Last year the Desert Locusts, also known as the migratory locusts, swarmed the African, Middle Eastern and Asian landscape leaving behind devastation. The victims have individually decided to tackle the crisis using many tools. The tools can be categorized into two main sections; one section gives the combating and on-site fight against the locusts stampede in the shape of pesticides, smoke, sound and even using ducks and wattled starlings to eat away the insect invasion; the other section is the employment of pre-planning tools to launch a surprise maneuver to stop the trajectory of the locusts from one place to another and place a pause on their crop havoc.
The face-to-face battle is a conventional tool which is a hectic task and countries like Pakistan, India, Kenya, Nigeria and Iran have been unsuccessful in preventing the damage left by them. Thus, Kenya have shown a new way to tame these mini beasts by using machine learning and satellite imagery. A residential company have developed a software named, Kuzi, a Swahili name (Kenya’s national language) for the locust eating wattled starling. Kuzi is an initiative developed and funded by the Nairobi-based company Selina Wamucii, an online platform that integrates with cooperatives, farmers groups, agro-processors, and other organizations that work directly with Africa’s family farmers, pastoralists, and fishing communities to form a valuable link to markets around the world. Selina Wamucii is the pan-African marketplace for food and agricultural produce.
It generates a real-time heat map of locusts across Africa, displays all potential migration routes and offers a locust breeding index. This breeding index shows the magnitude and population density of where the locusts are mating and reproducing. It predicts migration and breeding routes across countries spanning the Horn of Africa and East Africa by using satellite data, soil sensor data, ground meteorological observation and machine learning. The data is used as training samples for the tool and thus proper machine learning takes place. More the data, more will be the accuracy in clarity of the map.
Kuzi is a real development in the fight against Locusts. Using satellite data, soil sensor data, ground meteorological observation, and machine learning, Kuzi can predict the breeding, occurrence and migration routes of desert locusts across African countries, and uses deep learning to identify the formation of locust swarms. Kuzi then sends farmers and pastoralists free SMS alerts 2-3 months in advance of when locusts are highly likely to attack farms and livestock in their areas.
The tool is currently available to users in Somali, Ethiopia, Kenya, and Uganda without any cost and plans are being made to roll out the tool to cover the rest of Africa. If same approach is prepared in a country like India and Pakistan where there is many opportunities to collect data, we surely can protect our crops for future use.
