Unsupported Browser

Your web browser appears to be outdated. Our website may not look quite right in it.

Please consider updating your browser to enjoy an optimal experience.

Dismiss this message

Blog Image

Smart Farming - Detect field zones

2610

Many or all of the products featured here can be from partners who compensate us. This may influence which products we write about and where and how the product appears on a page. However, this does not influencer our evaluations. Our opinions are our own.


Precision in output

Detect Field Zones from bird's-eye view in smart farming

In smart farming, it's not easy to detect field zones. Historical data, information on yield, plant emergence, and available soil nutrients are used to derive specifications for the site-specific application of seed, fertilizer, crop protection products, and for irrigation. The application rates per unit are displayed in so-called application maps and can be transferred to the tractor.

The various application maps represent so-called zones within the field, with the aim of enabling conformal soil cultivation and sowing or demand-oriented plant treatment and optimized harvesting. Criteria can be stored for each zone, which determine how much fertilizer or seed is to be applied within the zones, how deep the seed is placed or how much is harvested at different points. The control of the machines to regulate the application rates is ideally done via ISOBUS.

Meanwhile, there are more and more services that provide application cards, but often the acquisition of the capable machines is relatively expensive and the technical solutions are demanding (e.g. ISOBUS-capable machines).

The base of the application cards can be of different origin:

Satellite data show differences of the plot

Detect Field Zones via satellites from space in smart farming

ESA - Source: esa.int

ESA's Sentinel-2 is one of four Earth observation satellites recently launched into orbit. Its equipment includes a high-resolution multispectral camera. The data is available to the agricultural industry free of charge.

Remote sensing in agriculture

Die Techniken des Precision Farming d.h. der datengestützten Präzisionslandwirtschaft, gewinnen seit einigen Jahren immer mehr an Bedeutung. Insbesondere für die teilflächenspezifische Bewirtschaftungsmethoden auf Basis der Nutzung von Applikationskarten für eine dem Bedarf der Pflanzen angepasste Ausbringung von Betriebsmitteln bei gleichzeitiger Vermeidung von Überdüngung.

With the high-resolution geo-remote sensing provided by ESA's Sentinel satellites, extensive agricultural information (cleaned raw data) is available continuously and free of charge. Companies are using this information to develop innovative services for farmers. New satellite technologies and evaluation methods can provide information on the type and intensity of use, the condition and properties of soils over a wide area and with high spatial resolution.

Globe View

The Earth is regularly "scanned" in narrow strips - Source: esa.int

By 2020, ESA will have ten satellites for Earth and environmental observations, the so-called "Sentinels", in operation. Within this satellite fleet, the four satellites of particular interest to agriculture are: Sentinel 1A /1B (radar) and Sentinel 2A/2B (optical) of great interest. These two pairs of satellites cross Switzerland with a repetition rate of about 5 days. One orbit duration is about 90 minutes. The spatial resolution is 10x10 meters, i.e. the pixel size and thus recognizable differences in the field are limited to this square size. In order to obtain usable data, at least for the optical satellites an almost cloud-free sky is required. This results in the following benefits for agriculture:

  • Demand-based management (fertilization, plant protection, irrigation)
  • Harvest period and biomass estimation
  • Damage estimation e.g. crop damage due to severe weather (storm, flood)
  • Damage due to frost or drought

Since an existing aerial photograph does not have to be unambiguous (water deficiency or N deficiency), the statement regarding the cause of a target deviation is increased by including further information such as the development stage of the plants, soil maps, measures that have already been taken, weather patterns, plant density, etc.

This means that satellite images alone show differences, but do not provide explanations for the causes of the differences. The images of the Sentinel satellites are freely accessible via ESA's sentinel-hub.

However, they are in raw data and still need to be processed. There are several providers and service providers and below are some of these providers that offer services to use crop remote sensing.

Drones provides images from the air

The use of drones in agriculture is already common practice today. It is possible to create application maps for site-specific management. With a drone, inventories are made from the air, so that problem areas, defects or damage can be detected earlier. Compared to satellite images, it has the advantage that daily updated application maps can be created and made independent of cloud cover.

Using drones to detect Field Zones in smart farming

Drones in the air - Source: ufarevue.ch

Current biomass and application maps can be generated for the areas of application, for example N fertilization, growth regulators or plant protection measures. Meanwhile, maps can be generated directly in the field and thus be made available at short notice.

There are complete providers for drone overflights or it can also be generated independently with a good quality commercial drone recordings. However, the use of an application software is recommended, which plans the flight paths of the corresponding field.

When using drones, the applicable rules and legal requirements regarding insurance; flight bans and flight permits must also be observed.

Drone software for evaluation

Bringing machine data into a graphic

Yield mapping

Yield maps show how much was harvested at which point in the field. This information is intended to provide the basis for targeted cause research and a decision-making basis for future management measures.

The crop yield in the field is usually not the same in all places. Causes for yield differences within a field are on the one hand natural factors such as the soil and on the other hand consequences of management measures such as soil compaction by increasingly heavy machinery, nutrient deficiencies or drainage problems. Yield mapping reveals these often astonishingly large yield differences within the areas and presents them in maps. The yield map shows which zones have produced which yield and how much the areas within a field differ. Especially the evaluation of yields over several years helps to identify zones with different yield potentials.

Corresponding sensors in the discharge pipe system of the machines determine quantities and moisture in a geo-referenced manner.

Using data to detect field zones in smart farming

The use of these cards also has its limitations:

  • The difference in individual years due to weather and natural forces is often the greatest influence
  • As a result, yield patterns partially cancel each other out
  • due to many factors influencing the yield, it is difficult to refer to the current year from past values

Therefore, it is recommended to assess the current plant emergence for crop management measures.

Optical plant sensor

This method uses multispectral cameras that detect the green component (chlorophyll) of the plant by reflecting the light emitted by the plants in their sights. From this, the vegetation index IRMI is calculated, which reflects the current nitrogen supply of the plants. The system measures the amount of N that the plant has already taken up to that point, compares it to the target value of the current EC stage, and thus determines the nutrient difference that is missing or needs to be compensated. The possible areas of application are:

Fertilization

  • Balance biomass (promote weak areas stock)
  • Quality biomass (strong areas promote stock)

Plant protection

  • Growth regulator (treat weak stand less)
  • Fungicide (treat dense stand more, treat thin stand less)

On this basis, the ISOBUS-capable fertilizer spreader or crop protection sprayer is controlled and applies the required quantity in line with demand.

Power consumption visible on card

The recorded power of the tractor during field work is georeferenced and displayed on a map. The information shows the different power applications within a plot, which can be used for further decisions. Thus, the following applications can be adapted to a current situation of the soil accordingly.

Include geological and soil maps

Georeferenced soil sampling

To ensure that sampling is representative and takes place within the boundaries of the subplot, it is necessary to define the sampling lines prior to sampling. This also ensures that the repetition of sampling, e.g. in the rotation of a crop rotation, can take place along the same sampling line or sampling points and that comparability is given.

Soil conductivity

Different methods and procedures allow the measurement of the conductivity of the soil. This is to make a conclusion on compaction and water saturation of the soil.

EM 38 Conductivity measurement

The EM-38 measurement is an electromagnetic induction method, which is used, among other things, to determine and map zones with the same soil properties, especially in (highly) heterogeneous soils. Furthermore, it serves as a basis for the formation of the zones within the plots.

Top Soil Mapper

With the Topsoil Mapper, it should be possible to record inhomogeneity in the soil over a large area and without contact and to map soil parameters such as soil type, water saturation and compaction on the basis of electromagnetic induction.

Topsoil Mapper to detect field zones in smart farming

Soil map

Knowledge of soil properties is an important prerequisite for site-adapted and careful management. Soil maps provide information on the properties and spatial distribution of soils. The soil maps provide information about the soil type, water balance, plant-useful depth and grain size.

Create cards yourself

QGIS

QGIS is a geoinformation system for viewing, editing and capturing spatial data and is available as open source free of charge.

The generated application maps are provided as shapefiles (.shp) and can be used in different ways. If the tractor has an ISOBUS terminal with "Task Controller geo-based" (TC-GEO) functionality, for example, a fertilizer map can be transferred straight to the terminal, ready for immediate use. In the event that no such terminal is available, an Android application for smartphones and tablets is being developed.

paint.t3rra

T3RRA design is mainly for leveling terrain. It works together with John Deere applications. Click here to visit the website.

Creation of an application map

The soils in our area vary greatly in terms of their green cover and root zone, even within the same field. Thus, the nutrient content in the soil and the resulting yield potential varies within a few steps. This is where site-specific management comes into play.

Various technologies from hyper- and multispectral images provide evaluations that provide conclusions about the potential for crop yield. From this, a zoning of the plots takes place. Supplemented by further information from soil analyses and the knowledge of the farmer, management strategies can be set up and application maps created accordingly.

Creation of an application map to detect field zones in smart farming

Special knowledge around application cards

Spectral Technology

Multi- and hyperspectral technology

The sensors work with sunlight reflected back from the earth's surface. It is a passive recording method of remote sensing. Spectral digital recording systems record the sunlight reflected from the earth separated by wavelength in so-called recording channels. This means that the imaging sensors each have an imaging channel for different ranges of light. Multispectral remote sensing sensors have one channel each for the red, green and blue and then up to 10 additional spectral channels. Hyperspectral cameras operate in ranges between 10 and several hundred spectral channels, which explains the high price of these devices.

Satellite images are converted into a raster, which are arranged in columns and rows. The individual cells in such a raster become the pixels.  The values reflect the intensity of the light reflected from the earth. In another process certain colors are assigned to the different wavelength ranges (= channels), which leads to the corresponding color images.

Using Vegetation indices to detect field zones in smart farming

Vegetation indices

NDVI

The vegetation index is a value which is determined on the basis of the measurement of the reflection of the light of the earth's surface or its vegetation towards the sensor (e.g. satellite). With the help of a vegetation index it is possible to quickly identify areas on the earth's surface covered with vegetation and their condition. Now it is a question of recognizing the condition of a plant by differences in the back radiation (reflection).

Using Vegetation indices to detect field zones in smart farming

NDVI stands for "normalized difference vegetation index" (also "normalized density vegetation index"). It is currently the most frequently used vegetation index and is calculated on the basis of satellite image data, among other things.

Using Vegetation indices to detect field zones in smart farming

Healthy plants reflect relatively little radiation in the visible spectral range (RGB), but relatively much in the infrared range (IR). This means that the healthier a plant is, the higher the reflection in the near infrared range (NIR). These two spectral ranges are put in relation to each other. Soil or even matured plants, on the other hand, do not show a clear difference in the reflectance of both ranges. Measurements of these spectral ranges using sensors from meteorological satellites therefore make it possible to distinguish areas on the Earth's surface that are covered with vegetation from uncovered areas. They also allow conclusions to be drawn about photosynthetic activity, vitality and the density of the vegetation cover.

GNDVI

GNDVI = green normalized difference vegetation index. It is similar to NDVI except that instead of the red spectrum, the green spectrum is measured in the range of 0.54 to 0.57 microns. This is an indicator of photosynthetic activity of vegetation cover. It is most commonly used to assess moisture content and nitrogen concentration in plant leaves using multispectral data that do not have an extreme red channel. Compared to the NDVI index, it is more sensitive to chlorophyll concentration. It is used to assess depressed and aged vegetation.

ENDVI

ENDVI includes a comparison of green and blue light in addition to NIR and red for a more sensitive result. This isolates the indicators of plant health and can be used to assess the presence and health of a crop.

EVI

EVI = enhanced vegetation index. Designed to improve NDVI by optimizing the vegetation signal in areas with a high leaf area index (LAI). The index uses the blue reflectance region to correct for background signals in the ground and reduce atmospheric effects, including aerosol scattering. Most useful in regions with high LAI where NDVI may be saturated. EVI values should be between 0 and 1 for vegetation pixels. It is used to assess variability in plant development under both dense vegetation cover and sparse vegetation.

Using enhanced vegetation index to detect field zones in smart farming

Isobus

Isobus was explained in detail under the following article: FoodPorty Blog - Data Transmission

References


You may also like