top of page

Cassava Mosaic Virus Detection Using AI and Remote Sensing

Vietnam, Cambodia and Thailand are all leading exporters of cassava products, owing to their favourable tropical climate and fertile soil, which provides ideal conditions for cassava cultivation. However, in recent years, the emergence of the cassava mosaic virus (CMV) has posed a significant threat to cassava production. CMV is primarily transmitted by the vector whitefly (Bemisia tabaci). These whiteflies can rapidly spread the virus through a cassava field or to neighbouring fields, particularly in the presence of high whitefly populations.

The early detection of this virus has important implications for farmer incomes and provides valuable intelligence to governments, crop traders and crop buyers on the likely size of the crop and hence the direction of prices.

CMV can cause significant yield losses in cassava crops, which can range from 30 to 50 percent, from the secondary infection. The virus can cause a variety of symptoms in cassava plants, including distortion of the leaf lamina, chlorotic mosaics, yellowing, mottling of leaves, as well as stunted growth and reduced tuber production. Likewise, tubers' starch and protein contents can be affected, with the protein content being reduced by 15% compared with healthy tubers. In severe cases, CMV can cause the death of the cassava plant, leading to total crop loss. As cassava is a vital crop for many farmers in developing countries, the impact of CMV can be devastating, resulting in food shortages and economic hardship.

An artistic rendering of CMV-infected cassava plants with yellowing leaves. Source: Wikimedia Commons

Fortunately, advances in technology, such as the use of AI and remote sensing, offer promising solutions to detecting and managing CMV outbreaks. Using these technologies, we are able to detect early signs of infestation, allowing us to implement control measures that are more targeted and effective. For example, remote sensing technology can detect changes in the health of cassava plants that may indicate the presence of the virus. This can be done through satellite data that captures the spectral signatures of the plants, allowing for the detection of any changes that may be indicative of disease.

Furthermore, AI and remote sensing technologies can help promote sustainable farming practices. With the help of machine learning algorithms, precision agriculture techniques can target specific areas that require treatment, reducing the amount of chemicals used and minimising the impact on the environment. This approach not only benefits farmers by improving crop yields and quality but also helps to ensure the long-term sustainability of cassava cultivation.

Adatos has developed a CMV model using historical CMV data in Asia, consisting of a total of 211 ground truth samples from 2018 to 2020. Through this analysis, we identified high-risk areas for CMV infection over a large area. By accurately identifying the presence of CMV, the model can facilitate early detection and help prevent the further spread of the virus, particularly in high-risk areas. It can also generate valuable data for insurance companies offering crop insurance policies to cassava farmers, including risk assessment, loss prevention, and claims management services.


bottom of page