Review of Recent Machine-Vision Technologies in Agriculture

This paper reviews the machine-vision techniques available for image acquisition and their processing-analysis in agricultural automation up to now according to the essential base and the core work, focusing on 2 types of image, i.e., visible image and infrared image. Results from previous studies have shown two important topics on agriculture application, i.e., visual navigation and behaviors surveillance. Each topic introduces important and major applications in respective subfields. Furthermore, each one expresses technical discussions, accordingly. Proper combination of different kinds of machine-vision element for different objects is a normal research way. The accuracy, the robustness and the real-time reveal 3 major obstacles in applications. Further to this, many blanks are not yet developed, and the level of some results is preliminary. Thus, additional research is needed to fully realize this potential.