Smart farming

Precision agriculture aims to optimize field-level management with regard to

- crop science: by matching farming practices more closely to crop needs (e.g. fertilizer inputs);
- environmental protection: by reducing environmental risks and footprint of farming (e.g. limiting leaching of nitrogen);
- economics: by boosting competitiveness through more efficient practices (e.g. improved management of fertilizer usage and other inputs).

Precision agricultural is not a single technology, but rather a set of many component technologies from which farmers can select to form a system that meets their unique needs and management style.

Soil spatial inventory techniques

Analysing Productivity Patterns in Agricultural Fields

By knowing inner-field variability and applying precision farming techniques,  the amount of necessary pesticides can be reduced to the necessary minimum as well as the over-fertilization of soils can be avoided. Conventional land management, that is focussing on the production of large amounts of high-quality agricultural products, can therefore combine economic benefits with environmental sustainability.  To dertermine variability, we are using earth observation data  such as photos and radar images taken from remote-sensing satellites, aircrafts or drones and we develop decision-support tools based on processed information, such as maps and models. By integrating the results of positioning systems, variation detection and feedback sensors and decision models, we are able to reach different levels of accuracy: field, plot, plant or leaf.

Precision application

Fertilizer spreading

Poor application of fertilizer leads to poor crop performance and environmental damage, such as eutrophication. For the best utilization of fertilizers, the nutrients have to be spread in the right amount at the right spot according to soil characteristics and crop conditions in the field. Spread patterns of centrifugal spreaders can be determined using a combination of measurements via image acquisition and processing , combined with ballistic flight modelling (hybrid approach). Using a high speed binocular stereovision system, static (size, shape) and dynamic (3D velocity) properties are determined for the fertilizer grains leaving the vanes. These measurements are then used as inputs of the ballistic flight model, allowing to determine the spread pattern of the spreader. As such, this method can be used for the fast and accurate calibration of spreaders at farm level or for design of new spreaders. When implemented on the spreader, the system could provide the machine real-time feedback on the distribution of fertilizer in combination with e.g. Variable Rate Technology equipment.

Precision spraying

Precision spraying allows zones or individual spray nozzles to be regulated by a map-based controller, potentially saving chemicals, fuel, and time during the application process. In addition to the private benefits to the adopting farmer, society may also benefit through reduced agrichemical pollution, and ultimately, through reduced cost of food and fiber. For the measurement of spray characteristics, we are using image acquisition systems and image processing algorithms for 2D motion estimation and size determination. These tools allow us to determine macro and micro spray characteristics.

Use of hyperspectral data from remotely piloted aerial systems for weed detection and management

Field aerial hyperspectral images of weed are collected throughout spring and the effective spectral wavelengths are used to discriminate the spectral response of weeds and crops by using Successive Projection Algorithm (SPA). The textural features of crops and weeds are obtained through the Gray-level co-occurrence matrix (GLCM) algorithm. The morphological parameters of species (i.e. area, perimeter, skeleton length, Hu factors and roundness) can be analyzed using Object-based image analysis (OBIA) algorithms and the traditional Otsu's threshold method. The different combined feature sets will be regarded as inputs to different discrimination models e.g. Support Vector Machines (SVM) or Linear Discrimination Analysis (LDA) to classify species. The accuracy of the resulting models will be assessed by overall classification accuracy.

To facilitate precision spraying, relevant task maps are generated that delineate the spray zones combined with the appropriate doses. An economic analysis in terms of costs of plants protection products and labor costs can be added.

monitoring Plants

Sensor monitoring and big data analysis enable growers to make better and faster decisions about the management of their crops. Plant-based control and stress detection systems allow monitoring of multiple plant related variables and provide for accurate tools to assess the real plant state.

Greenhouse gas emission control

Development of measuring techniques for ventilation and ammonia emission rates from naturally ventilated animal houses

Because of the impact of ammonia emissions and consequently nitrogen deposition originating from agriculture, a pressing need exists to develop new and affordable emission reducing techniques. However, measuring the emission rate in naturally ventilated animal houses to quantify the reduction potential of such techniques is not straightforward. This is mainly due to the temporal and spatial variability of both gas concentration and air velocity profiles. This research aims at developing a measuring technique both for velocity profiles in naturally ventilated animal houses and for the accompanying ammonia emissions. To this end, a naturally ventilated test installation where conditions, except those that are weather related, can be controlled, has been built. In this construction, an accurate measuring technique for the airflow rate / velocity profile has been developed, implemented and validated using simultaneous emission measurements. These accurate measuring techniques are now being adapted for transfer and application to commercial animal houses. 

Optimization of air scrubbers to reduce ammonia, odour and greenhouse gases from animal housing facilities.

The livestock sector in Flanders has largely increased over the last decades, which implies an increasing potential impact of the individual farms on their surroundings. Since 2004, newly built animal housing facilities in Flanders are legally required to be emission-low with respect to ammonia, according to one of the approved housing systems (Ministerial Decree of 19 March 2004). The most common application for pig housing facilities are air scrubbers, in which ammonia from the exhaust air is removed by absorption in water, followed by chemical and/or biological conversions. Despite the widespread use of chemical, biological and combined air scrubbing systems, there is still insufficient knowledge on the relation between removal efficiency, energy and water consumption on the one hand and process design and control on the other hand. We deal with these process engineering aspects, in order to exploit the full potential of air scrubbers. Particular attention is paid to the relationships between different types of emissions, not only ammonia but also odour and greenhouse gases (CH4 and N2O). The applied methodology concerns physical-based modelling and simulation, complemented with full-scale monitoring campaigns and lab-scale  experiments.