README.md

Automatic Parking Detection

This repository contains:

• C++ code
• Python prototype
• 1 video for testing

Running

DetectParking.exe <Video_Filename or Camera_Number> <ParkingData_Filename>

Where:

• <Camera_Number> can be 0, 1, ... for webcam or usb camera.
• <ParkingData_Filename> should be simple txt file with lines of: id x1 y1 x2 y2 x3 y3 x4 y4. The x,y are the quadrilateral 4 points which mark the parking spot. See \datasets\parkinglot_1.txt for example.

Required Dlls from OpenCV: opencv_core300.dll, opencv_ffmpeg300.dll, opencv_highgui300.dll, opencv_imgcodecs300.dll, opencv_imgproc300.dll, opencv_videoio300.dll.

Parameters

You can play with the threshold in Parking.h.

#define PARK_LAPLACIAN_TH 2

You can turn on/off the parking detection or saving output video in main.cpp by changing:

#define DETECT_PARKING true
#define SAVE_VIDEO false

Building

Linux: I compiled it on Raspbian (Debian) with GCC 4.9 using the following command:

g++ -Wall -g -std=c++11 `pkg-config --cflags --libs opencv` ./*.cpp -o ./detect-parking

Windows: Add the files to a Visual Studio solution. Add the OpenCV header and libraries files and build.

Screenshots

Video 1