shyanasri / ParkIT-hardware

Machine Vision Algorithm on a Raspberry Pi - OpenALPR

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

IoT Component

Raspberry Pi, USB Webcam, PIR Sensor, OpenALPR, OpenCV, Tesseract, Leptonica

Bash scripts

  • for interpreting USB webcam
  • for calling OpenAPLR library and returning license plate possibilites + confidence (percentage value)

Python Script

  • Upon movement detected by PIR Motion sensor, script takes a picture by calling Bash script
  • Picture is passed OpenALPR Bash script, with entire terminal result kept
  • first license plate (most confident guess) is truncated -- try/catch exception block used to analyze case where no license plate was detected
  • A HTTP POST request is sent to ParkIT server with following features: license (plate number), time (date of photo), and action (for our demo purposes since only one camera was in use, this was hard set to leaving -- usually will be based on which camera recorded plate)

OpenALPR & Raspberry Pi

  • Raspberry Pi does not have a X86, AMD64 or X86_64 processor so OpenALPR had to be built from it's source code to fit the Raspberry Pi architechture, compiling all indiviual dependencies

Future Steps

  • Optimize the application: with the usage of two bash scripts being called, there is some lag in analyzing pictures. This can be done through: -using Cloud Api for OpenALPR - fixing Python bindings for OpenALPR library

About

Machine Vision Algorithm on a Raspberry Pi - OpenALPR


Languages

Language:Python 89.1%Language:Shell 10.9%