AI-Powered Laser System Targets Mosquitoes

Steven Cheng built an autonomous system that detects, tracks, and eliminates mosquitoes using a deep-learning model and a laser mounted on a high-precision industrial rotary stage. The AI identifies mosquitoes in real time, relays target coordinates to a gimbal, and fires the laser only when a mosquito is confirmed.

Interesting Engineering

The project merges computer-vision advances, industrial robotics, and laser technology that have become increasingly accessible to hobbyists. Cheng’s work shows how off-the-shelf imaging equipment and consumer-grade GPUs can be combined to create specialized automated solutions outside traditional research labs.

Interesting Engineering

Cheng is a computer-vision and robotics enthusiast who documented the prototype online. He designed the custom mosquito image dataset, trained the detection model, and integrated safety subsystems, describing the device as the “ultimate mosquito killer.”

Interesting Engineering

• Month 1-2: Photographing mosquitoes with a DSLR and high-magnification zoom lens; building a dataset of thousands of images.

Interesting Engineering

• Month 3: Training the deep-learning model on a graphics card; calibrating the laser.

Interesting Engineering

• Month 4: Integrating the laser with the rotary stage and gimbal; conducting simulation tests and live-room trials.

Interesting Engineering

• Final night: System eliminated all resident mosquitoes in a single test.

Interesting Engineering

The dataset comprises thousands of annotated mosquito photos captured under controlled lighting. Cheng used a graphics card that “really put my graphics card through its paces” to train the model. The laser assembly sits on an industrial-grade tracking platform with a gimbal for rapid orientation, while a secondary wide-angle camera monitors the surrounding area for safety.

Interesting Engineering

A dual-camera setup continuously checks for humans or flammable objects within the laser’s engagement zone. If any such object is detected, the system aborts firing, a safeguard introduced after simulation testing. Cheng reports that the prototype performed as intended during these safety checks.

Interesting Engineering

Cheng states that the detection performance of the final model was “quite good,” and that the safety features successfully prevented accidental discharge. He confirms that after a night of operation, all mosquitoes in his residence were “successfully eliminated.” The system remains a personal prototype, not a commercial product.

Interesting Engineering

• “countless mosquito bites all over my body,” — Steven Cheng, project developer

Interesting Engineering

• “really put my graphics card through its paces,” — Steven Cheng

Interesting Engineering

• “instantly turn mosquitoes into roasted ones.” — Steven Cheng

Interesting Engineering

• “quite good.” — Steven Cheng (on detection performance)

Interesting Engineering

• “successfully eliminated” — Steven Cheng (on mosquito removal)

Interesting Engineering

The prototype illustrates that sophisticated AI-driven pest control can be engineered by individuals, potentially expanding the scope of DIY robotics. It also raises considerations about safety, regulatory oversight, and the ethical deployment of autonomous laser systems in domestic settings.

Interesting Engineering

Cheng emphasizes that the project showcases the growing accessibility of AI, computer vision, and robotics, suggesting that similar personal-automation initiatives may emerge as hardware and software tools continue to democratize advanced engineering.

Interesting Engineering