Assessment Of Canopy Vigor Information From Vine Crops Based On Digital Surface Models From Unmanned Aerial Vehicle Imagery

Grape and kiwifruit are important cash value crop and their cultivation has become a major agricultural industry that promotes the economic development of China.Rapid and effective acquisition of information about canopy vigor and growth of these kind of vine crops are critical to assess the potential impacts of biotic or abiotic stresses on plant development.Therefore,it is significant to acquire canopy vigor information of vineyard and kiwifruit orchard rapidly and accurately,as well as making quick and timely judgment on crop growth condition,which are beneficial to estimate frost damage loss,improving management and increasing yield.By implementing a Digital Surface Model(DSM)to imagery obtained using Unmanned Aerial Vehicles,it is possible to filter canopy information effectively,which provides an efficient method to extract canopy from ground background.Therefore,this paper takes the main cash fruit grape and kiwifruit in arid and semi-arid regions of Northwest China as the research objects,and systematically studied the monitoring method of crop canopy vigor based on DSM.The main research contents and conclusions are as follows:(1)Several image processing algorithms were applied to extract the canopy region of the vine crops.According to the height difference between the canopy area of crops and the surrounding ground background,several image segmentation algorithms were applied to DSM of vineyard and kiwifruit orchard,the algorithms include global threshold method,roughness algorithms,adaptive threshold algorithms and bottom cap algorithms.The experimental results showed that the segmentation result of the global threshold method is not effective because of the obvious surface fluctuation and complex background in the study area;roughness algorithm has a good effect on the extraction of vineyard,the extraction error is 5.1%,and the coincidence degree with the original canopy area is 94.6%,but it cannot extract the canopy area of kiwifruit where connected by canopy layer on adjacent planting rows;the adaptive threshold algorithm can extract the canopy areas of kiwifruit more effectively,the extraction error is 12.7%,and the coincidence degree with the original canopy area is 84.6%,but there are many phenomena where the ground background is misjudged as a crop;the difference between the bottom cap algorithm and other algorithms is the biggest,but it is rare to misjudge ground background information as crop.(2)Canopy vigor of the vine crops was graded and evaluated according to the extracted canopy area of the vine crops.According to the row spacing and plant spacing of grapevine and kiwifruit,a set of reference points for canopy vigor monitoring and the regions of interest corresponding to the reference points were generated by linear interpolation method on the obtained image of crop canopy area,the thresholds of canopy vigor division for grapevine and kiwifruit canopies were determined,and canopy vigor monitoring for vine crops was implemented,finally,the accuracy was evaluated.The results showed that the accuracy of canopy vigor classification using different algorithms were different.For vineyard,the best overall accuracy of canopy vigor classification is based on the roughness algorithm which performed an accuracy of 92.0%,and for kiwifruit orchard,the best overall accuracy of canopy vigor classification is based on the adaptive threshold algorithm which performed an accuracy of 89.5%.(3)Geographical maps were carried out based on the classification results of canopy vigor of vine crops.Firstly,the results of canopy vigor classification were registered,which transformed the image coordinates into projection coordinates.Then spatial data of the results were generated which could be displayed,analyzed and managed in GIS software,the geographic process has been completed.Finally,Kriging interpolation method was used to analyze the stress degree of the grapevine and kiwifruit canopy according to the relationship between the canopy vigor level and biotic or abiotic stresses.Canopy stress maps were generated and the results were displayed in a visual form.