Hackathon winner's drone uses visual recognition to help hurricane victims

Previous IBM Call for Code hackathon winner's DroneAid is being open sourced to work more efficiently toward widespread implementation.

The key to open source sustainability is good old-fashioned self-interest Don't look to donations to solve open source maintainers' money woes. They just need to be connected to the companies that need them to make a buck.

Call for Code 2018 Puerto Hackathon winner Pedro Cruz is getting his shot at making his winning DroneAid concept a reality. In his IBM Code and Response blog post on Thursday, Cruz outlined his inspiration behind DroneAid and announced the decision to open source his project.

The annual Call for Code event is a collaborative project between IBM and humanitarian organization David Clark Cause. The project challenges developers across the globe to create applications based on cloud, data, and artificial intelligence (AI) to help mitigate the effects of natural disasters. The 2019 Call for Code finalists addressed issues including the prioritization of emergency calls, physical and mental health care, and toxic smoke inhalation during catastrophic situations. 

SEE: How to build a successful developer career (free PDF) (TechRepublic)

pedro-cruz.png

Pedro Cruz with using the DroneAid technology

Image

Cruz's winning project uses visual recognition technology to detect and count SOS icons on the ground via overhead drone streams. After detection, the drones automatically contact first responders, communicating that area's specific needs for victims. 

The inspiration behind DroneAid came from Cruz's personal experiences during Hurricane Maria's wrath on Puerto Rico in 2017. During the disaster, rural Puerto Rican communities painted "water" and "food" on the ground as an SOS, hoping helicopters or planes would see their messages and get them help, Cruz wrote in the post. 

DroneAid is able to visually detect a standard set of icons—drawn with spray paint, chalk, etc.—and understand the specific needs by symbol. Specifically, the symbols can indicate immediate help needed, no help needed, water needed, food needed, shelter needed, first aid kit needed, area with children in need, or area with enderly in need. 

How DroneAid works

From the technical side, DroneAid uses a visual recognition AI model that is trained on the standardized icons. IBM's cloud annotations tool help strain AI using the IBM Cloud object storage. The drone surveys areas via video, and the video frames are analyzed for symbols, and if a symbol is detected, it is plotted on a map with the location and number of people in need, according to the post. 

The system is able to run on any drone using the steps in the source code repository. 

Moving to open source

In partnership with IBM's Code and Response, DroneAid is now being made into an open source project. Code and Response is a $25 million, four-year initiative that will build, test, and launch open tech solutions to help communities that have fallen victim to natural disasters. Winners of Call for Code are able to launch their programs with Code and Response. 

Cruz's decision to open source DroneAid was inspired by the mission to make this vision a reality. Making the project open source means the original source code is freely available and able to be modified by anyone.

"The standardized icon approach can be used around the world in many natural disaster scenarios (i.e., hurricanes, tsunamis, earthquakes, and wildfires) and having developers contribute by training the software on an ongoing basis can help increase our efficiency and expand how the symbols can be used together," Cruz noted in the post. 

Click here to contribute to the open source project. 

Also see 

agriculture.jpg

DroneAid application in action

Image: DroneAid