WO2020111934A1

WO2020111934A1 - A method and system for detection of natural disaster occurrence

Info

Publication number: WO2020111934A1
Application number: PCT/MY2019/050101
Authority: WO
Inventors: Amru Yusrin BIN AMRUDDIN; Wooi Kin Goon; May Fern KOH; Muhammad Awis Jamaluddin BIN JOHARI; Mohd Marzuq Ikram BIN MOHD HELMI
Original assignee: Mimos Berhad
Priority date: 2018-11-29
Filing date: 2019-11-26
Publication date: 2020-06-04

Abstract

A system (100) and method to detect an occurrence of a natural disaster comprises of a real time visioning analytic module (110) and a sensor analyser module (120), configured to collect real time data and sensing data on the natural disaster from a target area and from areas surrounding the target area; and an information collector and analyser module (130), configured to collect data related to the natural disaster occurrence in the target area and the areas surrounding the target area from a plurality of data sources. The system (100) further comprises a deep learning analytic module (140), configured to retrieve and process the collected data for determining a natural disaster score on the occurrence of the natural disaster in the target area and the areas surrounding the target area.

Description

A METHOD AND SYSTEM FOR DETECTION OF NATURAL DISASTER

OCCURRENCE

FIELD OF THE INVENTION

This invention relates to a detection of natural disaster occurrence and, more particularly, to a method and system for monitoring, early detection and forecasting of the natural disaster occurrence. BACKGROUND OF THE INVENTION

Natural disasters are unexpected events that can destroy people’s lives and their livelihoods. For an instance, flood is an overflow of water that submerge dry land and is a hazard in many areas in the world. The floods are caused due to water flushed from river, lake or pond usually due to the heavy rains.

A reliable flood detection system is necessary to provide advance warning to assist in protecting life and property. Most of the current flood detection system use weather sensing systems which are based on wind, humidity, cloud observations, rain and temperature measurements. The weather sensing systems includes rain gauge, Doppler radar and satellite telemetry. The rain gauge measures only continuous precipitation at specific locations. The Doppler radar works well only with large-scale weather features such as frontal systems. Further, the Doppler radar is limited to flat terrain, because radar coverage is restricted by beam blockage in mountainous areas. The satellite telemetry is only a representative of cloud coverage and not actual precipitation at ground level. All of these sensing systems require a processing module to translate sensed data into reliable flooding forecasts. However, none of these provide any real-time indication about the actual state of a flood situation and are thus generally ineffective for detecting and predicting the floods.

There are several prior arts disclosing the visual based system for automatically detecting the flood conditions, some of which are listed below for reference. The patent application number US20170193305A1 discloses a system for monitoring and detection of flash flooding events. The system comprising a plurality of visual markers for placement on open area ground surfaces; a plurality of video cameras for obtaining captured visual images of at least one of the visual markers; a plurality of video analytics units for analyzing the captured visual images of the visual markers and for detecting surface water covering of one or more of the visual markers; and a logic unit, for correlating data from at least one of the video analytics units and at least one of the video cameras, for relating surface water distributions on at least one of the visual markers to at least one flash flooding condition, and for issuing at least one notification relating to the flash flooding condition.

Another cited prior art, US 20170277954A1 discloses a system for early detecting disasters based on an Support Vector Machine (SVM) comprising an input unit configured to decode a plurality of input images and convert the decoded images into shared data; a shared data management unit configured to manage the shared data provided from the input unit; a processing unit configured to analyze the shared data provided from the shared data management unit based on an SVM learning algorithm, and detect whether a disaster situation occurred; and an output unit configured to output the detection result of the processing unit.

There are few existing technologies to detect the flood situations in real time by providing the sensor information for automatic processing. However, the aforementioned technologies are not based on imaging sensing devices and automated analytic methods. The imaging sensing devices when coupled with the analytics offers the advantage of not only automatically detecting the flood conditions visually for early warning, but can also be used simultaneously and subsequently to visually inspect the situation in real time. Therefore, there is a need for an effective visual based system for accurately monitoring and predicting the flood conditions.

None of the above-cited prior arts discloses a system and a method for high accuracy automated natural disaster detection in a target area using deep learning analytics and to issue an alert to areas surrounding the target area. SUMMARY OF THE INVENTION

The present invention provides a computer implemented method to detect an occurrence of a natural disaster, the method comprises the steps to: collect, by a real time visioning analytic module and a sensor analyser module, real time data and sensing data related to the natural disaster from a target area and from areas surrounding the target area; collect, by an information collector and analyser module, data related to the natural disaster occurred in the target area and the areas surrounding the target area from a plurality of data sources, wherein the data sources include data from a web media, data from a national weather department, data from a national irrigation and drainage department and historical data; and retrieve and process, by a deep learning analytic module, all the collected data, characterised in that, process all the collected data comprises the steps to: assign a value to each of the retrieved data from the real time visioning analytic module, the sensor analyser module and the information collector and analyser module based on the occurrence of the natural disaster; compute a first score by averaging the assigned values of the data retrieved from the information collector and analyser module, whereby the first score indicates whether the natural disaster has occurred in the areas surrounding the target area; and compute a natural disaster score by summing the first score and the assigned values of the data retrieved from the real time visioning analytic module and the sensor analyser module, whereby if the natural disaster score is greater than a pre determined value, it indicates the natural disaster occurred in the target area and if the natural disaster score is less than or equal to the pre-determined value, it indicates the natural disaster has not occurred in the target area.

Preferably, the steps to collect real time data related to the natural disaster by the real time visioning analytic module includes: capture an image of the target area and the areas surrounding the target area; and analyse the captured image in every predetermined time interval, to monitor whether the natural disaster has occurred in the target area. Preferably, the steps to collect the sensing data related to the natural disaster by the sensor analyser module includes: retrieve the sensing data of the environment in the target area and the areas surrounding the target area; and analyse the retrieved data in every predetermined time interval to detect changes in the retrieved data.

Preferably, the steps to analyse the captured image by the real time visioning analytic module comprises the steps to: translate basic shape of the captured image into an ontology based image, wherein the ontology based image is represented by a series of color codes with different percentage; extract the color code used in the ontology based image; sort the color code based on its percentage; compare the ontology based image with the captured image; and determine if there is a significant change in the basic shape and the color code percentage in the image based on the comparison result.

Preferably, the method further comprises the steps to: retrieve, by the deep learning analytic module, location details of the target area and the areas surrounding the target area from a location service provider module; calculate, by the deep learning analytic module, distance between the target area and the areas surrounding the target area based on the analysis of the natural disaster occurrence; and average the calculated distance, by the deep learning analytic module, to determine distance of the areas affected by the natural disaster surrounding the target area.

Preferably, the method further comprises the steps to broadcast a low alert signal, by the output module along with the calculated distance of the areas affected due to the natural disaster surrounding the target area, if the natural disaster score is less than or equal to one, wherein this condition indicates the natural disaster is about to occur in the target area; and broadcast a high alert signal, by the output module, if the natural disaster score is greater than one, wherein this condition indicates the natural disaster has occurred in the target area.

The present invention also provides a system to detect an occurrence of natural disaster comprises: a real time visioning analytic module, configured to collect real time data related to the natural disaster from a target area and from areas surrounding the target area; a sensor analyser module, configured to collect real time sensing data on the natural disaster from the target area and from the areas surrounding the target area; an information collector and analyser module, configured to collect the data related to the natural disaster occurred in the target area and the areas surrounding the target area from a plurality of data sources, wherein the data sources include data from a web media, data from a national weather department, data from a national irrigation and drainage department and historical data; and a deep learning analytic module configured to retrieve and process all the collected data, characterised in that, the deep learning analytic module configured to assign a value to each of the retrieved data from the real time visioning analytic module, the sensor analyser module and the information collector and analyser module based on the occurrence of the natural disaster; compute a first score by averaging the assigned values of the data retrieved from the information collector and analyser module, whereby the first score indicates whether the natural disaster has occurred in the areas surrounding the target area; and compute a natural disaster score by summing the first score and the assigned values of the data from the real time visioning analytic module and from the sensor analyser module, whereby if the natural disaster score is greater than a pre-determined value it indicates the natural disaster occurred in the target area and if the natural disaster score is less than or equal to the pre-determined value, it indicates the natural disaster has not occurred in the target area.

Preferably, the real time visioning analytic module is configured to capture an image of the target area and the areas surrounding the target area; and, analyse the captured image in every predetermined time interval, to monitor whether the natural disaster has occurred in the target area.

Preferably, the sensor analyser module is configured to retrieve the sensing data of the environment in the target area and the areas surrounding the target area; and analyse the retrieved data in every predetermined time interval for detecting changes in the retrieved data. Preferably, the real time visioning analytic module further configured to: translate basic shape of the captured image into ontology based image; extract the color code used in the ontology based image; sort the color code based on its percentage; compare the ontology based image with the captured image; and determine if there is a significant change in the basic shape and the color code percentage in the image based on the comparison result.

Preferably, the deep learning analytic module is further configured to: retrieve location details of the target area and the areas surrounding the target area from a location service provider module; calculate distance between the target area and the areas surrounding the target area based on the analysis of the natural disaster occurrence; and calculate average of the calculated distance for determining distance of the areas affected by the natural disaster surrounding the target area.

Preferably, the system further comprises an output module configured to broadcast a low alert signal along with the calculated distance of the areas affected due to the natural disaster surrounding the target area, if the natural disaster score is less than or equal to one, this condition indicates the natural disaster is about to occur in the target area; and broadcast a high alert signal, if the natural disaster score is greater than one, this condition indicates the natural disaster has occurred in the target area

BRIEF DESCRIPTION OF DRAWINGS

These and other features, aspects, and advantages of the present invention will become better understood, when the following detailed description is read with reference to the accompanying drawings in which like characters represent like parts throughout the drawings, wherein:

Figure. 1 is a block diagram of a system to detect an occurrence of a natural disaster, in accordance to an embodiment of the present invention.

Figure. 2 is a flow chart illustrating a method to detect the occurrence of the natural disaster in a target area, in accordance to an embodiment of the present invention.

Figure. 3 is a flow chart illustrating the steps to monitor a captured image of the target area, in accordance to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

The present invention provides a system and method for high accuracy detection of a natural disaster occurrence using deep learning analytics. Further, the system and method is able to detect the current situation of the natural disaster in the target area and also able to estimate distance of the area surrounding the target area effected due to the natural disaster event.

The term“natural disaster” used in this document can refer to a flood, hurricane, earthquake, tsunami, typhoon, landslide and any other natural disaster known in art.

Referring to figure 1, illustrates a system for the detection of the natural disaster occurrence comprises of: a real time visioning analytic module (110), a sensor analyser module (120), an information collector and analyser module (130), a deep learning analytic module (140) and an output module (150).

The real time visioning analytic module (110) can be a video analytics module that can collect real time data related to the natural disaster in the target area and the areas surrounding the target area by capturing the images of the target area and the areas surrounding the target area in real time using a digital video camera, camera feeds, and uploaded video. Preferably, the real time visioning analytic module (110) is arranged in the target area and the areas surrounding the target area. Further, the real time visioning analytic module (110) is configured to analyze the captured image at every pre-determined time interval. The real time visioning analytic module (110) is configured to translate the basic shape of the captured image into an ontology based image based on the concept of ontology, where the translated image is represented by series of color code with different percentage. The real time visioning analytic module (110), is also configured to extract the color code used in the translated image and sort the color code based on its percentage. Further, the real time visioning analytic module (110), is configured to compare the ontology based image with the captured image. The real time visioning analytic module (110), is further configured to determine if there is any change in the basic shape and the color code percentage used based on the comparison result.

The sensor analyser module (120) is configured to collect real time sensing data related to the natural disaster based on environment of the target area and the areas surrounding the target area for every pre-determined time interval. Further, the sensor analyser module (120) is configured to analyse the collected data to determine any changes in the environment of the target area and the areas surrounding the target area. Preferably, the sensor analyser module (120) includes a water level sensor, a rain fall sensor and any other sensors known in the art arranged in the target area and the areas surrounding the target area.

The information collector and analyser module (130) is configured to collect the data related to the natural disaster occurred in the target area and in the areas surrounding the target area from various data sources which includes web media, a national weather department, national irrigation and drainage department, historical data and any other data sources known in the art. Preferably, the information collector and analyser module (130) collects the data from the various data sources by using combinations of keywords with the nearest location. Further, the information collector and analyser module (130), is configured to calculate a term frequency for the keywords used to retrieve the data based on the results and stores in it. The calculated term frequency determines the significance of that particular keyword within the data. The data is retrieved from the web media includes web, wiki, news, books, blog and any other related known in the art. Further, the data is also retrieved from the historical data on the previous natural disaster events in the target area and in the areas surrounding the target area.

The deep learning analytic module (140) is a type of machine learning that performs tasks like organizing the data to run through predefined equations and deep learning sets up basic parameters about the data. The deep learning analytic module (140), is configured to retrieve the collected data from the real time visioning analytic module (110), the sensor analyser module (120) and from the information collector and analyser module (130). Further, the deep learning analytic module (140) is configured to compare and analyze the retrieved data from the real time visioning analytic module (110) and the sensor analyser module (120) with the retrieved data from the information collector and analyzer module (130), so as to provide high accuracy of the natural disaster occurrence. Preferably, the deep learning analytic module (140) is configured to harmonize and organize the retrieved data into structured information as shown in example table 1. This structured information is used for further analysis. Based on the analyzed result of the natural disaster occurrence, the deep learning analytic module (140) is configured to assign a value for each retrieved data as shown in example table 1, the value assigned is either one or zero. The value one indicates the natural disaster has occurred, and the value zero indicates the natural disaster hasn’t occurred. For example, a first value is assigned to the data retrieved from the real time visioning analytic module (110), a second value is assigned to the data retrieved from the sensor analyser module (120) and likewise for the each data retrieved from the information collector and analyser module (130). If the natural disaster has occurred, then the value is assigned as one. If the natural disaster hasn’t occurred, then the value is assigned as zero. The deep learning analytic module (140), is configured to compute a first score by averaging of the assigned values of the data retrieved from the information collector and analyser module (130), whereby the first score indicates whether the natural disaster has occurred in the target area and in the areas surrounding the target area. Then, the deep learning analytic module (140) is configured to compute a natural disaster score by summing of the first score and the assigned values of the data retrieved from the real time visioning analytic module (110) and from the sensor analyser module (120).

The output module (150), is configured to broadcast an alert based on the natural disaster score calculated by the deep learning analytic module (140). Preferably, the output module (150) is configured not to broadcast the alert if the value of the natural disaster score is equal to pre-determined value which is zero.

Natural disaster score = 0

Preferably, the output module (150) is configured to broadcast a low frequency alert if the value of the natural disaster score is greater than zero and less than equal to one. This condition indicates the natural disaster is about to occur in the target area.

Natural disaster score > 0 & <=1

Preferably, the output module (150) is configured to broadcast a high frequency alert if the value of the natural disaster score is greater than one. This condition indicated the natural disaster has occurred in the target area.

Natural disaster score > 1.

Further, the system (100) further includes a location provider module configured to collect the exact location details of the areas surrounding the target area via global positioning services, mobile tower, wireless network signal or any other known location provider services. The location details from the location provide module are retrieved by the deep learning analytic module (140), to calculate the accurate distance of the areas surrounding the target area affected by the natural disaster based on the analyzed result of the natural disaster occurrence. The accurate distance is determined by average of the calculated distances by the deep learning analytic module (140).

Referring to figure 2, is the flow chart illustrating the processing steps (210 to 270) implemented by the system (100) to detect the natural disaster occurrence in the target area comprises the steps of: (Step 210) collect real time data and sensing data by the real time visioning analytic module (110) and the sensor analyser module (120) related to the natural disaster in the target area and the areas surrounding the target area, where the real time visioning analytic module (110) and the sensor analyser module (120) are arranged in the target area and the areas surrounding it to monitor those areas.

The real time visioning analytic module (110) captures the real time images or video data of the target area. Next, the real time visioning analytic module (110) analyzes the captured data for every pre-determined time interval to determine whether the natural disaster occurred in the target area.

The sensor analyser module (120) collects the sensing data related to the environment of the target area and the areas surrounding the target area for every pre-determined time interval. Next, the sensor analyser module (120) analyzes the collected sensing data to determine any changes in the environment of the target area and in its surrounding areas.

Next, (Step 220) collect the data related to the natural disaster occurred in the target area and in the areas surrounding the target area, by the information collector and analyser module (130), from the plurality of data sources. Where the various data sources can be web media, a national weather department, national irrigation and drainage department, historical data and any other data sources known in the art. Preferably, the steps to collect the data by the information collector and analyser module (130) comprises search by using combinations of keywords and nearest locations to collect the natural disaster related data from the various data sources. Next, calculate the term frequency for the keywords used to retrieve the data based on the results. The term frequency determines the significance of that particular keyword within the data used in the search. And, retrieve the natural disaster data based on the location.

Next, (Step 230) retrieve, by the deep learning analytic module (140), the collected data from the real time visioning analytic module (110), the sensor analyser module (120) and the information collector and analyser module (130). In this step, the data is harmonized and organized in form in structured information as shown in example table 1. This structured information is used for further analysis.

Next, (Step 240) assign a value to each of the retrieved data based on the analyzed result of the natural disaster occurrence. For example, the first value is assigned to the data retrieved from the real time visioning analytic module (110), second value assigned to the data retrieved from the sensor analyser module (120) and likewise for every data retrieved from the information collector and analyser module (130). The value assigned is either one or zero depending on the analysis. If the natural disaster has occurred, then the value is assigned as one. If the natural disaster didn’t occurred, then the value is assigned as zero.

Next, (Step 250) compute the first score, by averaging the assigned values of the data retrieved from the information collector and analyser module (130), by the deep learning analytic module (140), whereby the first score corresponds to occurrence of the natural disaster in the areas surrounding the target area.

Next, (Step 260) compute the natural disaster score, by summing the first score and the assigned values of the data retrieved from the real time visioning analytic module (110) and from the sensor analyser module (120), by the deep learning analytic module (140), whereby if value of the calculated natural disaster score is greater than a pre -determined value, then it indicates the natural disaster has occurred in the target area and if the value of the calculated natural disaster score is less than or equal to the pre-determined value, then it indicates the natural disaster is about to occur in the target area.

Then, (Step 270) broadcast the alert signal by the output module (150) based on the natural disaster score.

The method further includes the steps to calculate accurate distance of the areas affected by the natural disaster surrounding the target area is by collecting exact location details of the areas surrounding the target area by the location provider module. The location provider module collects the location details via global positioning services, mobile tower, wireless network signal or any other known location provider services. Next, retrieve the location details by the deep learning analytic module (140), to calculate the distance between the areas surrounding the target area affected by the natural disaster and the target area. Next, calculate average of the calculated distances by the deep learning analytic module (140), to determine the accurate distance of the areas affected by the natural disaster surrounding the target area.

Referring to figure 3, the flow chart illustrates the processing steps (221 to 225) implemented by the real time visioning analytic module (110) to monitor the captured image for every pre-determined time interval comprises (Step 221) translate the basic shape of the captured image into the ontology based image. Then, (Step 222) extract the color code used in the ontology based image and sort it based on the percentage. Then, (Step 223) compare the ontology based image with the captured image. Next, (Step 224) determine if any changes in the basic shape and color code percentage.

In an exemplary embodiment, the present invention is now explained with reference to the flood detection. However, the present invention is not limited only to the flood detection.

The real time visioning analytic module (110), collects the captured real time images from the target area and the area surrounding the target area. Then, the captured images are analyzed by the real time visioning analytic module (110) to detect the flood occurrence in the target area. The analyzed data is represented as A.

The sensor analyser module (120), collects the sensing data on the environment of the target area and the areas surrounding the target area. Then, the collected sensing data is analyzed to detect the changes in the sensing data due to the flood. The analyzed sensing data is represented as B.

The information collector and analyser module (130), collect the flood related data in the areas surrounding the target area from the plurality of data sources. The data sources include web media, historical data, national weather department and national irrigation and drainage department. For example, the flood related data collected from the web media is represented as C. The flood related data collected from the historical data based on the previous flood events is represented as D. The flood related data collected from the national weather department is represented as E. And, the flood related data collected from the national irrigation and drainage department is represented as F.

The deep learning analytic module (140), retrieves the data A to F from the real time visioning analytic module (110), the sensor analyser module (120) and the information collector and analyser module (130). The deep learning analytic module (140) analyzes the data A to F, where the data A to B is compared and correlated with the data C to F to detect on flood occurrence. The deep learning analytic module (140) harmonizes and organizes the data A to F as flood occurred and flood not occurred. Then, the deep learning analytic module (140) assigns the value for each data A to F based on the analyzed result. Next, the deep learning analytic module (140), computes the first score by the average of the assigned values of the data C to F. Then, the deep learning analytic module (140), computes the flood score by the sum of the first score and the assigned values of the data A and B. The flood score can be expressed as an equation shown below:

Flood score = (A + B) + [(C+D+E+F)/4]

A to F = 1, if flood detected

A to F = 0, if flood is not detected

The deep learning analytic module (140) retrieves the location details from the location service provider module to calculate distance between the target area and the areas surrounding the target area based on the analysis of the flood occurrence. The accurate distance of the areas affected by the flood occurrence from the target area is determined by calculating the average of the calculated distance. The accurate distance can be expressed as an equation shown below.

Accurate distance = (C+D+E+F)/4

Table 1 : Structured information

Flood Score = (0+0) + [(l+l+l+l)/4]

Flood Score = 1

Accurate distance = (l + 4 + 4 + 3)/4

Accurate distance = 3km

The symbol X in the first row and in second row of the table 1 indicates which data discloses that flood has occurred and which data discloses that the flood hasn’t occurred yet.

Example: Alert is issued by the output module (150) as the flood happened 3km away (average km from the target area) within 1 hour the target area can be effected by the flood. The present disclosure includes as contained in the appended claims, as well as that of the foregoing description. Although this invention has been described in its preferred form with a degree of particularity, it is understood that the present disclosure of the preferred form has been made only by way of example and that numerous changes in the details of construction and the combination and arrangements of parts may be resorted to without departing from the scope of the invention.

Claims

1. A computer implemented method for detecting an occurrence of a natural disaster, the method comprising the steps of:

collecting, by a real time visioning analytic module (110) and a sensor analyser module (120), real time data and sensing data related to the natural disaster from a target area and from areas surrounding the target area;

collecting, by an information collector and analyser module (130), data related to the natural disaster occurred in the target area and the areas surrounding the target area from a plurality of data sources, wherein the data sources include data from a web media, data from a national weather department, data from a national irrigation and drainage department and historical data; and

retrieving and processing, by a deep learning analytic module (140), all the collected data, characterised in that processing of the collected data comprises the steps of:

assigning a value to each of the retrieved data from the real time visioning analytic module (110), the sensor analyser module (120) and the information collector and analyser module (130) based on the occurrence of the natural disaster;

computing a first score by averaging the assigned values of the data retrieved from the information collector and analyser module (130), whereby the first score indicates whether the natural disaster has occurred in the areas surrounding the target area; and

computing a natural disaster score by summing the first score and the assigned values of the data retrieved from the real time visioning analytic module (110) and from the sensor analyser module (120), whereby if the natural disaster score is greater than a pre determined value, it indicates the natural disaster has occurred in the target area and if the natural disaster score is less or equal to the pre determined value, it indicates the natural disaster has not occurred in the target area.

2. The method according to claim 1, wherein the steps of collecting real time data related to the natural disaster by the real time visioning analytic module (110) includes:

capturing an image of the target area and the areas surrounding the target area; and

analysing the captured image in every predetermined time interval, for monitoring whether the natural disaster has occurred in the target area. 3. The method according to claim 1, wherein the steps of collecting the sensing data related to the natural disaster by the sensor analyser module (120) includes:

retrieving the sensing data of the environment in the target area and the areas surrounding the target area; and

analysing the retrieved data in every predetermined time interval for detecting changes in the retrieved data.

4. The method according to claim 2, wherein the steps of analysing the captured image by the real time visioning analytic module (110) comprises the steps of: translating basic shape of the captured image into an ontology based image, wherein the ontology based image is represented by a series of color codes with different percentage;

extracting the color code used in the ontology based image; sorting the color code based on its percentage;

comparing the ontology based image with the captured image; and determining if there is a significant change in the basic shape and the color code percentage in the image based on the comparison result.

5. The method according to claim 1, further comprises the steps of:

retrieving, by the deep learning analytic module (140), location details of the target area and the areas surrounding the target area, from a location service provider module; calculating, by the deep learning analytic module (140), distance between the target area and the areas surrounding the target area based on the analysis of the natural disaster occurrence; and

averaging, by the deep learning analytic module (140), the calculated distance, for determining distance of the areas affected by the natural disaster surrounding the target area.

6. The method according to claim 1, further comprises the steps of:

broadcasting a low alert signal, by the output module (150), along with the calculated distance of the areas affected due to the natural disaster surrounding the target area, if the natural disaster score is less than or equal to one, wherein this condition indicates the natural disaster is about to occur in the target area; and

broadcasting a high alert signal, by the output module (150), if the natural disaster score is greater than one, wherein this condition indicates the natural disaster has occurred in the target area.

7. A system (100) for detecting an occurrence of natural disaster comprising:

a real time visioning analytic module (110) configured to collect real time data related to the natural disaster from a target area and from areas surrounding the target area;

a sensor analyser module (120) configured to collect real time sensing data on the natural disaster from the target area and from the areas surrounding the target area;

an information collector and analyser module (130) configured to collect the data related to the natural disaster occurred in the target area and the areas surrounding the target area from a plurality of data sources, wherein the data sources include data from a web media, data from a national weather department, data from a national irrigation and drainage department and historical data; and

a deep learning analytic module (140) configured to retrieve and process all the collected data, characterised in that the deep learning analytic module (140) configured to: assigns a value to each of the retrieved data from the real time visioning analytic module (110), the sensor analyser module (120) and the information collector and analyser module (130) based on the occurrence of the natural disaster;

computes a first score by averaging the assigned values of the data retrieved from the information collector and analyser module (130), whereby the first score indicates whether the natural disaster has occurred in the areas surrounding the target area; and

computes a natural disaster score by summing the first score and the assigned values of the data retrieved from the real time visioning analytic module (110) and from the sensor analyser module (120), whereby if the natural disaster score is greater than a pre determined value it indicates the natural disaster has occurred in the target area and if the natural disaster score is less than or equal to the pre-determined value, it indicates the natural disaster has not occurred in the target area.

8. The system (100) according to claim 7, wherein the real time visioning analytic module (110) is further configured to capture an image of the target area and the areas surrounding the target area and analyse the captured image in every predetermined time interval, to monitor whether the natural disaster has occurred in the target area. 9. The system (100) according to claim 7, wherein the sensor analyser module

(120) is configured to retrieve the sensing data of the environment in the target area and the areas surrounding the target area and analyse the retrieved data in every predetermined time interval for detecting changes in the retrieved data.

10. The system (100) according to claim 8, wherein the real time visioning analytic module (110) is further configured to translate basic shape of the captured image into an ontology based image; extract the color code used in the ontology based image; sort the color code based on its percentage; compare the ontology based image with the captured image; and determine if there is a significant change in the basic shape and the color code percentage in the image based on the comparison result. 11. The system (100) according to claim 7, wherein the deep learning analytic module (140) is further configured to retrieve location details of the target area and the areas surrounding the target area from a location service provider module; calculate distance between the target area and the areas surrounding the target area based on the analysis of the natural disaster occurrence; and calculate average of the calculated distance for determining distance of the areas affected by the natural disaster surrounding the target area.

12. The system (100) according to claim 7, further comprises an output module (150) configured to: broadcast a low alert signal along with the calculated distance of the areas affected due to the natural disaster surrounding the target area, if the natural disaster score is less than or equal to one, this condition indicates the natural disaster is about to occur in the target area; and broadcast a high alert signal, if the natural disaster score is greater than one, this condition indicates the natural disaster has occurred in the target area.