Detection of Agricultural Microbial Colonies Based on Image Morphology and Otsu Method

  • Xiaoyu Xu
Keywords: Agriculture Microorganism, Intelligent Irrigation, Image Morphology, Otsu Method

Abstract

Microorganism has played an important role in the health, growth and development of agricultural animals and
plants, the improvement and restoration of water, soil and atmospheric environment. So to learn about the
impact of microorganism on agricultural production activities, the approaches it exerts the effect, and the
regulation of plant-related microbial community structure is of great significance to enhance the applications of
microorganism in agricultural production and then increase crop yield, save relevant costs, and promote the
development of green agriculture and eco-agriculture.According to the information of microbial colonies, it can
also realize timely and appropriate intelligent irrigation for farmland, and play an important role in the growth of
crops and the construction of intelligent agriculture.This paper proposes the method for detecting agricultural
microbial colonies based on image morphology and Otsu algorithm. This method, mainly based on the
determination result of a single frame of image in the video of colonies of agriculture microorganisms and the
related information of continuous video frames, records different key states at different moments and in
different spaces, and through the switch of different key states, maximizes all environment information obtained
in order for accurate detection, tracking and counting statistics of microbial colony target.It demarcates all
possible targets detected in every frame of image, proceeds multi-frame tracking on every target, and combines
with the results of follow-up frames for comparison. Every target object must go through confirmation of
different states, including detection state, storage state, foreground extraction state, statistic state, deletion state
and display state, so as toremoveinaccurate detection targets, preserve, track and count the accurate ones. The
simulation experiment proves that this method is effective.

Published
2020-02-01