Integrated Colormap and ORB detector method for feature extraction approach in augmented reality

Augmented Reality (AR) is a technology that addition of virtual objects into the real-world environment. AR technology uses images recognition approaches to recognize objects. The objects can be easily recognized if rich in details, have good contrast, and have no repetitive patterns. A feature-based technique called Natural Feature Tracking (NFT) system can be used to recognize physical objects in markerless AR. The features such as blob, edge, and corner in the object are extracted by the feature detector and descriptor before recognizing process. The extraction feature is the most important thing in the recognition process because it can determine accurate results. ORB detector is a feature extractor were suitable for real-time tracking in AR because it has speed, efficiency, and a high quantity of features detected and extracted. However, before detecting and describing the features, ORB detector uses the Grayscale Image Generation (GIG) process to change color images into grayscale images. We found some features extracted using the GIG process not extracted perfectly. ORB detector is influenced by the intensity of the grayscale pixel to find the candidate corner. The proposed integration of the Colormap technique and ORB detector method can enhance feature extraction for improving features detection in AR.