Cassava disease recognition from low‐quality images using enhanced data augmentation model and deep learning

Abstract

Improvement of deep learning algorithms in smart agriculture is important to support the early detection of plant diseases, thereby improving crop yields. Data acquisition for machine learning applications is an expensive task due to the requirements of expert knowledge and professional equipment. The usability of any application in a real-world setting is often limited by unskilled users and the limitations of devices used for acquiring images for classification. We aim to improve the accuracy of deep learning models on low-quality test images using data augmentation techniques for neural network training. We generate synthetic images with a modified colour value distribution to expand the trainable image colour space and to train the neural network to recognize important colour-based features, which are less sensitive to the deficiencies of low-quality images such as those affected by blurring or motion. This paper introduces a novel image colour histogram transformation technique for generating synthetic images for data augmentation in image classification tasks. The approach is based on the convolution of the Chebyshev orthogonal functions with the probability distribution functions of image colour histograms. To validate our proposed model, we used four methods (resolution down-sampling, Gaussian blurring, motion blur, and overexposure) for reducing image quality from the Cassava leaf disease dataset. The results based on the modified MobileNetV2 neural network showed a statistically significant improvement of cassava leaf disease recognition accuracy on lower-quality testing images when compared with the baseline network.The model can be easily deployed for recognizing and detecting cassava leaf diseases in lower quality images, which is a major factor in practical data acquisition.

Drone-Based Water Level Detection in Flood Disasters

“…On the other hand, we apply the noise-adding data augmentation technique to the selected and annotated images by introducing random noise to the original images. This is done to generate synthetic low-quality versions of the original image that contribute during training to improve the skills of the deep models to cope with blurred images that might be captured at the online stage (as proved in [ 33 , 38 , 39 , 40 ]).…”

EfficientNet: A Low-bandwidth IoT Image Sensor Framework for Cassava Leaf Disease Classification

“…Lightweight models, similar to our framework, which are more practical for developing countries, include a mobile implementation using a single-shot detector called MobileNet to identify foliar symptoms of cassava leaf diseases (12) and a recognition model using low-quality images that are augmented with enhanced data techniques. (13) The mobile implementation (12) requires sufficient training examples from diverse data sources to improve its detection accuracy and better capture the diversity of cassava leaf diseases that occur in the real world. The methodology using low-quality images (13) attempts to solve the recognition problem by using a novel image color histogram transformation technique to generate synthetic images, which is based on the convolution of Chebyshev orthogonal functions with the probability distribution functions of image color histograms.…”

“…(13) The mobile implementation (12) requires sufficient training examples from diverse data sources to improve its detection accuracy and better capture the diversity of cassava leaf diseases that occur in the real world. The methodology using low-quality images (13) attempts to solve the recognition problem by using a novel image color histogram transformation technique to generate synthetic images, which is based on the convolution of Chebyshev orthogonal functions with the probability distribution functions of image color histograms. Our method is similar in that we use color transformations for normalization but do not rely on color histograms.…”

Classification of Plant Leaves Using New Compact Convolutional Neural Network Models

“…Detection of Cassava plant disease by [ 36 ] achieved an accuracy of 96.75% with the deep residual neural network model. Alli et al [ 37 ] used a data augmentation method to achieve an accuracy of 99.7% for cassava plant disease classification using MobileNetV2. Pearl millet disease classification with an automated method of collecting the pearl millet data from the farm and classifying the disease with a Custom-Net model with an accuracy of 98.78% [ 38 ].…”