

[0002] With the rapid development of smart grids, the safety and reliability of power equipment have become the key to the stable operation of the power grid. However, power monitoring, as an important means to ensure the stable operation of the power grid, optimize resource allocation, and prevent power accidents, has a wide variety of data, covers a wide range, and is updated in real time, resulting in a very large amount of data information. When monitoring personnel view it, it is difficult to quickly and accurately retrieve the required data.

[0003] To solve the above problems, the Chinese patent CN111143294B specification discloses a data retrieval method, device and electronic device for a power collection terminal, including storing the message data to be retrieved in multiple storage files according to a preset storage size; retrieving the multiple storage files to retrieve target data within the current retrieval cycle, and when the target data is not retrieved within the current retrieval cycle, recording the depth information of the retrieved current storage file, the depth information including the position information of the current message data in the retrieved current storage file; in the next retrieval cycle, determining the position information of the current message data in the current storage file according to the depth information, and retrieving the unretrieved storage files starting from the position information of the current message data in the current storage file, and repeating the previous step until all the target data are retrieved. The present invention can improve the retrieval speed, greatly reduce the retrieval time, and improve the accuracy of the retrieval.

[0004] Due to the huge amount of power data information and its wide variety, the existing search method that simply uses keywords is slow, which easily leads to a large amount of similar data that does not meet the needs, making it difficult for personnel to screen. On the other hand, it is difficult to fully retrieve various types of data related to the needs, making it difficult to timely and truly reflect the actual operating status of the equipment. Summary of the invention

[0005] In view of the above-mentioned prior art, the technical problem to be solved by the present invention is that when viewing power monitoring data, the existing retrieval means have problems such as large amount of data screening, difficulty in screening, incomplete retrieval results, etc., making it difficult to timely and truly reflect the operating status of the equipment.

[0006] In order to solve the above problems, the present invention provides a retrieval system for multi-source heterogeneous data fusion, including a data calibration fusion system and a calibration retrieval system carried in a retrieval terminal, the data calibration fusion system including a heterogeneous data receiving module, a diversified feature extraction module, a feature calibration module, a calibration total vocabulary and a time-sharing calibration database, the heterogeneous data receiving module is used to receive multiple types of source data transmitted from an equipment monitoring center, the calibration total vocabulary includes a normal vocabulary and an abnormal vocabulary;

[0007] The heterogeneous data receiving module is connected to an active data database, which includes an environment sensor database, an equipment sensor database, an electric power database and a monitoring database. The environment sensor database stores the environmental temperature and humidity data around the equipment obtained by real-time monitoring. The equipment sensor database stores the equipment temperature and humidity data obtained by real-time monitoring. The electric power database stores the equipment electric power data obtained by real-time monitoring, and the data includes voltage, current and power. The monitoring database stores the fixed-point thermal image data obtained by real-time shooting with an infrared thermal imager and the multi-point thermal image data obtained by dynamic inspection equipment inspection.

[0008] The calibration retrieval system includes a search word library, a same-word search module and a similar-word search module. The search word library includes a state word library and a device word library, and the state word library contains all the words in the calibration total word library.

[0009] As a further supplement to the present application, the method for using the data calibration fusion system includes the following steps:

[0010] Step 1: After receiving various types of source data transmitted from the equipment monitoring center, the diversified feature extraction module extracts and analyzes key features of all source data to determine whether the data is abnormal;

[0011] Step 2: According to the judgment result, at least one word is selected from the calibration total word library through the feature calibration module to perform state calibration on the data;

[0012] Step 3: According to the source data collection time and calibration results, the calibrated data are classified and packaged, and stored together in the time-sharing calibration database, that is, heterogeneous data in the same time period and with similar calibration status are grouped together for packaged storage.

[0013] A retrieval system for multi-source heterogeneous data fusion, wherein the search method comprises the following steps:

[0014] S1. When searching using the same-word search module, enter at least one word contained in the search word library in the interface, then selectively enter a time range, and after confirming the search, search the data in the time-sharing calibration database through the search module;

[0015] When the calibrated word of a data is the same as the input word, and the time range overlaps with the input time range, it means that the data meets the search requirements and will be displayed on the search result interface;

[0016] S2. When searching using the term search module, enter the text you want to express in the interface, optionally enter the time range, and then use the search module to search for data in the time-sharing calibration database;

[0017] When the calibrated term of a certain data has the same semantics as the input text content, or contains the same words, and the time range overlaps with the input time range, it indicates that the data meets the search requirements and will be displayed in the search result interface.

[0018] As a further supplement to the present application, the data calibration fusion system also includes a periodic recycling module, which is used to clean up obsolete data in the time-sharing calibration database and source data corresponding to the obsolete data in the source data library. The criterion for judging obsolete data is: the interval between the data entry and storage date and the current date exceeds X days.

[0019] As a further supplement to the present application, a historical database and a normal database are also provided in the retrieval terminal. The historical database stores the time range when the equipment undergoes normal data changes when affected by external factors and the data range after the changes, and the above data range includes the historical abnormal equipment temperature and humidity data range and the historical abnormal equipment power data range. The normal database stores the normal temperature and humidity data range and the normal power data range of the equipment when the equipment operates normally.

[0020] As a further supplement to the present application, in step 1, when determining whether the two source data, the device temperature and humidity data and the device power data, are abnormal, the following operations are performed in sequence:

[0021] Step 1-1, first compare the source data of the equipment temperature and humidity data and the equipment power data with the normal temperature and humidity data range of the equipment and the normal power data range of the equipment in the normal database, and judge whether they are abnormal. When the source data is within the above range, it indicates that the source data is normal. On the contrary, when the source data is not within the above range, proceed to step 1-2;

[0022] Step 1-2: When a source data has an anomaly, it is compared with the corresponding historical anomaly data in the historical database. When it is not within the range of historical anomaly data, it indicates that the received source data has an anomaly. On the contrary, when the source data is within the range of historical anomaly data and the collection time of the source data is within the time range in the historical database, it indicates that the source data is normal.

[0023] As a further supplement to the present application, the retrieval terminal is also provided with an initial thermal imaging database, which stores normal thermal images of the equipment when it is operating normally. In step 1, when determining whether the three source data of the ambient temperature and humidity data, the fixed-point thermal imaging data, and the multi-point thermal imaging data are abnormal, the following operations are performed:

[0024] Compare the ambient temperature and humidity data with the current meteorological data. If they are within the range of the meteorological data, it indicates that the ambient temperature and humidity data are normal. Otherwise, it indicates that the ambient temperature and humidity data are abnormal.

[0025] The fixed-point thermal imaging data and the multi-point thermal imaging data are compared with the normal thermal image. When the difference value exceeds the predetermined range, it indicates that there is an abnormality in the two. Otherwise, it indicates that the data of the two are normal.

[0026] As another improvement of the present application, the data calibration fusion system further includes a pre-self-check module, and the method for using the pre-self-check module includes the following steps:

[0027] After completing the data status calibration process in step 2, the calibration vocabulary of multiple source data in the same time period is checked through the pre-self-check module according to the collection time of the source data. When the check result shows that at least two heterogeneous source data are calibrated with vocabulary in the abnormal vocabulary, the retrieval terminal will issue an equipment risk warning and display the above heterogeneous source data on the retrieval terminal.

[0028] To summarize, by first analyzing a variety of heterogeneous source data, lexically calibrating their states, and grouping and storing a variety of heterogeneous source data within the same time period and with similar calibration states, a heterogeneous data group is formed. In the process of data query and retrieval, the required heterogeneous data group can be accurately and quickly retrieved according to the vocabulary. Compared with the existing retrieval method, the present application does not need to perform a separate search for a single type of data, and can quickly obtain different types of heterogeneous data that meet the needs. By integrating various heterogeneous data, the operating status of the equipment can be quickly known, and the equipment safety can be improved. At the same time, by setting a search vocabulary and a calibrated total vocabulary that reflect the data status, it can provide guidance for the retrieval process, effectively solving the problems of inaccurate retrieval vocabulary and time-consuming and laborious thinking about the retrieval content in the existing retrieval process. In addition, after calibrating the data status, the calibrated vocabulary of multiple source data within the same time period is checked. When a certain identical probability is reached, an equipment risk warning is issued, which can further improve the equipment safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is an overall system diagram of the first implementation method of this application;

[0030] Figure 2This is a system diagram of the source data repository of the first implementation mode of the present application;

[0031] Figure 3 This is a flow chart of calibrating multiple heterogeneous data in the first implementation mode of the present application;

[0032] Figure 4 This is a flow chart of searching for multiple power monitoring data in the first embodiment of the present application;

[0033] Figure 5 This is a system diagram for analyzing and judging a variety of heterogeneous data in the first embodiment of the present application;

[0034] Figure 6 This is a flow chart for performing risk warning on a device according to data status in the second implementation mode of the present application;

[0035] Figure 7 This is a tree diagram of this application. DETAILED DESCRIPTION

[0036] The following describes two implementation modes of the present application in detail with reference to the accompanying drawings.

[0037] The first implementation method:

[0038] The present invention provides a retrieval system for multi-source heterogeneous data fusion, please refer to Figure 1 , including a data calibration fusion system and a calibration retrieval system installed in a retrieval terminal, the data calibration fusion system including a heterogeneous data receiving module, a diversified feature extraction module, a feature calibration module, a calibration total vocabulary and a time-sharing calibration database, the heterogeneous data receiving module is used to receive various types of source data transmitted from a device monitoring center, and the calibration total vocabulary includes a normal vocabulary and an abnormal vocabulary;

[0039] See also Figure 3 and Figure 7 The method for using the data calibration fusion system comprises the following steps:

[0040] Step 1: After receiving various types of source data transmitted from the equipment monitoring center (including power data, temperature and humidity data, image data, etc., for specific classification, please refer to the source data library below), the diversified feature extraction module is used to extract and analyze key features of all source data to determine whether the data is abnormal;

[0041] Step 2: According to the judgment result, at least one word is selected from the calibration total word library through the feature calibration module to perform state calibration on the data;

[0042] When the source data is in a normal state, appropriate words are selected from the normal word library to calibrate the source data, such as "normal", "standard", etc. When the source data is in an abnormal state, appropriate words are selected from the abnormal word library to calibrate the source data, such as "overheated", "too low", etc. The specific words are selected appropriately according to the data type. For example, for power data such as voltage and current, "standard", "too low", "too high", "slightly low", etc., you can choose "normal", "overheated", etc. for sensor data such as temperature and humidity. This is common knowledge for those skilled in the art.

[0043] Step 3: According to the source data acquisition time and calibration results, the calibrated data is classified and packaged, and stored together in the time-sharing calibration database, that is, heterogeneous data in the same time period and with similar calibration status are grouped together for packaged storage;

[0044] For example, when a current data is calibrated as "too high" and a temperature data is calibrated as "overheated", these two calibration states both reflect data abnormalities, and the degree of abnormality is similar. At the same time, the collection time of these two data is in the same time period. At this time, the two data are grouped together and packaged and stored to form a heterogeneous data group (combined with Figure 3 shown).

[0045] See also Figure 2 , the heterogeneous data receiving module is connected to an active data database, and the source data database includes an environmental sensor database, an equipment sensor database, a power database and a monitoring database;

[0046] The environmental sensor database stores the environmental temperature and humidity data of the surrounding area of ​​the equipment obtained through real-time monitoring, which can be obtained through monitoring by the temperature and humidity sensor;

[0047] The device sensor database stores device temperature and humidity data obtained through real-time monitoring, which can also be obtained through monitoring by temperature and humidity sensors;

[0048] The power database stores the equipment power data obtained through real-time monitoring, including voltage, current and power, which not only directly reflects the supply of electric energy, but also serves as an important basis for evaluating the operating status of power equipment and the quality of power supply;

[0049] The monitoring database stores fixed-point thermal imaging data obtained through real-time shooting by infrared thermal imagers and multi-point thermal imaging data obtained through dynamic inspection equipment inspections. In the prior art, fixed-point installed infrared thermal imagers are generally used to perform infrared detection on power equipment to check whether there are temperature abnormalities. Due to the distribution characteristics of power equipment, obstructions may be formed between equipment. Fixed-point installed infrared thermal imagers are difficult to perform comprehensive detection on multiple devices. Therefore, in the actual monitoring process, multi-point thermal imaging data collected by dynamic inspection equipment (such as drones or inspection robots with infrared thermal imagers) can be used to conduct comprehensive and comprehensive monitoring of power equipment to obtain more accurate monitoring data.

[0050] Combination Figure 1 and Figure 4 As shown, the calibration retrieval system includes a search vocabulary, a same-word search module and a similar-word search module. The search vocabulary includes a status vocabulary and a device vocabulary, and the status vocabulary contains all the words in the calibration total vocabulary. In this way, when searching using the words in the search vocabulary, the source data calibrated with the same words can be accurately searched in the time-sharing calibration database.

[0051] See also Figure 4 , a retrieval system for multi-source heterogeneous data fusion, the search method of which comprises the following steps:

[0052] S1. When searching using the same-word search module, enter at least one word contained in the search word library in the interface, and then selectively enter a time range (when entering a time range, the data can be further screened to improve the search accuracy). After confirming the search, the search module searches for data in the time-sharing calibration database;

[0053] When the calibrated word of a data is the same as the input word, and the time range overlaps with the input time range, it means that the data meets the search requirements and will be displayed on the search result interface;

[0054] S2. When searching using the term search module, enter the text you want to express in the interface, optionally enter the time range, and then use the search module to search for data in the time-sharing calibration database;

[0055] When the calibrated term of a certain data has the same semantics as the input text content, or contains the same words, and the time range overlaps with the input time range, it indicates that the data meets the search requirements and will be displayed in the search result interface.

[0056] By first analyzing a variety of heterogeneous source data, then performing vocabulary calibration for the normality of the data, and grouping and packaging a variety of heterogeneous source data within the same time period and with similar calibration status for storage, a heterogeneous data group is formed. In the process of data query and retrieval, the required heterogeneous data group can be accurately and quickly retrieved according to the input search content. Compared with the existing retrieval method, the present application does not need to perform a separate search for a single type of data, and can quickly obtain different types of heterogeneous data that meet the needs. By combining various heterogeneous data, the operating status of the equipment can be quickly known, and the safety of the equipment can be improved. At the same time, by setting a search vocabulary library and a calibration total vocabulary library that reflect the data status, and the two are in an inclusion relationship, it can provide guidance for the retrieval process, and effectively solve the problems of inaccurate retrieval vocabulary and time-consuming and laborious thinking about retrieval content in the existing retrieval process.

[0057] The data calibration fusion system also includes a periodic recycling module, which is used to clean up obsolete data in the time-sharing calibration database and source data corresponding to the obsolete data in the source data library. The criterion for judging obsolete data is: the interval between the data entry and storage date and the current date exceeds X days. By regularly cleaning up the data, the data storage volume in the retrieval terminal is effectively reduced, and the operating efficiency of the retrieval terminal is improved. At the same time, during the data retrieval process, the accuracy and retrieval rate of data retrieval are also improved.

[0058] Combination Figure 5 As shown, the retrieval terminal is also provided with a historical database and a normal database. The historical database stores the time range when the equipment undergoes normal data changes when affected by external factors and the data range after the changes, and the above data range includes the historical abnormal equipment temperature and humidity data range and the historical abnormal equipment power data range. Due to various factors such as the external environment and the equipment's own operating load, the equipment's temperature and humidity data and power data sometimes also have normal changes. For example, when the external environment temperature rises, heat is transferred to the power equipment, which will cause the temperature of the equipment to rise as well. However, this temperature change is temporary and difficult to avoid. Alternatively, due to a sudden increase in equipment load in a certain period of time, the equipment current data increases, which is required for normal operation. Therefore, this type of normal change data is not calibrated as abnormal data. The normal database stores the equipment's normal temperature and humidity data range and the equipment's normal power data range when the equipment is operating normally.

[0059] Therefore, according to the above historical database and normal database, in step 1, when judging whether the two source data, the equipment temperature and humidity data and the equipment power data, are abnormal, the following operations are performed in sequence:

[0060] Step 1-1, first compare the source data of the equipment temperature and humidity data and the equipment power data with the normal temperature and humidity data range of the equipment and the normal power data range of the equipment in the normal database, and judge whether they are abnormal. When the source data is within the above range, it indicates that the source data is normal. On the contrary, when the source data is not within the above range, proceed to step 1-2;

[0061] Step 1-2, when a certain source data has an abnormality, it is compared with the corresponding historical abnormal data in the historical database. When it is not within the range of historical abnormal data, it indicates that the received source data has an abnormality; conversely, when the source data is within the range of historical abnormal data and the collection time of the source data is within the time range in the historical database, it indicates that the source data is normal, that is, as recorded in the above content, this is a short-term normal change required for the normal operation of the equipment or an unavoidable one.

[0062] Combination Figure 5 As shown, the retrieval terminal is also provided with an initial thermal imaging database, and the initial thermal imaging database stores normal thermal images of the equipment when it is operating normally. In step 1, when determining whether the three source data of the ambient temperature and humidity data, the fixed-point thermal imaging data, and the multi-point thermal imaging data are abnormal, the following operations are performed:

[0063] Compare the ambient temperature and humidity data with the current meteorological data. If they are within the range of the meteorological data, it indicates that the ambient temperature and humidity data are normal. Otherwise, it indicates that the ambient temperature and humidity data are abnormal.

[0064] The fixed-point thermal imaging data and the multi-point thermal imaging data are compared with the normal thermal image. When the difference value exceeds the predetermined range, it indicates that there is an abnormality in the two. Otherwise, it indicates that the data of the two are normal.

[0065] Second implementation method:

[0066] This implementation mode adds the following contents based on implementation mode 1: Figure 6 The data calibration fusion system further includes a pre-self-check module, and the method for using the pre-self-check module includes the following steps:

[0067] After completing the state calibration process of the data in step 2, the calibration vocabulary of multiple source data in the same time period is checked through the pre-self-check module according to the collection time of the source data. When the check result shows that at least two heterogeneous source data are calibrated with the vocabulary in the abnormal vocabulary, the retrieval terminal issues an equipment risk warning and displays the above heterogeneous source data on the retrieval terminal;

[0068] Using words in the abnormal vocabulary to calibrate the data indicates that the data has certain abnormal conditions. Therefore, in the same time period, when two or more different types of source data are abnormal, it can indicate to a certain extent that the power equipment has an abnormal condition or is about to have an abnormal condition. For example: in the same time period, the equipment temperature data is calibrated as "overheated" and the current data in the equipment power data is calibrated as "too high". Both data are abnormal, which indicates that the equipment is already in an overloaded operating state or is about to be overloaded and burned. Therefore, at this time, the pre-self-check module issues an equipment risk warning on the retrieval terminal to alert the monitoring personnel so that they can conduct further inspections on the equipment in a timely manner to further improve the safety of the equipment.

[0069] In view of current practical needs, the above-mentioned implementation mode adopted in this application is not limited to the scope of protection. Various changes made within the knowledge scope of technical personnel in this field without departing from the concept of this application still fall within the scope of protection of the present invention.

Claims

1. A retrieval system for multi-source heterogeneous data fusion, characterized by: It includes a data calibration fusion system and a calibration retrieval system installed in a retrieval terminal, wherein the data calibration fusion system includes a heterogeneous data receiving module, a diversified feature extraction module, a feature calibration module, a calibration total word library and a time-sharing calibration database, wherein the heterogeneous data receiving module is used to receive various types of source data transmitted from a device monitoring center, and the calibration total word library includes a normal word library and an abnormal word library; The heterogeneous data receiving module is connected to an active data database, which includes an environment sensor database, an equipment sensor database, an electric power database and a monitoring database. The environment sensor database stores the environmental temperature and humidity data around the equipment obtained by real-time monitoring. The equipment sensor database stores the equipment temperature and humidity data obtained by real-time monitoring. The electric power database stores the equipment electric power data obtained by real-time monitoring, and the data includes voltage, current and power. The monitoring database stores the fixed-point thermal image data obtained by real-time shooting with an infrared thermal imager and the multi-point thermal image data obtained by dynamic inspection equipment inspection. The calibration retrieval system includes a search word library, a same-word search module and a similar-word search module. The search word library includes a state word library and a device word library, and the state word library contains all the words in the calibration total word library.

2. A retrieval system for multi-source heterogeneous data fusion according to claim 1, characterized in that: The method for using the data calibration fusion system comprises the following steps: Step 1: After receiving various types of source data transmitted from the equipment monitoring center, the diversified feature extraction module extracts and analyzes key features of all source data to determine whether the data is abnormal; Step 2: According to the judgment result, at least one word is selected from the calibration total word library through the feature calibration module to perform state calibration on the data; Step 3: According to the source data collection time and calibration results, the calibrated data are classified and packaged, and stored together in the time-sharing calibration database, that is, heterogeneous data in the same time period and with similar calibration status are grouped together for packaged storage.

3. A retrieval system for multi-source heterogeneous data fusion according to claim 2, characterized in that: The search method includes the following steps: S1. When searching using the same-word search module, enter at least one word contained in the search word library in the interface, then selectively enter a time range, confirm the search, and search the data in the time-sharing calibration database through the search module; When the calibrated word of a data is the same as the input word, and the time range overlaps with the input time range, it means that the data meets the search requirements and will be displayed on the search result interface; S2. When searching using the term search module, enter the text you want to express in the interface, optionally enter the time range, and then use the search module to search for data in the time-sharing calibration database; When the calibrated term of a certain data has the same semantics as the input text content, or contains the same words, and the time range overlaps with the input time range, it indicates that the data meets the search requirements and will be displayed in the search result interface.

4. The retrieval system for multi-source heterogeneous data fusion according to claim 1, characterized in that: The data calibration fusion system also includes a periodic recycling module, which is used to clean up obsolete data in the time-sharing calibration database and source data corresponding to the obsolete data in the source data library. The criterion for judging obsolete data is: the interval between the data entry and storage date and the current date exceeds X days.

5. The retrieval system for multi-source heterogeneous data fusion according to claim 2, characterized in that: The retrieval terminal is also provided with a historical database and a normal database. The historical database stores the time range when the equipment undergoes normal data changes when affected by external factors and the data range after the changes, and the above data range includes the historical abnormal equipment temperature and humidity data range and the historical abnormal equipment power data range. The normal database stores the normal temperature and humidity data range and the normal power data range of the equipment when the equipment is operating normally.

6. A retrieval system for multi-source heterogeneous data fusion according to claim 5, characterized in that: In step 1, when determining whether the two source data, the device temperature and humidity data and the device power data, are abnormal, the following operations are performed in sequence: Step 1-1, first compare the source data of the equipment temperature and humidity data and the equipment power data with the normal temperature and humidity data range of the equipment and the normal power data range of the equipment in the normal database, and judge whether they are abnormal. When the source data is within the above range, it indicates that the source data is normal. On the contrary, when the source data is not within the above range, proceed to step 1-2; Step 1-2: When a source data has an anomaly, it is compared with the corresponding historical anomaly data in the historical database. If it is not within the range of the historical anomaly data, it indicates that the received source data has an anomaly; On the contrary, when the source data is within the range of historical abnormal data and the collection time of the source data is within the time range in the historical database, it indicates that the source data is normal.

7. The retrieval system for multi-source heterogeneous data fusion according to claim 2, characterized in that: The retrieval terminal is also provided with an initial thermal imaging database, which stores normal thermal images of the equipment when it is operating normally. In step 1, when determining whether the three source data of the ambient temperature and humidity data, the fixed-point thermal imaging data and the multi-point thermal imaging data are abnormal, the following operations are performed: Compare the ambient temperature and humidity data with the current meteorological data. If they are within the range of the meteorological data, it indicates that the ambient temperature and humidity data are normal. Otherwise, it indicates that the ambient temperature and humidity data are abnormal. The fixed-point thermal imaging data and the multi-point thermal imaging data are compared with the normal thermal image. When the difference value exceeds the predetermined range, it indicates that there is an abnormality in the two. Otherwise, it indicates that the data of the two are normal.

8. The retrieval system for multi-source heterogeneous data fusion according to claim 2, characterized in that: The data calibration fusion system further includes a pre-self-check module, and the method for using the pre-self-check module includes the following steps: After completing the data status calibration process in step 2, the calibration vocabulary of multiple source data in the same time period is checked through the pre-self-check module according to the collection time of the source data. When the check result shows that at least two heterogeneous source data are calibrated with vocabulary in the abnormal vocabulary, the retrieval terminal will issue an equipment risk warning and display the above heterogeneous source data on the retrieval terminal.

Images