A major contributor to crop loss is plant disease. While many farmers can identify some of the plant diseases, they lack access to identify treatment to save the crop in time. PEAT – Progressive Environmental and Agricultural Technologies, a German startup has developed a mobile app called Plantix which is a massive database of pictures of plant disease which can be used for comparison. This helps in identification and subsequent diagnosis and treatment. PEAT aims to support farmers across the world, specifically developing countries to use technology to enhance their agricultural output through timely and informed disease treatment.
PEAT’s Charlotte Schumann spoke to Networked India about how the Plantix app brings together technology like geodata, image recognition, artificial intelligence, agroecology and algorithms and presents these in a cool package via an app for farmers. She also tells about PEAT’s collaboration with Hyderabad-based ICRISAT to assist farmers in Telangana and Andhra Pradesh through an extended network.
Networked India (NI): Diagnosing plant disease through an app – sounds very futuristic! How does it work?
Charlotte Schumann (CS): We have developed a technology that is able to identify over 50 plant diseases, pest & nutrient deficiencies just with the help of a photograph of the infected plant. In principle, this software can be implemented in any system with a camera and – at least temporary – access to the internet. In a first step, we implemented this technology into a smartphone app called Plantix. Our software has literally learned how to identify different plant damages by identifying the specific optical patterns of that disease, pests and nutrient deficiencies left on the plants – just like an human expert. We teach our software with the help of learning datasets consisting of several hundred pictures of every specific plant damage. Thanks to our users, our database is growing from day to day, and every week there are new plant damages that can be identified via Plantix.
Plantix app screens
NI: The Plantix app uses some very cool technology like Geodata, digital library and other nifty tools. Tell us some more about them.
CS: PEAT believes that Artificial Intelligence will boost human capacities to a new level. We take cutting edge technologies from the field of image recognition and deep learning and bring it to a field where we think smart technologies can really make a difference for everyone. Agriculture can grow smart and sustainable when supported by real time and tailor made information. Our app Plantix does not only help to identify plant damages, it also supports farmers to treat them adequately. We are giving advice for biological and chemical treatment as well as preventive measures to prevent the occurrence of plant diseases or at least contain them at an early stage.
Beyond identifying damages, we want to push forward regionalised early warning systems related to upcoming plant diseases. This will also be possible with the help of PEAT’s software. With every picture sent to our servers, we receive geodata on where and when the picture was taken. In this way we can map the spread of plant diseases and send out warning messages if we know that a pathogen is around. But also we can get deeper insights on how geofactors like soil types and weather conditions are related to the occurrence and spread of plant diseases. Based on such information we can provide our users with really customised advice for the prevention of plant damage and the reduction of pesticide use. One of our focuses is on integrated pest management.
Charlotte collecting data on the banana crop
NI: How did PEAT come into existence? What is PEAT’s goal?
CS: Plantix is the first product of PEAT, a young German StartUp founded 2015 in Hanover, Germany. PEAT’s founders shared working experience before in different contexts, for example by working together in a German-Brazilian research project in Amazonia, dealing with greenhouse gases (GHG), sustainability and land use change. It was these kinds of field experiences that triggered the idea of tackling the problem of plant disease recognition in remote areas with the help of digital technologies. The vast majority of Brazilian small scale farmers had a local name for the pathogens but lacked the spot knowledge on treatment options and background information on how to prevent infections. Large scale farmers on the other hand did not have the time and resources to be on the field and watch their vast plantations. Both groups can be helped with automated disease recognition. Plantix’s goal is to support farmers worldwide in an early and informed disease treatment, which will lead to an optimisation of resources and chemical use as well as reduce severe crop shortfalls.
NI: Who are the brains behind PEAT – introduce us to the founding team?
CS: PEAT stands for Progressive Environmental and Agricultural Technologies, and that’s already the company’s main orientation. The team combines expertise in the area of agroecology, IT and artificial intelligence, geodata management as well as a profound experience in fieldwork with farmers and development cooperation. Coming from social as well as natural science, the strength of PEAT’s founders is based on their diversity of perspectives on agriculture, based on a shared striving to build a world worth living in for all of us. This is our team:
Simone Strey – CEO. Simone is a geographer specialized in geobotany and soil science – and our CEO.
Alexander Kennepohl – Plant Pathology & Data. Alex is our expert for plant diseases and databases. His expertise in geographic information systems (GIS) helps PEAT to develop our geostatistics and maps.
Charlotte Schumann – Knowledge Management. Charlotte is responsible for international cooperations and knowledge management in PEAT. She recently finished her PhD in Social Anthropology and currently secures the quality of PEATs overall claim to engage in an empowering dialogue with farmers worldwide.
Robert Strey – Programming. Rob sits at the interface between ecology and machine learning within PEAT. He prepares the food for thought of our beloved algorithms and finally pushes the button to train them.
Pierre Munzel – PR/Marketing. He is socially medial, always up to date and loves sharing: with style and passion. In PEAT, he is responsible for public relations, administration and legal issues.
Korbinian Hartberger – Communication. Korbinian’s area in PEAT is communication – he is our language expert.
Bianca Kummer – International Cooperations. Bianca studied engineering for landscape planning and recently graduated from the renowned Center for Advanced Training in Rural Development (SLE). In PEAT, she coordinates local cooperations of PEAT in different African countries.
NI: Tell us about PEAT’s India connection and Icrisat, the Hyderabad-based international institute for research in crops for semi-arid tropics.
CS: India is an incredible important field if you want to innovate in the field of agriculture. Here, over 700 million people make a living on agriculture and at the same time, India’s proneness to technology is far more than a myth. Smartphone penetration in rural areas is far more developed than in comparable countries, and network coverage is constantly being expanded.
So, India was the place to be. In a first step, we looked for partners to adapt our database to Indian crops and diseases, which differ from other countries where PEAT’s app Plantix is rolled out already. In ICRISAT, we found a strong partner with an impressive expertise on crops planted in the semi arid tropics. In collaboration with ICRISAT then, a database is built up related to mainly pulses. In addition to the scientific support, plans are made to use PEAT’s technology within ICRISAT’s extension network that assists farmers in Telangana and Andhra Pradesh. Extension officers supported by ICRISAT can help us to adapt this technology right to the needs of local farmers. We are looking forward to extending this collaboration in more countries and assist even more farmers with real time, tailor made information on their crops and diseases.
NI: Given many Indian farmers are poor and barely manage to access basic mobile services, how do you envisage them utilising PEAT’s tech-enabled services?
CS: In a first step, we will work with public and private extension services in order to adapt Plantix’s usability and the information delivered. In that field, smartphone usage is pretty common in India. But we also believe it is just a matter of time until the penetration of smartphones will also increase in the rural areas and we want to provide intelligent options for usage beyond simply communication – smartphones can be very mighty tools if they are equipped with the right information tools and a lot of new apps for farmers are pushing on the market. They are ranging from weather apps over sowing to market price apps. We would like to get connected to other publishers of apps to connect all this information and create a more integrated approach, e.g. also giving information on where to obtain products for treatment.
NI: PEAT has a massive database of pictures of plant diseases which can help farmers detect what ails their crops. How did PEAT build up this database? Any specific collaborations or crowd-sourcing efforts?
CS: PEAT’s database consists of over 90,000 pictures today. A lot of them were collected through crowd-sourcing science, by people looking for answers to their problem. Additionally, we gather pictures with our private and public partners so that we are extending our database day by day. This way, the number of damages recognized by our software grows rapidly.
Alex collecting information on crops
When we started one and a half year ago, our software could recognize five different nutrient deficiencies on tomato. This was our proof of principle but we wanted to extend our database on other crops and diseases. So we built up Plantix, our picture-based mobile plant disease guide. Plantix today is a – very successful – data mining tool. In the first growing season in Germany, as a bootstrapped project, the app was downloaded over 25,000 times. Today we have more than 30,000 sessions a month in countries like Germany, India and Brazil. We aim to increase this number to gather more data and to extend our database on new diseases so that every farmer in India and all over the world can use our app to identify what’s wrong with his or her plant.
NI: What does future like for PEAT? Any India-specific goals?
CS: PEAT’s vision is to define a new standard for the automated recognition of plant diseases worldwide. After Germany and some other European countries, the two next big roll outs for Plantix are set in Brazil and India, and we provide interfaces to our software to providers of rural apps in Tunisia and Kenya as well.
In India, Plantix will be adapted to regional needs starting with a pilot in Telangana and Andhra Pradesh, important areas for the production of rice, maize, cotton etc. In collaboration with private and public partners, we will built up a content database in Telugu and train our recognition software on major diseases occurring in these two states. After this, next states will follow, first connections have been established in Maharashtra for example.