AUTHOR=Li Xudong , Zhou Yuhong , Liu Jingyan , Wang Linbai , Zhang Jun , Fan Xiaofei 

TITLE=The Detection Method of Potato Foliage Diseases in Complex Background Based on Instance Segmentation and Semantic Segmentation

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 13 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2022.899754

DOI=10.3389/fpls.2022.899754

ISSN=1664-462X

ABSTRACT=Abstract: Potato early blight and late blight are devastating diseases affecting potato planting and production. Precise diagnosis of the diseases is critical in treatment application and management of potato farm. Traditional computer vision technology and pattern recognition methods have certain limitations in detecting crop diseases. In recent years, the development of deep learning technology and convolutional neural networks has provided new solutions for the rapid and accurate detection of crop diseases. This paper devised an integrated framework that combines instance segmentation model, classification model and semantic segmentation model to realize the segmentation and detection of potato foliage diseases in complex backgrounds. In the first stage, Mask R-CNN was used to segment potato leaves in complex backgrounds. In the second stage, VGG16, ResNet50 and InceptionV3 classification models were employed to classify potato leaves. In the third stage, UNet, PSPNet and DeepLabV3+ semantic segmentation models were utilized to segment potato leaves. In the end, the three-stage models were combined to segment and detect the potato leaf diseases. The experimental results proved that the average precision (AP) obtained by the Mask R-CNN network in the first stage was 81.87%, and the precision was 97.13%. The accuracy of the classification model in the second stage was 95.33%. The mean intersection over union (MIoU) of the semantic segmentation model in the third stage was 89.91%, and the mean pixel accuracy (MPA) was 94.24%. It provides a new model framework for the identification and detection of potato foliage diseases in natural environment, and provides a theoretical basis for potato disease assessment and classification.