A Novel Object Detection Method Based on Fuzzy Sets Theory and SURF

In this paper a novel object detection with the combination of fuzzy concepts and SURF is presented. This method contains several steps which are fuzzy C-means (FCM) clustering, Gaussian filter, edge detection and speed-up robust feature (SURF). The image contains the target object which can be taken from a camera more than 2 meters away. SURF is used to detect the feature points and extract the feature descriptor that shows an excellent performance on robust and accurate object detection; however, it is prone to fail when the target object is rather far away from the camera such as the distance more than 2 meters. The proposed method addresses to solve the problem that is achieved by some pre-processing for the image as follows. Using FCM to shows up the colourful parts and using fuzzy inference system (FIS) to detect the edges in the image. The above two processes can improve the accuracy of feature detection and reduce the calculation of the descriptor in SURF. The experiment shows the proposed method performs the target extraction efficiently and accurately with the distance 2 meters between the target object and the camera.