The Food and Agriculture Organization (FAO) of the United Nations projects that by 2050 humanity’s ranks will likely have grown to nearly 10 billion people. Farmers will need to produce more with less, while preserving our environment for future generations. And society has a duty to help them achieve this. Although agriculture is perceived as a traditional economic sector, precision agriculture technologies have already boosted crop yields significantly in the last decades. What else can be done?
What drones have to offer to agriculture? You might argue: Unmanned Aerial Vehicles (UAVs) have been around for about a century, so what’s new?
Robust investments in the commercial drone sector have made the technology cheaper, lighter, safer and more sophisticated. Drones can fly on autopilot. The unique combination of the above-mentioned factors makes the utilization of drones by smallholder farmers affordable.
Early pest detection is a major application of drones in agriculture. Depending on the crop, agricultural producers survey their land several times per week. For example, a potato producer is going in the field at least 3 to 4 times per week during the growing stage of the crop. No surprise here, as pests like the Colorado potato beetle, could spread extremely fast and destroy hundreds of hectares per day.
The earlier you catch the problem, the cheaper to contain it. Once you identify a small spot, you can fix it right away, as opposed to having it spread. This leads to big time and labor savings.
AgroHelper, a Bulgarian startup is working on developing a web-based solution that helps farmers process drone captured images and detect, in real time, zones with potential crop health issues. The platform is powered by a cloud infrastructure and does not require any specific hardware to be present on the farmer’s local machine (or a fast internet connection).
Currently, most state-of-the-art software solutions, use a process called “stitching” to create an orthophoto map from hundreds of individual overlapping aerial photos. Each individual photo captured by the drone camera contains different terrain features like crop rows, tractor trails or buildings. As the photos overlap, each individual feature is captured by the drone camera multiple times from different angles and perspectives.
AgroHelper’s Health Map feature indicates zones with potential issues
Stitching, as the name suggest, is a mathematical process that matches the photos to solve the puzzle and create one high-resolution map. This is the most precise solution available today. However, the process takes from 5 to 10 hours, requires heavy computing power (which translates in high cost) and fails fairly often.
The AgroHelper team aims to solve this problem by offering a real-time tool and eliminate stitching from the process. The results are less precise than stitching in terms of resolution but allow farmers to pinpoint an area that requires attention in real time. As the process is optimized, the cost is contained, and AgroHelper provides farmers the opportunity to process 3 maps per month for free.