Resources

Contestants will have to provide image processing algorithms that will correctly identify situations where a moving construction site is placed on the road. The solutions will be evaluated using the following:

  • Hardware:
    Processor: Intel Core i5-3320M CPU @ 2.60 GHz
    RAM: 10 GB
    Operating System: Windows 7×64
  • Software:
    Matlab 8.3.0.532 (R2014a)
    Toolboxes (see MatlabVersion)
    C/C++ Algorithms must compiled for specified hardware

Interfaces mandatory, according to the specifications:

The Interfaces can be used as templates for the preferred language.
The C/C++ Interfaces must be callable from command line by the following command:

>> ProgrammName “PathToSequences” “SequenceName”
e.g. >> runConstructionSiteDetection “C:\Sequences” “Sequence1″

The main function of all interfaces has to create the following folder/file structure in the current folder:

./SequenceName
./SequenceName/resImg
./SequenceName/result.txt

The folder resImg is the output folder for the result images with detection overlay. Each img should have the same filename as the input frame (the image format is left open).
The result.txt should contain the detection result of each frame in the format:

inputImageName detectionResult

e.g. img0001.png 0
img0002.png 1
img0003.png 1
img0004.png 2

After Registration, you will be provided the video files, 10 Sequences without Labels for algorithm development and calibration for each Sequence.

File format is 16 Frames per second, PGM-File with 12 bit grayscale values in ASCII Coding (P2-pgm).
http://en.wikipedia.org/wiki/Netpbm_format

Regarding the Image Format: Size is 1176 x 640, Raw Bayer pattern image with „grbg“ pixel alignment .
http://en.wikipedia.org/wiki/Bayer_filter

For full access please Register or Login.