About LAR Citizen Science (LARCS)
What is Citizen Science?
Citizen science is the practice of public participation and collaboration in scientific research to increase scientific knowledge. Through citizen science, people share and contribute to data monitoring and collection programs. Usually this participation is done as an unpaid volunteer.
Collaboration in citizen science involves scientists and researchers working with the public. Community-based groups may generate ideas and engage with scientists for advice, leadership, and program coordination. Interested volunteers, amateur scientists, students, and educators may network and promote new ideas to advance our understanding of the world.National Geographic
The LARCS project establishes, within the new Lincoln Agri-Robotics (LAR) centre, a citizen science initiative for collecting agricultural and horticultural data on local crops in order to create a database that can be used by LAR researchers, as well as others in the scientific community, to develop sensor-based models of crops. Data (camera images) will be solicited from citizens within the Lincolnshire community (and beyond) to help monitor agri-food growth. The LARCS project will not only engage the professional farming community around Lincolnshire and potentially more broadly across the UK, but also reach out to home gardeners and potentially introduce gardening enthusiasts to aspects of agricultural and horticultural research.
Within the scientific communities around agricultural technologies, including computer science and robotics, there has been a rapid increase over the last 5-10 years in the use of camera imagery to identify weeds, segment weeds from crops, count aspects of crop output (e.g. fruits or seed heads), track growth and estimate water retention. The most common approaches to analysing these data include engineered (rule-based) image processing and machine learning, for which “deep learning” techniques are currently extremely popular. However, in order to apply a machine learning technique successfully, a large and robust data set must be provided to the computer-based process which “learns” a model of the data. Such models are used for the types of counting, identification and prediction tasks outlined above; but a model is typically tailored to the task to which the model will be applied, which means that a substantial amount of data must be available in order to develop such a range of models.
A typical way to collect data for such machine learning research is for scientists to gather the data themselves, frequently in collaboration with industry partners. For example to gather data on a particular crop, such as strawberries, the researchers will collaborate with a commercial berry grower and set up cameras to collect the data at agreed-upon intervals. These cameras may be operated by the scientists or may be operated by trained personnel who work for the commercial grower. However, during the extraordinary circumstances which commenced in the UK with a national lockdown on 23rd March 2020, the opportunities for scientists to collect data on site are minimal at best and largely non-existent. This project allows commercial growers, as well as general citizens in the community, to collect data for us and contribute these data to our repository.