In the Field
UAV surveys will need big data management tools
April 27, 2015 By Raima Inc.
The vast North American prairies have been producing grain crops in global volumes for many decades. They helped feed America and Canada as they grew, and generate large amounts of foreign revenues from the export markets.
From the 1960s onward, aerial and even satellite photography of the fields have helped drive efficiency into planting, irrigation and harvesting. However, such surveys were difficult and expensive to execute and the results were often crude.
Not surprisingly then, the advent of the modern UAV has moved agricultural surveying onto an entirely new plane. Today’s UAVs are small, simple and cheap to operate, and can carry multiple types of surveying equipment. They allow survey data to be collected in a continuous stream throughout the flight and instantly uploaded to a server for immediate analysis. Where previously aerial surveys were occasional and crude, now they can be frequent and detailed!
Farmers are now talking about moving into an era of ‘precision agriculture’, which luckily ties in with economists’ predictions that worldwide population growth over the next two generations will mean that food production will have to at least double. UAV surveys will play an important role in ensuring the necessary increase in productivity.
As well as conventional cameras, the survey UAVs can carry specialist equipment such as ultraviolet, infrared and thermal imaging cameras, hyperspectral sensors and lidar scanners. (Hyperspectral sensors provide ‘big picture’ or wide-view information to typically 10mm resolution. Lidar is a cross between lasers and radar which automatically focuses on planted areas and scans them with up to 500,000 laser pulses per second to collect ultra-high resolution data.) These data streams will have to be accurately married with GPS (global positioning systems) and other information for complete analysis.
It is inevitable that farmers will want their surveys to provide instant information, so the survey data will have to upload instantly to the cloud, be processed, then downloaded – to both mobile devices in the field and to a central computer in the farm office.
Dr Kevin Price of Kansas State University predicts that in the medium-term future (5–10 years) over three-quarters of UAV usage will be agriculture-driven, and that this will grow into a $100 bn industry. UAVs will be used year-round to monitor crop health, growth rates, yields, the presence of invasive species, livestock and their feed bins, water resources, etc. Daily surveys and 24×7 surveys will not be unusual, and by comparing agricultural information to market forecasts, farmers will be able to determine the optimum time to harvest. In order to provide this outcome, vast amounts of data will have to be handled automatically, so high-performance, always-on database management systems will become as important to agriculture as the tractor or combine harvester.
Many other industries are also adopting UAVs for security surveillance, battlefield monitoring, filming, crowd control, traffic management, assessing the condition of buildings, metrological forecasting, geological surveys, flood management, etc. The big data solutions developed for agriculture will transfer directly into these fields.
Much of the hardware necessary for this sort of blanket surveying is already available at near-commodity prices. But for precision agriculture and other professional uses, data quality and post-processing are critical, and this is a specialist role for dedicated data services providers.
With its RDM (Raima Database Manager) embedded data management technology, Raima already has a proven track record in big data applications across many industries. Its solutions offer specialised capabilities that enable intuitive management and manipulation of real-time information.
RDM is optimised for embedded, real-time, in memory and mobile applications, delivering flexible and reliable solutions for collecting and storing large volumes of data. It provides intuitive and efficient methods for managing and navigating through information quickly.
The product has a highly modular structure, so that only the components that you really need are included in your embedded application. As well as multiple APIs, RDM includes developer tools to help you manage databases on an embedded platform, and these tools can also be built into your application if desired.
Designed for distributed architectures in resource-constrained applications, RDM can store data locally on a mobile or embedded device, which may be disconnected from any networked database servers, and then replicate this data to a server when a connection becomes available. This capability also allows RDM to buffer data on a mobile device during a phase of rapid data acquisition, and then transmit it across a network with relatively low bandwidth. This feature can be useful in agricultural applications, where low-cost processors and low-bandwidth networks are common.
Because RDM provides sophisticated data management on the embedded device itself, it is also possible to filter the data in real time on the embedded device, thereby reducing the amount of data that has to be transmitted to remote servers.
RDM provides a SQL implementation that is small enough for resource-constrained systems, while providing the features most needed by an embedded application, such as pre-compiled database schemas, pre-compiled stored procedures, user-defined functions written in C and “virtual tables” for accessing any kind of source data through SQL (e.g. real-time data fed from sensors).
RDM also has lower-level, non-SQL APIs for applications where SQL would be too big an overhead. On the other hand, if you do use SQL, the same database can be accessed through SQL and non-SQL APIs.
Importantly, RDM’s modular structure, and its support for distributed data and scalability through data partitioning, make it a very flexible product. As well as enabling it to run on small systems these characteristics allow it to handle larger databases, supporting data mining applications.
Data mining allows users, such as farmers, to look beyond simple factual information and consider the consequences of each decision in a wider context. For instance, data mining can highlight implicit, previously unknown and potentially useful knowledge from large data sets. A real-world agricultural example of this would be that, if localised adverse weather occurs, data mining will identify similar weather patterns from previous years and predict how market prices will change.
There is no doubt that in future, data – and the information and knowledge that can be drawn from that data – will be one of the greatest assets available to farmers. Raima’s RDM data management technology offers agriculture the tools needed to achieve greater productivity and to be better able to react to changing market requirements.
Print this page