SURF-Based Multi-scale Resolution Histogram for Insect Recognition

Automatic insect recognition are time-saving and laborsaving and global feature extraction algorithms have been used for recognition. However, local features based automatic insect recognition is not studied. Therefore, in our research, SURF algorithm is used to extract local features of insect images and then the features are taken as input of multi-scale histogram algorithm. The experimental results show that the accurate recognition rate of SURF based multi-scale histogram method is 89%.