A PhD student from Russia has used a Raspberry Pi (opens in new tab) and some other innocent pieces of tech to design an innovative solution that neutralizes mosquitos using lasers.
Rakhmatulin Ildar from South Ural State University says he lost his cool when he learnt that mosquitoes kill over 700,000 people every year, and decided to do something about it.
Ildar's solution was to mount a 1-watt laser on a Raspberry Pi 3 and train it with machine vision and neural networks until it became a ruthless mosquito hunting killer automaton.
- These are the best Raspberry Pi distros (opens in new tab)
- Here are the best Linux laptops (opens in new tab) for running Linux
- These are the best laptops for programming (opens in new tab)
On a rampage
Tom’s Hardware intercepted Rakhmatulin's frighteningly detailed build instructions, noting that the project uses the Raspberry Pi Camera module to detect mosquitoes in its vicinity.
When one is spotted, the Pi asks a connected galvanometer to adjust the mirrors and orient the laser. Once the lasers are in position, the Pi powers it up and delivers a potent burst to zap the flying creten out of existence.
Rakhmatulin has shared details on how he created a convolutional neural network (CNN) to detect mosquitoes. To perfect the AI network he trained it using a GeForce GTX 1060 (opens in new tab) 6GB GPU with the Ryzen 5 3600 (opens in new tab) CPU.
He’s also published details about his use of the OpenCV library to track an identified mosquito and predict its flight pattern.
Buoyed by the success of his research, Rakhmatulin now wants to build a more powerful version that can track and eliminate more than two mosquitoes in a second.
- Subscribe to Linux Format magazine (opens in new tab) for more Linux and open source goodness
Via: Tom’s Hardware (opens in new tab)