Electron Microscope Image Processing Of Plant Pathogen

Image processing of plant pathogens and their length measurement play an important role in the study of the mechanism of action of pesticides.In order to further study the intelligent analysis of plant pathogen images,the main research work and results are as follows:(1)Use the YOLOv4 model to perform target detection on electron microscope images of plant pathogens.According to the morphological characteristics of pathogens before and after the binary fission period,they are divided into two categories.The position of each pathogen detection frame is obtained,the influence of the external environment is excluded,and better detection results are obtained through the analysis of evaluation indicators such as recall rate and m AP.(2)Using the improved Deep Lab V3+ image edge segmentation model to complete the edge segmentation of pathogen electron microscope images.(i)Use Mobile Net V2 as the model backbone feature extraction network,and use the depthwise separable convolution in the inverse residual structure to replace the ordinary convolution to reduce the model calculation amount;(ii)Improve the feature learning efficiency through the multi-scale connection and depth supervision structure To characterize the ability and accuracy,introduce shallow effective features in the enhanced feature extraction network,and fuse the effective feature layers of different jump layers to improve the segmentation accuracy of small-scale targets in the contour of pathogens;(iii)Introduce a channel attention mechanism to globally pool feature channels It becomes the global effective feature layer on the channel dimension to obtain the final feature layer,which strengthens the learning ability and generalization ability of the network model.Through experimental comparison with the classic image segmentation models Deep Lab V3+,Unet,PSPNet and other models,the results show that the improved Deep Lab V3+ model has improved segmentation accuracy and effect.(3)According to the contour of pathogenic bacteria after segmentation,find the minimum circumscribed rectangle of the contour of pathogenic bacteria according to the problem of different degrees of inclination of pathogenic bacteria,and measure its length based on the length and width of the rectangle;The distance between the pixel points between the endpoints is used as the basis for the calculation of the diaphragm width.The length,width and circumference of the two types of pathogens were measured respectively,and the actual length of the pathogens was obtained by converting the image size and pixel ratio,and the results were compared with the results of manual manual measurement.