Insects, due to their diversity, are the largest group of creature on earth, many of which are yet to be discovered right now. As the photographic hardware develops and gets pervasive rapidly, insect images or pictures have become an important approach for insect researchers to conduct scientific study. However, considering the huge volume of insect image data, researchers usually are not interested in all the images and instead, hope to categorize these images preliminarily and then hand them on for manual processing, which will significantly improve work efficiency and reduce manual workload. In this paper, color histogram and textural features are adopted to collect image features which then use clustering analysis method for automatic classification of insect species. The experimental result is quite good. And on the basis of this, more image features can be extracted to gain more accurate results, or program efficiency can be optimized through parallelization, during which a species feature database is gradually established to have automatic comparison and identification of insect features.
Discussion(0)
No comments yet. Be the first to comment.