论文部分内容阅读
地上部分生物量是水稻生长研究中的重要参数,传统的测量方法主要依靠人工剪取称重,不仅具有破坏性,而且费时费力。提出了一种改进的基于图像特征参数的生物量预测模型,并比较了其在分蘖期和拔节期的效果。优于使用单幅图像投影面积预测的方法,该模型使用多幅侧视图像投影平均值和顶视图像来降低植株不对称带来的影响。基于分蘖期和拔节期的两批数据,新模型的预测平均相对误差(MAPE)达到9.26%,决定系数(R2)为0.93,优于其他模型。实验结果还显示出,当水稻生育期跨度较大时会造成植株结构差别较大,进而影响生物量的预测效果。
Aboveground biomass is an important parameter in the research of rice growth. The traditional measurement methods rely mainly on manual cutting and weighing, which is not only destructive but also laborious and time-consuming. An improved biomass prediction model based on image feature parameters was proposed and its effects at tillering and jointing stages were compared. This method is superior to the method using a single image projected area, which uses multiple side-view image projection averages and top-view images to reduce the impact of plant asymmetry. Based on two batches of data at tillering and jointing stages, the predicted mean relative error (MAPE) of the new model reached 9.26% and the determination coefficient (R2) was 0.93, which was better than other models. The experimental results also show that when the growth period of rice is large, the plant structure will be greatly different, thus affecting the prediction effect of biomass.