Study On The Design Of Wheat Stripe Rust Summer Spores Microscopic Image Acquisition Device And The Image Processing Method

Wheat stripe rust has always been an important threat to the high quality and high yield of wheat in China.Due to the lack of early forecasting technology,the incidence of wheat stripe rust is widespread and serious.The number of rust spores of stripe rust is a direct factor influencing the incidence and spread of wheat stripe rust.Traditional wheat stripe rust spore counting requires labor and material resources,and it is difficult to achieve real-time online counting.In order to realize the remote real-time rapid determination of the number of spores in air,this study carried out the design of wheat stripe rust spores microscopic image acquisition device and image processing methods.The main research contents and conclusions are as follows:(1)The automatic acquisition and transmission device of spore microscopic image of wheat stripe rust was designed.AT89C52 RC MCU(Micro Control Unit),42BYG250 C hybrid two-phase four-wire stepper motor,TB6560 driver board,Intel NUC5 CPYH mini-host,power supply,etc were used to design an automatic acquisition and transmission device for spores microscopic images of wheat stripe rust based on AT89C52 RC MCU.This device controls the stepping speed,steering and delay time of the stepping electrons through a single-chip microcomputer.Through the Supereyes interface,the time interval for automatic electron microscope photographing can be set.After collecting spores,then the electron microscope captures the spore images and saves them to a local folder,which are transmitted to the cloud disk via the 3G/4G network.After debugging and optimization,clear spore microscopic images of stripe rust can be obtained,and the image can be uploaded to the cloud disk in real time.The experimental results show that under a good network condition,the transmission success rate is 100%,and the average transmission time is 1 minute.The microscopic image acquisition device and the image transmission platform can acquire a clear target image in real time,which lays a foundation for the source count of spores of wheat stripe rust.(2)The research on spore microscopic image processing methods of wheat stripe rust was carried out based on the machine vision technology.Through image enhancement,segmentation,and morphological processing methods,the spore target images with clear edges were obtained.The K-means segmentation algorithm based on L*a*b* color space,Otsu segmentation algorithm,and edge detection algorithm are proposed after comparison of multiple segmentation methods.The segmentation results show that the K-means segmentation algorithm based on L*a*b* color space works well,is suitable for most spores microscopic images,and can well identify active spores and inactivated spores.Morphological operation is used to process the segmented image,including morphological opening and closing operations,removal of small areas,hole filling,and other operations.Image denoising and spore edge smoothing were achieved,laying the foundation for accurate spore counts.(3)The method for counting spores is carried out.Distinguish signal spore from adherent spores by shape factor.The improved Harris corner detection and corner detection based on centroid-edge point distance are used for the count of adherent spores.Besides,a marker erosion method is proposed for spore counting.This method eliminates the need to distinguish signal spore from adherent spores and allows accurate counting of multiple adherent spores.The experimental results show that the average accuracy rate of spore counts based on the marker corrosion method is 94.46%,which is 3.46% higher than the improved Harris corner detection,and 1.86% higher than the corner point detection based on the centroid-edge point distance.This method can better solve the problem of counting counts of severely connected spores.(4)Using GUI compiler toolbox in MATLAB,wheat stripe rust spore microimage processing system is designed.The experiments show that this system can accurately achieve the comparison of multiple methods of image enhancement and segmentation,the process of image morphology processing,a variety of spore count comparison methods and other functions.