Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Application of Region-based Convolution Neural Network on Tea Diseases and Harming Insects Identification

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  2021 ASABE Annual International Virtual Meeting  2100872.(doi:10.13031/aim.202100872)
Authors:   Xue-Ming Chen, Chuan-Che Lin, Shiou-Ruei Lin, Shih-Fang Chen
Keywords:   Tea disease, foliar lesion, lesion identification, deep learning, Faster R-CNN, Cascade R-CNN


Tea diseases and harming insects are major factors that cause foliar lesions and lead to tremendous economic loss. It is essential to identify the lesion causes in a relatively early stage so that proper actions can be taken. Conventionally, lesions were identified manually by phytopathologists or entomopathogists. However, professionals are far below demand. Thus, an automatic identification approach is needed. This study aimed at developing a real-time identification system to identify the causes of lesions and the severity stages of the lesions for tea leaves. A dataset composed of 4295 images was collected. The causes of the lesions included three diseases and six harming insects, and they were categorized into 12 classes. Faster region-based convolution neural network (FRCNN) and cascade RCNN (CRCNN) were used in the study to identify the causes of the lesions. The FRCNN model achieved a mean average precision (mAP) of 74.0%. The CRCNN model achieved a mAP of 76.6%. The CRCNN model was hosted on a mobile device application to provide the service of tea pest identification to the public. The developed system provides a real-time mobile device application service to help a quick identification of tea foliar diseases and pest damage to the public.

(Download PDF)    (Export to EndNotes)