Drones, high resolution cameras and novel algorithms are paving new ways for the agricultural industry worldwide
Geospatial technologies have enabled massive imprments in agriculture sector in the last decade, and satellite imagery is playing a major role in this. However, satellites do not see through clouds, hence preventing their use when the clouds are persistent. Moreover, while orbiting around the earth they do not pass through the fields at the exact time when the farmer needs it the most. Also, satellites and their imagery are owned by governments and large corporations whose interests are not always aligned with farmers and require massive funding to increase the availability and coverage.
The big three advances
Three recent advances are giving birth to a new way to produce meaningful maps for farming applications. The first one is the development of fully autonomous, small, unmanned aerial systems (sUAS), spurred on primarily by the miniaturisation of autopilot components and the huge interest by consumers, enthusiasts and professionals for drones.
The second is the advancement in consumer grade cameras, providing extremely high resolution images in a lightweight package. The cameras are able to capture the near infrared wavelength for producing NDVI.
The third is the development of novel algorithms for automatically creating digital surface models from collections of overlapping images acquired with consumer-grade cameras. This combination of technologies makes it possible to do local-area 3D mapping and produce NDVI with a total non-recurring investment under $15,000. Repeat and timely acquisitions are enabled with a simple battery recharge.
The application of these novel drone mapping technologies started two years ago in the surveying world, led primarily by the needs of the mining industry. The main advantages were the safety that remote sensing allows, together with an incredible boost in efficiency, as 3D surveying projects that used to take days with ground methodologies would be completed within hours, and often with increased accuracy. Nowadays, all major mining groups are using these type of lightweight drones, either already in production or in research department for future larger scale deployment.
Democratising precision agriculture
One key element of this type of technology is its simplicity. There is no specialised training required for operating the sUAS or processing the software. The software package, such as Pix4Dmapper, takes advantage of the high redundancy of content in simple images to derive extremely accurate and relevant results. Most of the sUAS used in production weigh only a couple of hundreds of grams, and thus are inherently safe. On the legislation side, most governments have started putting restrictions on the usage of drones due to privacy and security reasons. These concerns are very relevant but they do not apply in the mining or farming industry where the drones are not flying over the general crowd and the land is a private property belonging to the farmer or the mining industry. Most regulation offices are aware about this matter and are putting in place exceptions for specific professional usages, such as mining and farming.
Simplicity is a key enabler in the farming industry. It allows farmers to keep and own the data they produce. They depend less on governments or large industries for satellite imagery. Current usage allows them to produce NDVI maps of the field within few minutes. Combined together with ground scouting, the agronomist can feed these maps into precision agriculture management software and produce an application map. These application maps are then transmitted to tractors that will apply the correct amount of nitrogen over the field. There have been numerous success stories in the last year using these simple tools that enhance both the efficiency and the accuracy of traditional scouting. In addition to NDVI, these devices produce detailed 3D elevation model which are used for irrigation planning.
Project with RGB and infrared imagery processed together in Pix4Dmapper
Drones are becoming the products of everyday use, easier to use and more affordable, thanks to the mass production of consumer drones. Last year, more than 500,000 drones were sold, led by companies like Parrot and DJI. Cameras and sensors are getting more reliable and precise. Multi- and hyper-spectral cameras are also easily available for ultra light sUAS. And the software, the key element that allows transformation of these highly repetitive simple measurements into meaningful maps, is leading these tools into the agricultural market by producing relevant data for farmers.
In near future, data delivered by a drone will be more relevant as it will generate immediate health status, treatment recommendation and yield estimates. It has the potential to become a critical part of the farming process. Moreover, these tools will be controlled by the farmers and agronomist directly, giving a true meaning to the term democratisation of precision agriculture.