Using a drone alone does not solve the problem, advanced techniques are needed to analyze aerial images

​With the advent of Unmanned Aerial Vehicles (UAVs), the challenge of developing computational programs for the acquisition, processing and analysis of captured images came to the fore.

04.10.2016 | 20:59 (UTC -3)

With the advent of Unmanned Aerial Vehicles (UAVs) - also known as drones - the challenge of developing computational programs for the acquisition, treatment and analysis of images captured by on-board cameras came to the fore, because the devices alone do not identify pests and diseases, as many believe.

"Powerful systems are needed to recognize patterns", says the researcher at Embrapa Instrumentação (São Carlos – SP), Lúcio André de Castro Jorge, who uses advanced geoprocessing techniques to interpret the images, transforming them into useful data for the producer.

The popularization of the drone has opened up a new range of opportunities in various fields. For the rural producer, who inspects the crops by walking through the plots, the equipment is an ally. However, according to the researcher, it is necessary to pay attention and not buy into the idea that the device solves everything and that any drone is good.

In so-called precision agriculture, a system that takes into account the spatial management of crops, the UAV has been used with the help of fundamental on-board accessories, such as high-resolution multispectral and hyperspectral photographic cameras, with infrared sensors, which allow pests to be identified. and diseases early.

The subject is on the agenda of discussions at the Brazilian Congress of Precision Agriculture (ConBAP), where Castro Jorge will outline an overview of aerial platforms for sensors – UAVs and drones – myths, truths and perspectives that still surround the agricultural use of these devices.

Held from October 4th to 6th, at the Federal University of Goiás, in Goiânia, the event is a biannual discussion forum that brings together scholars, researchers, extensionists, suppliers and users of the different techniques involved in the wide range of Precision Agriculture ( AP).

The researcher will also mediate a sector panel on images and remote sensing, while the coordinator of Embrapa's Precision Agriculture Network, Ricardo Inamasu, who makes up the Conbap scientific committee, will be responsible for another on "Data - who generates, who uses and who they belong to."

"We invested in creating image processing methodologies for low-cost drones capable of operating in adverse field conditions", says Castro Jorge, who has been developing software research for two decades for use, in principle, on images captured by any aircraft.

For him, the images taken by UAVs, combined with a good geoprocessing technique, accurately identify the existence of pests and flaws in the crop, soil problems, areas affected by erosion and silting of rivers, areas attacked by nematodes, water deficiency, because a computer program indicates with specific colors these problems that cause damage to properties.

Analysis techniques

The techniques developed by the researcher to process images are being used in sugarcane, citrus, corn and cotton crops in different regions of the country. In Gavião Peixoto (SP), the electronic engineer is conducting studies with UAVs to diagnose a disease that is considered the most important, severe, serious, destructive and devastating in the world's citrus industry – HLB (huanglongbing), known as greening.

In the State of São Paulo, between 2005 and 2012, the pest was the cause of the elimination of 26,7 million plants according to data released by the Agricultural Defense Coordination (CDA), based on reports of inspection and eradication of plants that producers delivered semi-annually.

Customized software is being used by an American multinational in the field of agriculture and biotechnology to evaluate corn hybrids, while one of the largest Brazilian cotton producing groups is studying the use of technology, on an experimental basis, to analyze the influence of nitrogen on the development of cotton or nutritional deficiency of the plant. More than two thousand photos have already been captured by drones on the group's property in Goiânia.

The photos will be processed with the help of Siscob software – used to analyze land cover – and others that are being developed, specifically for evaluating cotton crops.

Once obtained, the images can be processed individually with the aim of recognizing serious planting flaws that still allow correction. An example is gaps in the sugarcane planting line greater than 50 cm. With the "fault counter" software, which is in the validation phase with producers, it will be possible to interpret occurrences in the field.

Proper use

To make good use of the UAV, Castro Jorge's tips are to complete some steps, such as flight planning, overlapping flights, obtaining georeferenced images, image processing, mosaic generation, analysis in a GIS tool – a geographic information system - until the generation of reports with interpretations of the images captured by the device.

"With this information in hand, the producer will be able to prepare maps of recommendations, such as decompression, fertility and application of inputs at variable rates, helping in decision making", evaluates the researcher.

Some tools are available free of charge to farmers on the pagehttp://labimagem.cnpdia.embrapa.br/, such as the system for analyzing soil cover – Siscob – and the geolocated collection system – GeoFielder.

Fx Market

Unmanned aerial vehicles are the highlights of the moment. More than 40 countries have been working on the development of these devices for different markets, with billions of dollars. Study of The Global UAV Payload Marker 2012-2022, carried out by the company RnR Market Research, released at the end of 2012, points out that the global UAV market is worth 43,7 billion dollars, but the forecast is that the sector will move around of 68,6 billion in 2022.

The growth comes precisely in the wake of recent advances in computer technology, software development, lighter materials, global navigation systems, advanced data links, sophisticated sensors and miniaturization.

Pioneering in agriculture

In Brazil, the first reports of UAVs occurred in the 80s, when the Aerospace Technological Center (CTA) developed the Acauã project, specifically for military purposes. In the civil area, also in the 80s, the Helix project stands out, a mobile wing van that was deactivated in the following years. Later, the Renato Archer Information Technology Center (Cenpra) developed the Aurora airship project, which served to train the development team.

In agriculture, Embrapa Instrumentação is a pioneer in the use of unmanned aerial vehicles, having started research into the use of these devices in 1998. The proposal was to replace conventional aircraft, used to obtain aerial photographs, for monitoring agricultural areas and subject areas. to environmental problems, by small vans that carried out missions pre-established by users.

Since then, Embrapa has invested in the development of other platforms and aircraft that are capable of operating in adverse field conditions, such as agricultural areas, but with good performance and low risk.

The research is focused on two fronts, on the development of software and image capture systems suitable for different agricultural applications, for different scales, from small to large properties. Tests are being carried out on several aircraft models, such as multirotor systems, so-called multicopters, all embedded with free software, which are a cheaper technological option for small areas.

Some tests are carried out in the area of ​​the National Reference Laboratory for Precision Agriculture, in São Carlos, opened in 2013, as an unprecedented space to research and develop equipment, sensors, mechanical components and on-board electronics, in the same location.

Cultivar Newsletter

Receive the latest agriculture news by email

access whatsapp group
Agritechnica 2025