A young Chinese engineer named Steven Cheng has built an amazing system to kill mosquitoes. Their project makes use of a laser emitter connected to a camera and a computer. By combining computer vision algorithms with deep learning, the system is able to detect and fry mosquitoes with surprising accuracy. This is indeed the “artisanal” version of a commercial product from a Chinese startup, but there is a whole story behind this unique laser weapon.
From space missiles to lasers for mosquitoes. The idea of using lasers against mosquitoes was probably born by Lowell Wood, an astrophysicist who was also the architect of Ronald Reagan’s anti-missile defense system. In a meeting with engineer and inventor Nathan Myhrvold—former CTO of Microsoft—the proposal arose to create a similar device to kill mosquitoes and thus fight malaria.
failed attempt. Myhrvold got to work excited because mosquitoes weigh so little that a low-intensity laser was enough and could even be applied near (or on) humans and pets. Myhrvold’s company, Intellectual Ventures, created a prototype that showed in a TED talk in 2010but despite the attractiveness of the proposal, the idea and the product did not prosper.
The commercial leap. A Chinese startup recently picked up the gauntlet and began offering a product that precisely applied that same idea with current technology. The startup Photon Matrix Lab already markets what they call “Mosquito Air Defense”, a portable system that is available on Indiegogo and that promises to eradicate mosquitoes through a detection system and a laser that fries these insects. On the official website the price is about 600 eurosand deliveries promised to begin this June for the nearly 4,000 people who have invested in the crowdfunding campaign.
But there are those who cook it and put it together. Steve Cheng does not say what inspired him for the project, but he does sample in several tweets the videos in which you can see how your system has evolved. He also explains how for months he was collecting thousands of detailed images to feed his AI model. That, by the way, caused (always according to him) that he received various stings all over his body while he used high-resolution cameras to capture the insects in full feast.
Anti-mosquito deep learning. The heart of their system is a deep learning algorithm that processes images in real time to identify static or flying mosquitoes. The system uses GPUs to analyze the environment and detect the mosquito, separating it from any visual background noise. To detect mosquitoes, it uses a digital SLR camera with a telephoto lens that allows zooming to identify mosquitoes from a greater distance.
precision attack. The objective is to confirm that the target is indeed a mosquito before frying it with the laser. This engineer used a laser mounted on a gimbal capable of moving in milliseconds to act instantly. In the very striking video, you can see how the laser actually seems to kill the mosquitoes in a really satisfactory way.
Many unknowns. The tweets and the existence of that business model are promising, but we have seen in the past how crowdfunding campaigns did not end well by delivering products that did not live up to the promise. The real effectiveness of these systems is unknown, but the idea is of course especially striking, especially when we are already beginning to suffer high temperatures: mosquitoes will not take long to plague us day and night, if they do not do so already, and there is still no 100% effective remedy against what Bill Gates described as the deadliest animals on earth.
Image | Wolfgang Hasselmann | Steven Cheng
In Xataka | Anti-mosquito repellents have been effective for 40 years. Now mosquitoes are learning to appreciate them
GIPHY App Key not set. Please check settings