Our website use cookies to improve and personalize your experience and to display advertisements(if any). Our website may also include cookies from third parties like Google Adsense, Google Analytics, Youtube. By using the website, you consent to the use of cookies. We have updated our Privacy Policy. Please click on the button to check our Privacy Policy.

How does ReportFires app work?

ReportFires is a Free app and is available on the App Store,

https://apple.co/3lFc1cs

ReportFires helps user

  1. to detect nearby wildfire using AI smoke detection
  2. to get notification of wildfire in his/her area
  3. to report wildfire quickly by simple map clicks

We shared with you the What, the Why, and the How-to-use in the previous blogposts. Now let’s talk about How does it work.

With the advancement in consumer technologies, nowadays smartphones are powerful to do parallel computations. iPhone 6S and later models support iOS 13 and Core ML, the Machine Learning framework developed by Apple. Core ML models run strictly on the user’s device and remove any need for a network connection, keeping app responsive and users’ data private. These Core ML features allow us to do three things:

  • real-time predictions of live images on the device
  • run application without a network connection. It is important since network availability and high-speed connectivity are not must-haves in WUI (Wildland-Urban-Interface)
  • the image data never leaves the device for smoke detection purposes. User privacy is our top priority.
Apple Core ML
Figure 1: Apple Core ML integrates Machine Learning into apps (source: apple.com)

Referencing the InsightFD architecture, ReportFires uses the iPhone to play the role of InsightFD robot and edge computer. Users can use ReportFires app to locate wildfire manually or to detect wildfire smoke automatically. The wildfire information will be shared with the public.

Figure 2: the system architecture of InsightFD Early Wildfire Detection System
Figure 3: the system architecture of ReportFires app
  • Data capturing: iPhone rear camera takes images of the nearby forest. ReportFires app provides visual aims to help the user to position the iPhone to optimize its viewing angle and distance. The higher the vertical distance between the iPhone and treetops, the longer the wildfire detection range. An iPhone at 1500-ft (500 meter) vertical distance can detect wildfire smoke 1.5 miles (2.4 km) away from it.
  • Detection: ReportFires app is a stateful wildfire smoke detection software program. It includes a customized iOS version of the RoboticsCats AI Smoke Detection. The AI algorithm analyses images at regular intervals. ReportFires app does detection on edge to minimize data transmission between the iPhone and the centralized backend management system. ReportFires will track the iPhone GPS location, altitude, camera heading direction, and tilt angle. The AI detection algorithm is stateful to reduce false positives. When the ReportFires app detects wildfire smoke, it will estimate the 2D real-world locations of the wildfires. ReportFires app will send the detected wildfire information and images to ReportFires backend system running in the cloud. 
  • Alert and Reporting: ReportFires will generate a sound alert when it detects smoke. The ReportFires backend system will send wildfire alert notifications to those users who are within 6-mile (10 km) distance from the wildfires reported or detected by the app. The corresponding firefighting authority will receive the wildfire alerts too and can review, verify, and dismiss them, using the ReportFires mobile app or web app. The authority formulates the firefighting plan. Both ReportFires users and firefighting authority can take timely actions based on the information provided by the ReportFires app.  

The customized iOS version of the RoboticsCats AI Smoke Detection is running on Core ML. It detects wildfire smoke using Machine Learning-based (ML) image detection. 

Figure 4: RoboticsCats AI Smoke Detection is using Machine Learning algorithm
Figure 4: RoboticsCats AI Smoke Detection is using Machine Learning algorithm

InsightFD is a purpose-built integrated hardware and software system to early detect wildfires at long distances in a tough forest environment. ReportFires app resembles InsightFD in automatic detection of wildfire smokes within a few miles from your home.  

Figure 5: training data for ML image detection
Figure 6: training data for ML image detection
Figure 7: RoboticsCats AI Smoke Detection can detect wildfires in daytime and nighttime
Figure 7: RoboticsCats AI Smoke Detection can detect wildfires in daytime and nighttime

The core technologies we use to build the ReportFires app are Crowd-sourcing, Artificial intelligence, and the internet of Things.

Crowd-sourcing, Artificial intelligence, internet of Things.

Crowd-sourcing, Artificial intelligence, internet of Things.

Are you getting it?

Thank for our backers’ supports! ReportFires app Kickstarter campaign was successfully finished on 2020-10-09.

Unwire.pro, one of the most popular tech news media in Hong Kong, interviewed us last week. We had some in-depth discussions about ReportFires app. Here is their report:

https://unwire.pro/2020/10/09/robotics-cats/executive-interview/

 

Unwire.pro interview
Figure 8: Unwire.pro interview of roboticscats.com (source: unwire.pro)

ReportFires is a Free app and is available on the App Store. Hope you find it helpful!

Related Posts