Time-series change detection model of lettuce canopy area in acontrolled environment

摘要

Dynamical Information for crop growth is essential for exploring the mysteries of crop growth and assessing yield potential. However, due to the environmental characteristics of the controlled environment, the main difficulties in achieving monitoringcrop growth are the high cost and complicated process, which are required to improve the image acquisition equipment. Besides, there have been few studies on accurately monitoring of crop dynamical growth in the controlled environment. To solve this problem, we propose a new model for detecting the canopy leaf area of purple lettuce from time-series imagery under a controlled environment. In this study, purple lettuce was chosen as the research object Images of the purple lettuce were taken at 21: 00 every day from the top. At first, the source images are registered with the help of a 25cm * 25cm blue profile to evaluate the change of canopy leaf area on the same spatial scale. Then the segmentation of canopy leaves is based on the characteristics ofpurple lettuce at different growth stages and the relationship with the surrounding environment. Finally, the color card is used as a reference to estimate the real canopy leaf area. Experiment results show that the proposed model is affected by heterogeneous brightness. The central contribution of the paper is a prospective study on the monitoring of the canopy leaf area during the whole growing period. To some extent, this study has a positive effect to promote intelligent visual monitoring in a controlled environment.