Skip to content

joshuamataaraya/SoccerAnalysis

Repository files navigation

SoccerAnalysis

This is an analysis of soccer videos

Table of Contents

  1. Requirements
  2. OpenCV Installation
  3. DLLs Installation
  4. Web Server Installation

Requeriments

  • JDK 8.1 or above
  • Eclipse Neon Jee
  • OpenCV-3.1
  • DLLs

OpenCv Installation

  1. Download the library opencv-3.1 from the followin link: https://sourceforge.net/projects/opencvlibrary/files/opencv-win/3.1.0/opencv-3.1.0.exe/download
  2. Extract on the root of your HardDrive when you're installing it.
  3. Follow the instructions on the following link: http://docs.opencv.org/2.4/doc/tutorials/introduction/java_eclipse/java_eclipse.html#java-eclipse
  4. While you are installing opencv in your Eclipse IDE, it´s important to name it "opencv-3.1".
  5. Copy the file opencv\build\bin\opencv_ffmpeg310_64.dll into opencv\build\java\x64

DLLs Installation

DLLs requiered for running the tests or/and the server full functionality. Copy the DLLs from the folder Dependencies in the project and paste them on: C:\opencv\build\java\x64 The project wont work unless you do this last step.

Web Server Installation

This is used to deploy the UI interface, when it's deployed it can be accesed through "localhost:8080". We decided to use Jetty's and Maven integrated server, so here is the basic configuration

  1. Right Click on the project -> Run as -> Run Configurations

    Alt text

  2. Select Maven Build - New

    Alt text

  3. Fill in the next three boxes highlited by the red boxes.

    Alt text

  4. In the JRE Tab make sure its selected the workspace defualt JRE and it sshould be a JDK one as seen on the image, not a JRE

    Alt text

    Note if your default is on JRE, close the run configurations, go to Window -> Preferences -> Java -> Installed JRE's and add the jdk folder and/or check the jdk box and apply.

  5. In the same tab, add the next VM arguments: -Djava.library.path=C:\opencv\build\java\x64 hit apply and run.

    Alt text

***Note: since the WebSocket simulates the upload of your internet speed, its highly recommended to use files with low sizes. We recommend two files located in our drive: https://drive.google.com/drive/folders/0B1b_MSFDPh5ZRXcxMF9zUmhvLXM?usp=sharing one for running a JUnit test, which uses a big size file with more resolution (more on dependencies) and the second is a small size file .mp4 which generates fairly good results and can be used for the upload on the WebBrowser.

Also, the web page was designed for Google Chrome, take that into account. You need to select the file and press the "Subir video" button for the analysis to start. Also for the Groundtruth, select the video file and the groundtruth file and then upload the video with the "subir video" button under the Groundtruth Category. There's a small bug and problemas with optimization when doing multiple uploads on the same client. It's highly recommended to restart the server between uploads on these earlier versions.