DocScan

App for document scanning.

Authors

Fabian Hollaus, Florian Kleber, Markus Diem

Build Instructions for Android Studio

Preparation:

• You need to install Android NDK (with CMake) if you do not have it
• Download and unzip OpenCV for Android (https://opencv.org/releases.html)
• Clone the project and open it in Android Studio

Import the OpenCV module

• In Android Studio: 'File' -> 'New' -> 'Import Module'
• Set 'Source directory' to: {your_local_path}/OpenCV-android-sdk/sdk/java
• Set 'Module name' to: openCVLibrary (Note: Do not include any version number here. If Android studio is saying that there is such a module existing, delete the auto generated openCVLibrary folder in the app root folder.)
• Open 'openCVLibrary/build.gradle'
• Correct the version numbers of compileSdkVersion, minSdkVersion, targetSdkVersion so that it matches the version numbers in your app build.gradle (You can also delete the buildToolsVersion since this is not needed by gradle anymore)

Setup CMake file

• Copy the file app/src/CMakeListsSkel.txt and paste it to app/src/, name the pasted file CMakeLists.txt (do not add it to the git repository!)
• Open app/src/CMakeLists.txt
• Change the line include_directories(enter_your_opencv_path/sdk/native/jni/include) so that it contains your opencv path

Create a symlink

• Create a symlink named jniLibs in app/src/main that points to {your_local_path}/OpenCV-android-sdk/native/libs
• Fill in your required architecture in app/build.gradle under abiFilters

Sync and build the project

Visual Studio Project (C++ Library)

• Optional for testing the C++ module
• C++ lib for page segmentation and focus measure
• use CMake to create a Visual Studio Project
  • source code path: DocScan/app/src/main
  • binaries path e.g.: DocScan/build2015-x64
  • Specify OpenCV_DIR

App Dependencies

Google Play services need to be installed to use Firebase Jobdispatcher (https://github.com/firebase/firebase-jobdispatcher-android)