Abstract
It is important to obtain accurate information about kiwifruit vines to monitoring their physiological states and undertake precise orchard operations. However, because vines are small and cling to trellises, and have branches laying on the ground, numerous challenges exist in the acquisition of accurate data for kiwifruit vines. In this paper, a kiwifruit canopy distribution prediction model is proposed on the basis of low-altitude unmanned aerial vehicle (UAV) images and deep learning techniques. First, the location of the kiwifruit plants and vine distribution are extracted from high-precision images collected by UAV. The canopy gradient distribution maps with different noise reduction and distribution effects are generated by modifying the threshold and sampling size using the resampling normalization method. The results showed that the accuracies of the vine segmentation using PSPnet, support vector machine, and random forest classification were 71.2%, 85.8%, and 75.26%, respectively. However, the segmentation image obtained using depth semantic segmentation had a higher signal-to-noise ratio and was closer to the real situation. The average intersection over union of the deep semantic segmentation was more than or equal to 80% in distribution maps, whereas, in traditional machine learning, the average intersection was between 20% and 60%. This indicates the proposed model can quickly extract the vine distribution and plant position, and is thus able to perform dynamic monitoring of orchards to provide real-time operation guidance.
Highlights
This study proposes an effective plant location distribution model (PLD-M) to interpret kiwifruit remote sensing images obtained using a low-altitude unmanned aerial vehicle (UAV)
The random forest (RF) and support vector machine (SVM) are more sensitive to the fine features than the depth semantic segmentation, and the distribution range of the vine pixels in the correct region is more accurate
The use of a deep learning method increased the quality of the segmentation image, and segmentation of kiwifruit vines, and the prediction of the possible canopy distribution
Summary
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Different sensors that are carried by UAVs have been used to identify and count trees [4], and to determine the tree height and crown size [5,6] This form of information collection, which is characterized by low data accuracy, is often used in the field of forestry management, which does not require accurate management of each tree. The method used the target recognition and segmentation of the CNN to realize the extraction and calculation size of a bare apple tree crown. This study proposes an effective plant location distribution model (PLD-M) to interpret kiwifruit remote sensing images obtained using a low-altitude UAV.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
