基于嵌入式设备的作物长势监测系统

Crop Growth Monitoring System Based on Embedded Devices

传统的基于机器视觉的作物长势监测设备对环境要求较高,且体型较大,不易布置在实际生产的复杂环境中。本文提出了一套基于嵌入式设备的应用于实际生产环境的作物长势监测系统,该系统集成了图像采集、滤波、颜色空间转换、图像分割、形态学运算、特征量化的图像处理流程,依靠人工在嵌入式视觉系统中面向环境自定义算法提取作物长势特征,确定最优算法组合,设定流程,进行作物长势特征的自动提取并监测。利用该监测系统对拔节期玉米植株的株高进行连续监测,监测值与人工测量值的决定系数达到0.9072,随时间的变化趋势一致,表明该系统可用于复杂环境下的作物长势监测。

The traditional equipments monitoring crop growth based on computer vision are mostly big and have high requirements to environment, so they are difficult to use in complex agricultural environments. A set of crop growth monitoring equipment applied to the practical production environment was developed based on embedded devices in this paper. It integrated the picture processing flow including image acquisition, filtering, color space conversion, image segmentation, morphological operations, and feature quantification. The feature of plant growth was extracted by user-defined algorithm artificially in the embedded vision system. Through confirming the optimal algorithm combination, the crop growth feature could be monitored with this system. The plant height of maize grown in greenhouse at joining stage was continuously monitored using the crop growth monitoring system based on embedded devices. The data measured by embedded vision system showed a good correlation (R 2=0.9072) with that obtained by artificial measurement, and they showed similar chang trends with the extension of time. It indicated that the system based on embedded devices could be used for monitoring crop growth in complex agriculture environments.