Enhanced Field-Based Detection of Potato Blight in Complex Backgrounds Using Deep Learning

Rapid and automated identification of blight disease in potato will help farmers to apply timely remedies to protect their produce.Manual detection of blight disease can be cumbersome and may require trained experts.To overcome these issues,we present an automated system using the Mask Region-based convolutional neural network(Mask R-CNN)architecture,with residual network as the backbone network for detecting blight disease patches on potato leaves in field conditions.The approach uses transfer learning,which can generate good results even with small datasets.The model was trained on a dataset of 1423 images of potato leaves obtained from fields in different geographical locations and at different times of the day.The images were manually annotated to create over 6200 labeled patches covering diseased and healthy portions of the leaf.The Mask R-CNN model was able to correctly differentiate between the diseased patch on the potato leaf and the similar-looking background soil patches,which can confound the outcome of binary classification.To improve the detection performance,the original RGB dataset was then converted to HSL,HSV,LAB,XYZ,and YCrCb color spaces.A separate model was created for each color space and tested on 417 field-based test images.This yielded 81.4%mean average precision on the LAB model and 56.9%mean average recall on the HSL model,slightly outperforming the original RGB color space model.Manual analysis of the detection performance indicates an overall precision of 98%on leaf images in a field environment containing complex backgrounds.

Detection of the Progression of Anthesis in Field-Grown Maize Tassels:A Case Study

The tassel of the maize plant is responsible for the production and dispersal of pollen for subsequent capture by the silk(stigma)and fertilization of the ovules.Both the amount and timing of pollen shed are physiological traits that impact the production of a hybrid seed.This study describes an automated end-to-end pipeline that combines deep learning and image processing approaches to extract tassel flowering patterns from time-lapse camera images of plants grown under field conditions.Inbred lines from the SAM and NAM diversity panels were grown at the Curtiss farm at Iowa State University,Ames,IA,during the summer of 2016.Using a set of around 500 pole-mounted cameras installed in the field,images of plants were captured every 10 minutes of daylight hours over a three-week period.Extracting data from imaging performed under field conditions is challenging due to variabilities in weather,illumination,and the morphological diversity of tassels.To address these issues,deep learning algorithms were used for tassel detection,classification,and segmentation.Image processing approaches were then used to crop the main spike of the tassel to track reproductive development.The results demonstrated that deep learning with well-labeled data is a powerful tool for detecting,classifying,and segmenting tassels.Our sequential workflow exhibited the following metrics:mAP for tassel detection was 0.91,F1 score obtained for tassel classification was 0.93,and accuracy of semantic segmentation in creating a binary image from the RGB tassel images was 0.95.This workflow was used to determine spatiotemporal variations in the thickness of the main spike—which serves as a proxy for anthesis progression.

TasselGAN:An Application of the Generative Adversarial Model for Creating Field-Based Maize Tassel Data

Machine learning-based plant phenotyping systems have enabled high-throughput,non-destructive measurements of plant traits.Tasks such as object detection,segmentation,and localization of plant traits in images taken in field conditions need the machine learning models to be developed on training datasets that contain plant traits amidst varying backgrounds and environmental conditions.However,the datasets available for phenotyping are typically limited in variety and mostly consist of lab-based images in controlled conditions.Here,we present a new method called TasselGAN,using a variant of a deep convolutional generative adversarial network,to synthetically generate images of maize tassels against sky backgrounds.Both foreground tassel images and background sky images are generated separately and merged together to form artificial field-based maize tassel data to aid the training of machine learning models,where there is a paucity of field-based data.The effectiveness of the proposed method is demonstrated using quantitative and perceptual qualitative experiments.