Net Bloch, Leaf Scald and Healthy Malting Barley Leaves
Description
Net Bloch infected, Leaf Scald infected and Healthy Malting Barley Leaves dataset it collection of 3012 Malting Barley leaves that are healthy, net blotch infected and leaf scald infected leaves. The dataset is collected starting from August 2021 to December 2021. It is collected from Debre Berhan University Malting Barley farm. The two classes of diseases that are gathered in this dataset are the most fatal malting barley diseases causing a drastic loss in both quality and yield of the crop. The dataset is gathered by using smart phone camera and digital camera as per the conditional requirement while collection. To minimize the time to segment unwanted features out of the image, white paper is used as a background while taking the pictures. Using this dataset CNN based Malting Barley Disease detection is studies with an aim of increasing the performances of CNN models for detecting potential Malting Barley diseases. CNN pretrained models for mobile platforms such as EfficientNetB0, MobileNetV2 and NASNetMobile have been experimented to have an outperforming CNN model . Based on the experiment work EfficientNetB0 has scored a 96% accuracy which is 2 percent increase than the state-of-the -art CNN pretrained model for mobile platforms. In the research work the dataset is portioned in to two classes namely: training and testing dataset. the ratio of partitioning is 90% to 10% respectively. Then K-fold cross validation is used for building sampling independent robust model. Data augmentation technique is used to increase the number of training image dataset. After augmentation, the training image dataset transformed from 2,711 to 13,555. The augmentation technique comprises different arguments such as rotation range, zoom range and width shift. In order to use this dataset you should first understand the problem that you are longing to solve. If it is related with plant disease detection especially Malting Barley, congrats you have a very reliable dataset for your model input. The data can be used after some preprocessing activities based on the researcher's interest. Thank you very much.
Files
Steps to reproduce
Net Bloch infected, Leaf Scald infected and Healthy Malting Barley Leaves dataset it collection of 3012 Malting Barley leaves that are healthy, net blotch infected and leaf scald infected leaves. The dataset is collected starting from August 2021 to December 2021. It is collected from Debre Berhan University Malting Barley farm. The two classes of diseases that are gathered in this dataset are the most fatal malting barley diseases causing a drastic loss in both quality and yield of the crop. The dataset is gathered by using smart phone camera and digital camera as per the conditional requirement while collection. To minimize the time to segment unwanted features out of the image, white paper is used as a background while taking the pictures.