Attention: This app was created in 2016. I was a beginner to Android development and Computer Vision back then. So don't expect a perfect code please. In 2021 I updated the project to build with the latest Android Studio (2020.3.1), updated most dependencies and converted it to Kotlin, while the business logic remained unchanged.
- Takes the real time camera picture and uses it for finding a specified template image using a template matching approach with OpenCV for Android (Java library). Lines are drawn on top of the camera image where the searched object is expected. This can for example be used to find a logo.
- Be aware that template matching doesn't allow any rotations or zooming of the camera. Hold the camera without dip and at the 1:1 correct zoom. Otherwise the template image won't be found within the camera image.
- This app is using a gray scale live image for performance reason. This could easily be changed to search in color live images too.
- More computer vision projects at https://michaeltroger.com/computervision/
- IDE: Android Studio (tested with 2020.3.1)
- Android SDK
- Dependencies: OpenCV 3 library (included) License
- Template image location: res/drawable - Image is referenced in MainActivity
Used default template image:
Test with filming the logo at the top left at http://www.pxl.be/
- The template matching is based on the official OpenCV tutorial http://docs.opencv.org/2.4/doc/tutorials/imgproc/histograms/template_matching/template_matching.html Their version is using OpenCV 2 with static images files.