Research On Application Technology Of UAV-Study On Comprehensive Monitoring Techniques Of Pine Wood Nematode Disease

Pine trees are some of the most important tree species in Chinese forests.They are diverse in species and are widely distributed all over the country.It is believed that the total area occupied by these trees alone accounts for more than 70%of the total area of Chinese forestry plantations in China.In addition to their economic benefits,pine trees are also known for their ornamental and medicinal values.Pine wood nematodes/pine wilt nematodes commonly referred to as Bursaphelenchus Xylophilus are one of the natural enemies of pine tree species.These are the types of nematodes which infect pine trees and cause dangerous and stringent diseases.They have the characteristics of rapid spread,high mortality and difficulty to prevent and control.As a result,a large number of pine resources are infringed every year and are therefore called "pine cancers".Research has confirmed that the lack of early monitoring is one of the major factors causing the rapid spread of pine wood nematode disease as the best time for prevention and control is often missed.In the light of the above issues,this paper uses the UAV as a flight platform to carry out research on the comprehensive monitoring technology of pine wood nematode disease,including a series of research such as image acquisition,image splicing as well as pest and disease identification using drones.Test areas including Jiukuang Village,Xudun Town,Jianou City in Fujian Province are used to conduct this research using the aforementioned monitoring technology.The specific work conducted in this research study includes the following aspects:(1)An unmanned aerial vehicle(UAV)is used as an image acquisition platform.The operator determines the key parameters in route design such as the relationship between ground resolution(GSD)and altitude,heading overlap and speed,as well as the relationship between the degree of lateral overlap and the distance between routes.As soon as the images are collected,a thorough screening and removal of all the captured images is proceeded to ensure the image quality of the drone and the complete acquisition of all the necessary images of the test area.(2)Learning the general process of image splicing focusing on the study and research of image splicing technology based on 3D reconstruction.On the Pix4DMapper image splicing software platform,a point matching,three-dimensional sparse point cloud reconstruction,point cloud encryption,texture mapping as well as orthographic image output are performed on the images acquired by the drone to obtain the final panorama of the study area.(3)An image segmentation algorithm based on the combination of the hyper-green feature factor and the maximum between-class variance method(ExG+Otsu)is proposed.Based on the three-dimensional color space of RGB,the algorithm first performs image background processing obtained from the UAV and several actions are undertaken based on different background information in the image.For example,gray images are processed by the ExG algorithm to enhance the target.There is also a contrast between information and background information as well as the filtering and de-noising of the images.Finally,the largest between-class variance method(Otsu)to generate binary images containing only "0"and "1" information is used.(4)An analysis method of disease condition of pine wood nematode is proposed based on the color characteristics.This means that the area percentage of the pine trees occupied by the disease is calculated.Most importantly,this paper proposes different disease screening formulas and maps the calculated results to the disease severity level table to determine the degree of disease of the image for different background information contained in UAV images combined with processed binary images.The results show that this method can complete the identification of the disease and the degree of the disease using a single drone image.(5)A method for determining the disease level of panoramas is proposed and the algorithm is developed.Firstly,the panoramic image of the study area is divided by M×N image segmentation mode;then the disease acuity is discriminated for each segmented image using the aforementioned algorithm.Finally,different screening results are expressed in different colors to the panorama in the picture.Comparisons between calculated and visual observation results indicate that the accuracy of this method is about 97%.The introduction of drone technology has increased the reliability and timeliness of pine wood nematode disease surveillance data in forests,not only to prepare for the early monitoring and prediction of pine wood nematode disease but also to prevent pine wood nematode in the future.It also plays an important role in the prevention and treatment of diseases.