bnair2001 / driver-monitoring-system

An app which uses open CV to monitor if the driver is alert of asleep, and a state farm dataset trained model which scans for 9 different types of distraction.

Driver-monitering-system

Driver behaviour detecting app that warns if the behaviour is detrimental.

How it works:

• The app starts of with the camera running and it looks for the eyes in the video.
• It then looks for any distracting behaviour shown by the driver.
• If the behaviour is detected to be bad it warns the driver.
• The alert runs till the driver touches the screen and if it takes too long an sos is sent to his emergency contact number

Note:

The repo is divides into 3 files:

• The alert folder contains the app that triggers the alarm by flashing the screen red and also makes the flashlight of the camera to blink.
• The tfmobile folder consists of the android app that detects around 9 classes of distractions such as:
  • safe driving
  • texting - right
  • talking on the phone - right
  • texting - left
  • talking on the phone - left
  • operating the radio
  • drinking
  • reaching behind
  • hair and makeup
  • talking to passenger
  • Here is the accuracy and loss of the cnn model used to build this application:
• The vigilate folder contains the app that detects if the driver is sleeping or not.

• example:

Installation:

• First install openCv manager for android by using this link.
• Next open the folder of the app you wish to use via android studio and you are good to go.

Important note:

Th three apps are now seperate apps, if you would like to contribute plases help by combining the three into one.

Credits:

Bharath Nair,
Steve Paul

References:

tensorflow for poets
OpenCV for android
The ML model for the tfmobile app

driver-monitoring-system's Issues

Merge the three folders

Merge the three folders alert, vigilate and tfmobile which are 3 separate android apps built on Gradle. Your task is to put them into different activities of the same app.