Crop canopy temperature model of ditch-cultivated based on artificial neural network
Min Zhang1, Qiang Fan2, Fucang Zhang3, Xia Li1, Xuzhang Xue4, Guodong Wang1
1College of Science, Northwest A&F University, Yangling 712100, Shaanxi, China
2College of Water Resources and Architectural, Northwest A&F University, Yangling 712100, Shaanxi, China
3Key laboratory of Agricultural Soil and Water Engineering in Arid and Semiarid Areas of Ministry of Education, Northwest A&F University, Yangling 712100, Shaanxi, China
4National Engineering Research Centre for Information Technology in Agriculture; Beijing 100097, China
Aiming at the mechanism model are influenced by multiple random factors, this paper establishes canopy temperature models based on BP network and RBF network respectively. The models take the temperature, humidity, illumination, soil temperature and ditch depth in the closed greenhouse as input neurons and takes canopy temperature as the output neuron. The results show that both models can well predict ditch-cultivated crop canopy temperature. The mean error between the simulation value and measured value of BP network model is 0.8408℃, and root-mean-square error of 0.5789℃. Actual output and expected output of RBF network model differ little, mean error of 0.2236℃ and root-mean-square error of 0.3496℃. In contrast, RBF network model can more accurately predict crop canopy temperature of ditch-cultivated than BP network model.