Due to the drastic spreading of the COVID-19 virus , it is important to control and prevent further spreading of the virus to healthy individuals through hand contact at public places. Sanitization of hands is being done using Alcohol based sanitizers.Generally ,any person using the sanitizer should press the pump manually. However,this can cause cross infection between individuals using the same dispenser. The existing contact-less dispensers work using infrared ,proximity or motion sensors which get activated even when any object other than human hand is brought close to the dispenser. This leads to possible tampering of the unit or misuse by some irresponsible individuals.
What it does
In order to overcome the above shortfalls , a sanitizer dispenser which dispenses liquid only when “human hand” is brought near is needed. Our solution caters to these needs. The camera module connected to Raspberry Pi Zero captures video input.When the user brings their hand near the “service area” of the unit ,the object detection algorithm implemented using tensorflow object detection API , will detect “human hand”. If it is detected , using the Pi , a mini Push pull solenoid connected with the pump of the hand sanitizer is activated , which pushes the pump and dispenses the liquid . The push-pull solenoid is deactivated either after a threshold amount of time,say 2 sec,or after the “human hand” is out of the vicinity of the camera. Once the sanitizer liquid is depleted in the dispenser , using an ultrasonic sensor, it is found that the liquid level is low. This is intimated to an authority assigned to refill the sanitizer using SMS sent via the SIM800C GSM module.
How we built it
We built it using a Raspberry Pi Zero which will contain the hand detection deep learning model. A Pi camera is connected to Pi Zero which takes the video input. For activating the hand sanitizer dispenser , mini Push Pull solenoid connected to Pi zero is used. A GSM Module ,also connected to Pi is used to contact anyone regarding refilling the sanitizer. A HC-SR04 ultrasonic sensor connected to the Pi zero is used to get the level of sanitizer in the dispenser.
Challenges we ran into
1. Compatibility issues with Pi 2. Latency of the whole setup from input to output 3. Handling the empty dispenser situation
Accomplishments that we're proud of
- Refilling of the dispenser by intimating it to the authority using SMS sent by the GSM Module
- Using ultrasonic sensor to detect level of sanitizer in dispenser.
What we learned
- Using Raspberry Pi
- Using object detection deep learning model
What's next for Automatic Hand Sanitizer using Computer Vision
- Deploying it as a production level product
gsm-800simc, pi-camera, push-pull-solenoid, python, raspberry-pi, raspbian-os, tensorflow-object-detection-api, ultrasonic-sensor