Meteorological disaster early warning information quality control method and system

Abstract

The invention discloses a meteorological disaster early warning information quality control method and system. The method comprises the following steps: extracting a symbol part of meteorological disaster early warning information; obtaining a feature symbol of the meteorological disaster early warning information, and dividing the meteorological disaster early warning information into title information and text information according to the feature symbol; judging whether the symbol part has messy codes or not by utilizing a judgment device; respectively detecting whether a difference value between the time information in the title information and the time information of the real world exceeds a first threshold value or not and whether the early warning level is matched with the text information or not; and obtaining a keyword in the text information through an early warning information template, and judging whether the keyword is null or not and whether the keyword is correct or not. According to the invention, the quality inspection efficiency and accuracy can be balanced.

Description

Technical Field

[0001] The present invention generally relates to the field of quality control, and more specifically, to a method and system for quality control of meteorological disaster warning information. Background Art

[0002] The quality control of meteorological disaster warning information is the soul and core of the meteorological disaster warning issuance process. The warning data quality control technology aims to identify problems such as erroneous data and incomplete data and take effective measures to ultimately ensure that users obtain accurate data.

[0003] At present, the quality control technology of meteorological disaster warning information mainly involves automatic Chinese proofreading of meteorological warning information, and automatic word segmentation technology restricts the development of Chinese automatic proofreading technology. However, word segmentation technology has many shortcomings. First, the current mainstream word segmentation technology relies on deep learning models (such as BERT, BiLSTM-CRF, etc.), and its training and reasoning processes consume a lot of computing resources. For example, the model based on the Transformer architecture needs to process massive parameters in parallel during the training phase and relies on high-performance GPU clusters. However, in actual deployment, the large number of model parameters leads to high memory usage and is difficult to run on edge devices (such as meteorological monitoring terminals) or in low-power scenarios. In addition, weather disaster warning information usually requires real-time processing of bursty data streams. Traditional models cannot efficiently cope with peak loads due to the exponential growth of computational complexity. If lightweight model compression technology is used, the ability to recognize professional terms may be reduced due to the sacrifice of model capacity, resulting in a contradiction between resources and performance. Second, the weather disaster warning scenario has extremely high requirements for the timeliness of information processing, but the existing word segmentation technology generally has response delay problems. The rule-based word segmentation method needs to traverse a large number of predefined dictionaries and regular expressions, and the matching efficiency drops sharply when facing long texts (such as cross-regional disaster warning reports); although statistical models (such as HMM, CRF) improve efficiency through probability optimization, they still need to backtrack calculations when processing nested structures or unregistered words, which increases time consumption. Taking a provincial meteorological platform test as an example, the word segmentation system based on deep learning takes an average of more than 200 milliseconds to process a single warning message. When the number of concurrent requests surges, the system delay can reach seconds, seriously affecting the timeliness of warning release. Third, the general word segmentation model is prone to semantic misjudgment due to domain characteristics in weather disaster warning scenarios. Although the existing word segmentation technology performs well in general scenarios, it exposes core defects such as imbalance between resource consumption and performance, prominent response delays, misjudgment of professional terms, and disconnection from domain knowledge when faced with the high real-time, strong domain nature, and complex semantic structure of weather disaster warnings.

[0004] In summary, it is urgent to design a quality control solution that integrates the meteorological knowledge base, lightweight architecture and the characteristics of meteorological disaster warning information, so as to improve the efficiency and accurate analysis of disaster warning texts and provide reliable technical support for disaster prevention decision-making. Summary of the invention

[0005] In order to at least solve the technical problems described in the above background technology section, the present invention proposes a method for quality control of meteorological disaster warning information, which can improve the recognition efficiency and accuracy of meteorological disaster warning information. In view of this, the present invention provides solutions in the following aspects.

[0006] The first aspect of the present invention provides a method for controlling the quality of meteorological disaster warning information, comprising: extracting the symbol part of the meteorological disaster warning information; the symbol part includes the remaining character information except the text; obtaining the characteristic symbol of the meteorological disaster warning information, and dividing the meteorological disaster warning information into title information and text information according to the characteristic symbol; the title information includes information related to time and warning level, and the text information includes keywords describing the disaster situation; using a judgment device to judge whether the symbol part has garbled characters; if it contains garbled characters, the meteorological disaster warning information is judged to be unqualified; respectively detecting whether the difference between the time information in the title information and the time information in the real world exceeds a first threshold, and whether the warning level matches the text information; if the difference exceeds the first threshold or the warning level does not match, the meteorological disaster warning information is judged to be unqualified; obtaining the keywords in the text information through the warning information template, judging whether the keywords are empty, and whether the keywords are correct; if the keywords contain empty characters, or the keywords are wrong, the meteorological disaster warning information is judged to be unqualified.

[0007] In one embodiment, it also includes creating a warning information template: performing logical operations on multiple similar meteorological disaster warning information to generate the warning information template of this type, wherein the template includes a text template and a symbol template, which are used to represent the unchanged parts of multiple similar meteorological disaster warning information.

[0008] In one embodiment, performing a logical operation on a plurality of similar meteorological disaster warning information specifically includes taking the intersection of the plurality of similar meteorological disaster warning information and extracting information with an occurrence frequency greater than a second threshold as template information.

[0009] In one embodiment, using a judgment device to judge whether the symbol part contains garbled characters includes: converting the symbol part and the symbol template into digital signals through ASCII codes respectively and inputting them into the judgment device; using the judgment device to perform a subtraction operation on the symbol part and the symbol template to obtain a calculation result; and checking whether the calculation result contains a digital signal of an unexpected symbol.

[0010] In one embodiment, the judgment device can be implemented based on hardware or software, and its principle circuit is as follows, including: a first to a fifth resistor, a first operational amplifier and a second operational amplifier, one end of the first resistor is connected to a first input signal, the other end of the first resistor is connected to the inverting input of the first operational amplifier and one end of the second resistor, the non-inverting input of the first operational amplifier is grounded, the other end of the second resistor is connected to the output of the first operational amplifier and one end of the third resistor, the other end of the third resistor is connected to one end of the fourth resistor, the inverting input of the second operational amplifier and one end of the fifth resistor, the other end of the fourth resistor is connected to the second input signal, the other end of the fifth resistor is connected to the output of the second operational amplifier, and the non-inverting input of the second operational amplifier is grounded.

[0011] In one embodiment, the characteristic symbol is a colon.

[0012] In one embodiment, detecting whether the difference between the time information in the title information and the time information in the real world exceeds a first threshold value also includes: obtaining the current real time, the real time is a 24-hour time, including year, month, day and specific time; obtaining the extraction time in the title information, the extraction time is a 24-hour time, including year, month, day and specific time; subtracting the extraction time from the real time to determine whether the difference is greater than zero and less than the first threshold value.

[0013] In one embodiment, determining whether the keyword is empty and whether the keyword is correct includes: calculating the difference between the warning information template and the disaster warning information to obtain a key value; inserting a separator between the difference; determining whether the separators are empty, and if so, determining that the disaster warning information is unqualified; and checking whether the keyword appears unrelated to the weather information, and if so, determining that the disaster warning information is unqualified.

[0014] In one embodiment, it also includes using a sensitive word library to check whether the weather warning information contains sensitive information, and if it contains sensitive word information, the above information is judged to be unqualified.

[0015] A second aspect of the present invention provides a meteorological disaster warning information quality control system, which runs any of the meteorological disaster warning information quality control methods described above.

[0016] The present invention divides the warning information into symbols, text, and title according to the characteristics of meteorological disaster warning information, and adopts different detection methods for each part, so that the overall quality control can not only focus on format errors, but also cover the correctness of the content, while improving the efficiency and accuracy of detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0018] Figure 1 It is a method for controlling the quality of early warning information according to an embodiment of the present invention;

[0019] Figure 2 is a schematic diagram showing early warning information processing according to an embodiment of the present invention;

[0020] Figure 3 The figure shows the principle of the judgment device according to the embodiment of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0022] It should be understood that the terms "first", "second", "third" and "fourth" etc. in the claims, specifications and drawings of the present invention are used to distinguish different objects rather than to describe a specific order. The terms "include" and "comprise" used in the specification and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their collections.

[0023] It should also be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the claims, the singular forms of "a", "an" and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should also be further understood that the term "and / or" used in the specification of the present invention and the claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations.

[0024] As used in this specification and claims, the term "if" may be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" may be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0025] The key elements of meteorological disaster warning information generally include: warning issuing unit, warning category, warning level, warning start time, warning end time, filing (whether), warning title, warning content, test warning, etc.

[0026] Among them, different sources of meteorological disaster warning information generally have different warning information release standard formats, for example:

[0027] a) Provincial, state and municipal warning formats: Level IV, Level III, Level II, and Level I, with Level I being the highest level.

[0028] For example: XXX Meteorological Observatory issued a heavy rain level IV warning at XX:XX:XX on XX / XX / 20XX: It is predicted that in the next 24 hours our province will usher in a heavy rainfall weather process from north to south, and heavy rain will occur in XX Prefecture, XX City, XX Prefecture, XX City, and XX City. Please pay attention to the drainage of roads, farmland, and fish ponds, and the public should pay attention to safety when going out.

[0029] b) County-level meteorological warning signal formats: blue, yellow, orange, and red, with red being the highest level

[0030] For example: XXX County Meteorological Observatory issued a yellow rainstorm warning signal at XX:XX:XX on ...

[0031] c) Geological disaster meteorological risk warning format: four levels: IV, III, II, and I, with I being the highest level.

[0032] For example: XXX Department of Natural Resources and XXX Meteorological Bureau jointly issued a Level II warning for geological disaster meteorological risk at XX:XX:XX:XX on XX:XX:XX:XXXX: Affected by heavy rainfall, it is expected that in the next XX hours, the geological disaster meteorological risk level in XX Prefecture, XX City, and XX City will be Level II (high risk). Geological disasters such as landslides, mudslides, and collapses may occur, so please pay attention to prevention.

[0033] d) Flash flood disaster risk warning format: four levels: IV, III, II, and I, with I being the highest level.

[0034] For example: XXX Water Resources Department and XXX Meteorological Bureau jointly issued a flash flood disaster risk level II warning at XX:XX:XX:XX on XX:XX:XX:XXXX: Affected by heavy rainfall, it is expected that in the next XX hours, the flash flood disaster risk level in XX Prefecture, XX City, and XX City will be Level II (high risk). Flash flood disasters may occur, so please pay attention to prevention.

[0035] e) Forest fire meteorological risk warning format: There are four levels: Level III, Level II, and Level I, with Level I being the highest level.

[0036] For example: XXX Meteorological Observatory issued a forest fire weather level II warning at XX:XX:XX on XX / XX / 20XX: It is expected that in the next three days, XX Prefecture, XX City, XX City, XX City, and XX City will be sunny and hot with little rain. The forest fire weather level will reach an extremely high danger level, and forest fires are likely to occur. Please take precautions.

[0037] It can be found from the above examples that meteorological disaster warnings have special format restrictions. For example, the above information all has a separator, that is, the warning information is divided into two parts by a colon (:). One part is the title content, which briefly introduces the issuing unit, release time, and warning level of the warning; the other part is the text content, which is mainly used to describe the weather conditions in detail. In the present invention, different focus points are divided according to the characteristics of meteorological disaster warning information: that is, the release time and warning level of the information are concerned in the title part; the correctness of the keywords of the information is concerned in the text part. In addition, according to the characteristics of symbol problems that are prone to occur when filling in warning information in actual work, such as multiple spaces or garbled problems, the present invention adopts the method of extracting all non-text characters to obtain punctuation marks to check whether the symbols in the warning information are used correctly. In the prior art, software is generally used to exclude garbled symbols, and its reliability and speed are low. The present invention also adopts a circuit processing method, which is detected by transcoding to ASCII code, thereby improving the detection speed and accuracy.

[0038] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0039] The first aspect of the present invention provides a method for controlling the quality of meteorological disaster warning information. Figure 1 and attached Figure 2 The meteorological disaster warning information quality control method of the present invention is described below. Figure 1 Schematic diagram of a method for controlling quality of meteorological disaster warning information according to an embodiment of the present invention. Figure 2 This is a schematic diagram of early warning information processing in the present invention. Figure 1The disaster warning information quality control method of the present invention comprises:

[0040] Step S100: extracting the symbol part of the meteorological disaster warning information.

[0041] Specifically, the symbol part includes the remaining character information except the text.

[0042] Step S200: Acquire characteristic symbols of the meteorological disaster warning information, and divide the meteorological disaster warning information into title information and text information according to the characteristic symbols.

[0043] Specifically, the title information includes information related to time and warning level, and the text information includes keywords describing the disaster situation.

[0044] Step S300: using a judging device to judge whether the symbol part contains garbled characters.

[0045] Specifically, if garbled characters are included, the meteorological disaster warning information is judged to be unqualified;

[0046] Step S400: Detecting whether the difference between the time information in the title information and the time information in the real world exceeds a first threshold, and whether the warning level matches the text information.

[0047] Specifically, if the difference exceeds a first threshold or the warning level does not match, the meteorological disaster warning information is judged to be unqualified;

[0048] Step S500: obtaining keywords in the text information through the warning information template, and determining whether the keywords are empty and whether the keywords are correct.

[0049] Specifically, if the keyword contains space, or the keyword is wrong, it is determined that the meteorological disaster warning information is unqualified.

[0050] According to the attached Figure 2 In the present invention, the original meteorological disaster warning information is processed in two ways at the same time. On the one hand, the symbols in the warning information are extracted to check the correctness of the symbols; on the other hand, the title content and the text content in the warning information are extracted, and different contents are judged according to different judgment focuses. For example, for the title part, since the issuing unit usually does not change frequently, the judgment of the present invention is not performed, but the important time information and warning level are extracted and judged; and artificial intelligence (or manual) judgment is performed on the weather-related keywords in the text. The present invention saves computing resources by reasonably dividing the information and improves the correctness of the format and content in information detection.

[0051] In a preferred embodiment of the present invention, it also includes creating a warning information template: performing logical operations on multiple similar meteorological disaster warning information to generate the warning information template of this type, wherein the template includes a text template and a symbol template, which are used to represent the unchanged parts of multiple similar meteorological disaster warning information.

[0052] In specific implementation, templates can be set for different warning information categories, including provincial, state and city-level warnings, county-level meteorological warnings, geological disaster meteorological risk warnings, flash flood disaster risk warnings, and forest fire meteorological risk warnings, so as to speed up information processing. When obtaining the template, multiple similar meteorological disaster warning information is processed by performing logical operations and taking the intersection, and the information with a frequency of occurrence greater than the second threshold is extracted as the template information. The second threshold is, for example, characters with a frequency of occurrence greater than 90%. In the present invention, by setting the template and the subsequent difference (or subtraction) operation, only key information is extracted for word meaning judgment, which greatly reduces the content of the operation, thereby saving resource consumption for subsequent processing.

[0053] In a preferred embodiment of the present invention, using a judgment device to judge whether the symbol part contains garbled characters includes: converting the symbol part and the symbol template into digital signals via ASCII codes respectively and inputting them into the judgment device; using the judgment device to perform a subtraction operation on the symbol part and the symbol template to obtain an operation result; and checking whether the operation result contains a digital signal of an unexpected symbol.

[0054] In the present invention, considering that characters are prone to include empty codes, garbled codes, and other unexpected situations, errors are prone to occur through software processing, but the above scenarios can be correctly processed by converting to ASCII code, and the processing speed is fast. In a specific implementation, for example, the digital signal corresponding to the ASCII code of a comma is 00101100, the digital signal corresponding to the ASCII code of a period is 00101110, and the digital signal corresponding to the ASCII code of a colon is 00111010. When a combination of the above three digital signals appears in the symbol operation result extracted from the warning information, it means that an error has occurred in the symbol in the warning information.

[0055] In a preferred embodiment of the present invention, the above-mentioned judgment device can be implemented based on hardware or software, and its principle circuit is as shown in the attached Figure 3 As shown, including:

[0056] The first to fifth resistors, the first operational amplifier and the second operational amplifier, one end of the first resistor is connected to the first input signal, the other end of the first resistor is connected to the inverting input of the first operational amplifier and one end of the second resistor, the non-inverting input of the first operational amplifier is grounded, the other end of the second resistor is connected to the output of the first operational amplifier and one end of the third resistor, the other end of the third resistor is connected to one end of the fourth resistor, the inverting input of the second operational amplifier and one end of the fifth resistor, the other end of the fourth resistor is connected to the second input signal, the other end of the fifth resistor is connected to the output of the second operational amplifier, and the non-inverting input of the second operational amplifier is grounded.

[0057] According to the attached Figure 3 , we can get the input-output relationship of the circuit as:

[0058]

[0059] In a preferred embodiment of the present invention, detecting whether the difference between the time information in the title information and the time information in the real world exceeds a first threshold value also includes: obtaining the current real time, the real time is a 24-hour time, including year, month, day and specific time; obtaining the extracted time in the title information, the extracted time is a 24-hour time, including year, month, day and specific time; subtracting the extracted time from the real time to determine whether the difference is greater than zero and less than the first threshold value.

[0060] In the specific implementation, the release time of the warning information should be an event that has already occurred, so the above difference should be greater than zero; and the time interval between the time when the meteorological disaster occurs and the release of the warning information is required to be less than the first threshold. For example, if a meteorological disaster warning is required to be issued within 2 hours, the first threshold is set to 2 hours. Meteorological warnings have specific standards, and it is necessary to check whether the title warning level matches the specific weather conditions in the text. For example, the rainfall will reach more than 50 mm within 12 hours, or it has reached more than 50 mm and the rainfall may continue to be judged as a red warning for heavy rain. The above inspection can be used to determine the correctness of the meteorological warning content.

[0061] In a preferred embodiment of the present invention, judging whether the keyword is empty and whether the keyword is correct includes: calculating the difference between the warning information template and the disaster warning information to obtain a key value; inserting a separator between the difference; judging whether the separators are empty, if the separators are empty, judging that the disaster warning information is unqualified; checking whether the keyword appears unrelated to the weather information, if it appears unrelated to the weather information, judging that the disaster warning information is unqualified.

[0062] In specific implementations, checking keywords can be done by using a preset vocabulary, such as creating a meteorological disaster warning database, and checking whether the keywords are included in the professional terms of the meteorological disaster warning database; it can also be done through artificial intelligence or by manual judgment.

[0063] In a preferred embodiment of the present invention, it also includes using a sensitive word library to check whether the weather warning information contains sensitive information, and if sensitive word information is contained, the above information is judged to be unqualified.

[0064] It is understandable that the above-mentioned quality control method only describes the key points of the present invention, and can be expanded or deleted in combination with other quality control objectives and quality control methods as needed during specific implementation.

[0065] The second aspect of the present invention discloses a meteorological disaster warning information quality control system, which utilizes the meteorological disaster warning information quality control method disclosed above.

[0066] Although this specification has shown and described a plurality of embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will appreciate many changes, modifications and alternatives without departing from the thought and spirit of the present invention. It should be understood that in the process of practicing the present invention, various alternatives to the embodiments of the present invention described herein may be adopted. The attached claims are intended to define the scope of protection of the present invention, and therefore cover the modular compositions, equivalents or alternatives within the scope of these claims.

Claims

1. A method for quality control of meteorological disaster warning information, characterized in that: include: Extracting the symbol part of the meteorological disaster warning information; the symbol part includes the remaining character information except the text; Acquire characteristic symbols of the meteorological disaster warning information, and divide the meteorological disaster warning information into title information and text information according to the characteristic symbols; the title information includes information related to time and warning level, and the text information includes keywords describing the disaster situation; Using a judging device to judge whether the symbol part contains garbled characters; if the garbled characters are contained, judging that the meteorological disaster warning information is unqualified; Respectively detecting whether the difference between the time information in the title information and the time information in the real world exceeds a first threshold, and whether the warning level matches the text information; if the difference exceeds the first threshold or the warning level does not match, judging that the meteorological disaster warning information is unqualified; The keywords in the text information are obtained through the warning information template to determine whether the keywords are empty and whether the keywords are correct; if the keywords contain empty or the keywords are wrong, the meteorological disaster warning information is determined to be unqualified.

2. The meteorological disaster warning information quality control method according to claim 1 is characterized in that: It also includes templates for creating warning messages: A logical operation is performed on a plurality of similar meteorological disaster warning information to generate the warning information template of the same type, wherein the template includes a text template and a symbol template, and is used to represent the unchanged parts of the plurality of similar meteorological disaster warning information.

3. The meteorological disaster warning information quality control method according to claim 2 is characterized in that: The performing logical operations on the plurality of similar meteorological disaster warning information specifically includes taking the intersection of the plurality of similar meteorological disaster warning information and extracting the information with the occurrence frequency greater than the second threshold as the template information.

4. The meteorological disaster warning information quality control method according to claim 2 is characterized in that: Using the judging device to judge whether the symbol part has garbled characters comprises: The symbol part and the symbol template are converted into digital signals through ASCII codes and input into the judgment device; Using the judgment device to perform a subtraction operation on the symbol part and the symbol template to obtain a calculation result; Check whether the operation result contains a digital signal with an unexpected sign.

5. The meteorological disaster warning information quality control method according to claim 3 is characterized in that: The judging device can be implemented based on hardware or software, and its principle circuit is as follows, including: The first to fifth resistors, the first operational amplifier and the second operational amplifier, one end of the first resistor is connected to the first input signal, the other end of the first resistor is connected to the inverting input of the first operational amplifier and one end of the second resistor, the non-inverting input of the first operational amplifier is grounded, the other end of the second resistor is connected to the output of the first operational amplifier and one end of the third resistor, the other end of the third resistor is connected to one end of the fourth resistor, the inverting input of the second operational amplifier and one end of the fifth resistor, the other end of the fourth resistor is connected to the second input signal, the other end of the fifth resistor is connected to the output of the second operational amplifier, and the non-inverting input of the second operational amplifier is grounded.

6. The meteorological disaster warning information quality control method according to claim 1 is characterized in that: The characteristic symbol is a colon.

7. The meteorological disaster warning information quality control method according to claim 6 is characterized in that: Detecting whether the difference between the time information in the title information and the time information in the real world exceeds a first threshold value further includes: Get the current real time, which is a 24-hour system, including year, month, day and specific time; Obtain the extraction time in the title information, where the extraction time is a 24-hour time format, including year, month, day and specific time; subtract the extraction time from the real time to determine whether the difference is greater than zero and less than a first threshold.

8. The method for controlling the quality of meteorological disaster warning information according to claim 2, characterized in that: Determining whether the keyword is empty and whether the keyword is correct includes: Calculating the difference between the warning information template and the disaster warning information to obtain a key value; Insert a separator between the differences; Determine whether the separators are empty, if so, determine that the disaster warning information is unqualified; Check whether the keywords appear to be irrelevant to the weather information. If they appear to be irrelevant to the weather information, it is determined that the disaster warning information is unqualified.

9. The method for controlling the quality of meteorological disaster warning information according to any one of claims 1 to 8, characterized in that: It also includes using a sensitive word library to check whether the weather warning information contains sensitive information. If sensitive word information is contained, the above information is judged to be unqualified.

10. A meteorological disaster warning information quality control system, characterized in that: Run the meteorological disaster warning information quality control method as described in any one of claims 1-9.

Images