Analysis and design of video-based activity identification and classification system in agricultural industry
Kai Ma, Dong Ren, Zhiyong Huang, Chuanheng Sun · 2016
The agricultural production video record is an important primitive data in the establishment of agricultural product quality traceability system and digital monitoring system. This paper presents the feature extraction and automatic recognition system of the typical production activities in the agricultural production video record based on the machine learning theory. The system consists of the feature of production activity extraction module, production activity representation module, production activity classification module, which can identify and distinguish five typical agricultural production activities in video, including Planting, fertilizing, pest control, irrigation, harvesting. This work will help improve the authenticity and accuracy of agricultural product quality traceability system.