Українська   Русский
DonNTU   Masters’ portal

Abstract

Content

Introduction

Using precision farming technology is becoming more common not only in the developed countries of Western Europe and North America, but also in Asia, South America, Africa. In Ukraine, as necessary to implement and use new technologies, as they determine the level of development of agriculture and its competitiveness on world markets. The purpose of this direction is to maximize profits subject to save fertilizer, water, fuel, land management, environmental protection.

1. Theme urgency

The concept of precision agriculture (hereinafter PA) are representations of the existence of irregularities within the same field. Data heterogeneity can be identified by examining the soil. This information must be spatially linked to a cartographic basis, otherwise defeats the purpose of this definition of location, so necessary geographic information system (hereinafter GIS), which would be stored and processed on information soils mapping basis, plot boundaries, etc. The use of GIS in precision agriculture treated Boyko O. [2], Belavtseva T. [3], Barladin A. and Yaroshchuk P. [4], Bychkov I. [5], etc. In these studies proposed to use information systems with elements of GIS: SSToolBox, Agro‑Map PF, Agromenedzher, UrozhayAgro, MapInfo and AgroView, etc [5]. In the cartographic basis vectorised maps accepted scale 1 : 200000 or satellite images of Terra and Aqua by vectoring field boundaries with subsequent refinement according positioning systems GPS / GLONASS [4, 5].

2. Goal and tasks of the research

Satellite imagery used without processing, eliminating geometric distortions caused by terrain, shooting conditions and the type of camera, which reduces the precision of the location of their heterogeneity, and in the future may lead to an incorrect use of fertilizers, irrigation, agricultural machinery direction at harvest and so forth, that contrary to the objectives of precision agriculture. The main stage of technology in a particular sector of PA is to prepare maps, electronic maps. This operation is performed by scanning a suitable scale maps on paper or by computer processing of remote sensing data (aerial photography and satellite imagery in different frequency bands), obtained with the help of navigation equipment.

The aim of this work is to investigate the magnitude of geometric distortion of space images, after the elimination of distortions caused by terrain, shooting conditions and type of camera.

Main tasks of the research:

  1. Hold orthorectification of satellite images in the software package ENVI, using different digital models: ASTER GDEM, SRTM, GTOPO 30.
  2. Obtained measure and calculate the RMS deviation and systematic errors.
  3. Analyse the resulting accuracy of the processed images for use in precision farming technology.

Research object: geoinformation support in precision agriculture.

Research subject: remote sensing in precision agriculture.

3.Main part

For the study was chosen to supply satellite image data RPC in the city, Donetsk region, made satellite QuickBird. RPC – Rational Polynomial Coefficients‑rational polynomial coefficients (polynomials) – amendments for image processing based on a mathematical model of the satellite camera in the form of generalized approximating functions (rational polynomials) [6]. The main advantages are the QuickBird satellite broad band coverage (stage size – 16,5×16,5 km), high accuracy metric, offers complex polygon shapes, including extended objects 5 km wide.

The pixel size on the original photograph is 0,6 m on the ground, multispectral resolution – 2,44 m was previously made a preliminary analysis of the displacement in the image due to the topography, which in some places reached 25 m.

To eliminate geometric distortion tool was used – a software package ENVI, licenses provide the Department of Geodesy and geoifnormatiki Donetsk National Technical University by Pixel Solution, which is located in Kiev. In this program it is possible to carry out without the use of orthorectification (function Ortorectificatify QuickBird) and using (function Ortorectificatify QuickBird with Ground control) ground control points. Under orthorectification understand mathematically rigorous transformation of the source image in the orthogonal projection and elimination of distortions caused by terrain, shooting conditions and camera type [7]. For the study was selected the first option using a digital terrain model (hereinafter DEM). Information about the relief was obtained by downloading from public sites. As DEM models were chosen ASTER 2, SRTM 3, GTOPO 30. DEM data downloaded from sites http://reverb.echo.nasa.gov, http://srtm.csi.cgiar.org, https://lta.cr.usgs.gov/GTOPO30 respectively [8]. Orthorectification was conducted in the following sequence:

1) From the main menu were selected in March > Orthorectification > QuickBird > Ortorectify QuickBird (Fig. 1).

Main menu, select the function orthorectification

Figure 1 – Main menu, select the function orthorectification

2) In the opened window to download the required picture georeferenced by specifying the area and geodetic system, in which the shooting occurred. In this case, the area has been specified 37 and the system WGS84. Next loaded file with RPC coefficients (extension .rpc, .txt, .rpb).

3) In the next window shows the parameters orthorectification, which are selected to complete the operation (Fig. 2).

Settings window orthorectification

Figure 2 – Settings window orthorectification

The main parameters are – image resampling and input height. In image resampling is possible to select the required method resampling: in this case the selected method Bilinear bilinear interpolation image. Parameter input height determines how will be orthotransformation – using digital elevation model DEM (Digital Elevation Model) or midplane Fixed [9]. Transformation is performed using the DEM, so you need to add the DEM by clicking Select DEM File. A dialog box appears in which you select the file containing the DEM (Fig. 3).

Select the DEM

Figure 3 – Select the DEM

Orthorectification snapshot was performed using three digital models. For further research was selected transformed with DEM Aster, and adopted as the standard shot that was orthorectified in the software package ERDAS, using ground reference points. Reference photo has a local coordinate system, and treated – WGS84, which does not allow to evaluate the deviation obtained. To bring to the same coordinate system in the main menu tab Registration function is selected Image to image, then in the dialog box (Fig. 4) must specify which image is the reference, and which processed, not bound.

Determination of the reference image and georeferenced

Figure 4 – Determination of the reference image and georeferenced

Further, in both images were measured coordinates 20 characteristic points. For the operation, the default method is specified Coordinate – polynomial of first degree. Conversion above third degree is not recommended as it significantly decreases the reliability evaluation of the quality points. The mean square error was 3,2  m. To estimate the deviation on the reference picture and the treated were measured coordinates 1000 points (Fig. 5).

Measured point to the reference image (second frame) and orthorectified (third frame)

Figure 5 – Measured point to the reference image (second frame) and orthorectified (third frame)
(animation: 5 frames, 6 cycles of repetition, 147 kilobytes)

For each point were calculated deviation, the maximum X coordinate was 9 m, and by Y – 13 m. Mean square error (RMS Error) in orthorectified picture was 3 m. Received systematic errors close to 0: dX – 0,2 m , dY – 0,01 m. This suggests that the reduction in the same coordinate system have been made correctly.

Conclusion

  1. At the ENVI was made satellite image orthorectification QuickBird, the standard error was 3 m given accuracy is sufficient to determine the boundaries of the field, in accordance with the norm [2], which is 10 m.
  2. In precision agriculture for planning fertilizer application, yield monitoring, automatic data collection, accuracy is required – 1 m. Using more accurate digital models that are missing in the public domain, can afford to achieve the required accuracy.
  3. Using materials multispectral remote sensing data may parse irregularities soil properties for precision agriculture.

This master’s work is not completed yet. Final completion: January 2015. The full text of the work and materials on the topic can be obtained from the author or his head after this date.

References

  1. Н. И. Кобец Применение данных дистанционного зондирования земли в системах точного земледелия. – [Электронный ресурс] – http://www.ulrmc.org.ua.
  2. О. Г. Бойко Можливості використання ГІС / ДЗЗ технологій у точному землеробстві. – [Электронный ресурс] – pdaa.edu.ua.
  3. Т. М. Белавцева Технологии точного земледелия, их перспективы и возможность использования на мелиорированных землях. – М.: ФГНУ ЦНТИ Мелиоводинформ, 2009. – 110 – с.
  4. А. В. Барладин, П. Д. Ярошук Использование ГИС и ДЗЗ – технологий в сельском хозяйстве. – [Электронный ресурс] – http://repository.crimea.edu.
  5. И. В. Бычков, Л. В. Нефедьев, Г. М. Ружников, Н. Г. Луковников Внедрение геоинформационных технологий и навигационных систем в задачах точного земледелия. – [Электронный ресурс] – http://it.nsu.ru.
  6. Геометрическая обработка данных со спутника QuickBird. – [Электронный ресурс] – http://geopriz.webrost.lgg.ru.
  7. Обзор космических съемочных систем высокого разрешения. – [Электронный ресурс] –  http://vinek.narod.ru.
  8. Ортокоррекция космических снимков с использованием RPC. – [Электронный ресурс] – http://gis-lab.info.
  9. Выбор и характеристика DEM. – [Электронный ресурс] – http://gisa.ru.
  10. Учебное пособие «Программный комплекс ENVI». – М.: Компания Совзонд, 2007. – 265 с.