Wheat Ear Recognition&Counting Based On UAV Images

Wheat is one of the most important food crops in China.Maintaining high wheat yield is of great significance for improving the living standards of the Chinese people,ensuring social stability.and promoting development of the domestic economy.The use of information technology to quickly and accurately assess wheat yield has important research significance.Using image processing technology to identify and count wheat ears in remote sensing images collected by VAU can save a lot of time,manpower and material resources,and can also overcome weakness of subjective,empirical,unsustainable manual testing.This study is based on the use of Dajiang-Wu2 UAV with camera parameters 35mm-90mm,in the experimental field of Wenxian County,Henan Province,to collect remote sensing images.A number of different contrast images in different period were selected for research.Area estimation wheat identification and counting.This paper is to study the histogram in different color spaces,and select the RGB-grayscale,Lab-L,HSV-V color space with better segmentation effect.Multiple threshold segmentation algorithm was used to achieve accurate segmentation of wheat ears,land and wheat leaves.At the same time,combined with morphological characteristics of wheat ear,it can be further processed,which can further improve the accuracy of wheat ear segmentation.In order to verify the accuracy of the segmentation,the images were manually labeled and counted,and verified with the machine counting.A large number of experimental data show that the proposed segmentation method is feasible.the wheat wheat ear identification,and the average accuracy reaches 96.55%.The experimental results meet the experimental data requirements and provide technical support for the wheat yield estimation.