Image classification system for cassava disease detection

Abstract

As the second-- largest source of carbohydrates in Africa, cassava is a key food security crop grown by smallholder farmers because it can withstand harsh conditions. At least 80% of household farms in Sub Saharan Africa grow this starchy root, but viral diseases are major sources of p oor yields. With the help of data science, it may be possible to identify common diseases so they can be treated. In the agriculture environment, detection and classification of plant diseases plays an important role. Cassava farmers and Agricultural researchers in Uganda need to quickly adapt to new technologies for diagnosing diseases in the leaves at an early stage. One option is to switch to a supervised machine learning method that will perform disease diagnosis and results returned back to them in nea r real time. Good quality data is essential to creating good algorithms for the classification of cassava leaf diseases. Noise from the environment (leaves, people, dirt) might confuse the algorithm therefore deteriorating its performance. Plant Diseases c an be managed by identifying them as soon as they appear on the plant. In addition, with the rise of the internet and mobile technology worldwide, it's easy to access diagnosis information on a particular type of disease. As a result, the prevalence of smart phones with powerful cameras can help to scale up any type of solution that involves crop disease detection making it feasible and practical. We propose a machine learning model to automatically segment cassava leaf images and extract the background clutter thus ensuring good quality data is collected for disease classification.