A Study of Ultrasonic Sensors to Intelligent Estimation of Tree Canopy Volumes

Journal of Agricultural Machinery

M Eskandari¹  A Hosainpour¹ 

[1] University of Ilam;

关键词: Precision agriculture;    Variable rate technology;    Ultrasonic sensor;    Artificial neural network;    Tree canopy;   

DOI  :  10.22067/jam.v5i1.24956

来源: DOAJ

【 摘 要 】

Many research projects have been conducted about using ultrasonic sensors to estimate canopy volume. This study investigates using software applications such as artificial neural network (ANN) to improve the estimation of canopy volume by using ultrasonic sensors. A special experimental system was built. The system had three ultrasonic sensors mounted vertically on a wooden pole with an equal distance of 0.6 m. As the wooden pole moves with a constant speed, the ultrasonic sensors measure the thickness of tree canopy with sampling rate of 4 Hz. Experiments were conducted on 5 samples of Benjamin tree at three speed levels of 35,45 and 55 cm s-1 in three replications. The real volume of trees was measured manually with rectangular elements method. After a full passing of ultrasonic sensors, potential features such as canopy diameter, average width of tree canopy and height of the tree canopy were considered as the inputs to the ANN model and the manually volume as the output of the model. Optimal ANN model was selected based on mean square error and correlation coefficient. The results showed that 13-16-7-1 was the optimal neuron numbers in ANN topology for estimating canopy volume.

【 授权许可】

Unknown