ServicenavigationHauptnavigationTrailKarteikarten


Forschungsstelle
BLW
Projektnummer
10.20_1
Projekttitel
"ICT-AGRI"-Projekt "3D-Mosaic" (ERA-NET)

Texte zu diesem Projekt

 DeutschFranzösischItalienischEnglisch
Schlüsselwörter
Anzeigen
Anzeigen
-
Anzeigen
Kurzbeschreibung
-
-
-
Anzeigen
Methoden
-
-
-
Anzeigen
Projektziele
-
-
-
Anzeigen
Publikationen / Ergebnisse
-
-
-
Anzeigen

Erfasste Texte


KategorieText
Schlüsselwörter
(Deutsch)
ICT-AGRI, Entscheidungsunterstützungssystem DSS, Informations und Kommunikationstechnologien IKT/ICT, Photogrammetrie, Geoinformationssysteme GIS, Global Navigation Satellite System GNSS, Laserscanner LiDAR, Blattflächenindex LAI
Schlüsselwörter
(Englisch)
ICT-AGRI, decision support system DSS, information and communication technologies ICT, photogrammetry, geographic information system GIS, Global Navigation Satellite System GNSS, laser scanner LiDAR, leaf area index LAI
Schlüsselwörter
(Französisch)
ICT-AGRI, système d'aide à la décision DSS, technologies d'information et de communication TIC/ICT, photogrammétrie, système d'information géographique SIG/GIS, Global Navigation Satellite système GNSS, scanner laser LiDAR, indice de surface foliaire LAI
Kurzbeschreibung
(Englisch)

In the context of current global changes, assuring the supply of fresh produce and increasing economic viability are priority targets within plantation management. In the cultivation of tree crops, water is a critical input factor and irrigation is necessary in all European countries. As a matter of fact, input requirements in an orchard vary in space and time due to the variability of climate, soil and plant growth. Spatial patterns of soil and plant properties can be regarded as a 3D mosaic. Thus, an optimum orchard management has to address seasonal and spatial variability in soil and micro-climate that affect plant growth and fruit development. However, irrigation today is commonly managed at the orchard level and not at the tree level. Uniform irrigation frequently creates sub-optimum conditions with some parts of the orchard having insufficient water while other parts suffer from water logging and subsequent oxygen shortage. Consequently, water is wasted and/or yield and fruit quality are reduced.

Tree-scale leaf area index and fruit load and individual fruit-scale size, water content, and maturity-related pigmentation represent information vital for orchard management/irrigation decision making. Until now, these parameters have been underutilized due to lack of automation of their monitoring.

The target of 3D-Mosaic is to promote precision management of orchards by means of a decision support system (DSS) aiming to optimize efficiency of inputs including water and to diminish the environmental footprint of fruit production. The DSS will apply information and communication technologies (ICT) for precision management of the most economically relevant tree crops, apple and citrus. For this purpose, sensors, monitoring strategies, information processing and decision support systems will be developed. Together, these will produce maps for orchard management including irrigation.

The project is structured into six work packages:

1. Adaptation of an autonomous platform for transport of the mobile sensors and for implementation of efficient monitoring strategies

2. Development of a mobile image analysis system for detection of trees and fruits

3. Development of mobile fruit sensors for non-invasive fruit quality determination on the tree

4. Field trials for testing the system and data acquisition

5. Geographical information system for 3D data management

6. Development of a decision support system to generate management maps

3D-Mosaic will employ a horizontal approach bringing together work groups with multidisciplinary expertise, facilities, and infrastructure. Cooperative field tests and exchanges between partners of each work package will be conducted.

Field trials will be performed in Turkey (citrus) and Germany (apple). Orchards will be subjected to measurements throughout the season regarding climate, soil variability and induced irrigation differences as well as variability in plant parameters on tree and organ levels. Yield will be quantified and fruit quality evaluated.

During the harvest season two field tests of all partners are scheduled. Sensors based on multispectral, 2D and 3D photogrammetric techniques will be operated by an unmanned, autonomous platform for monitoring diagnostic plant parameters. Tree and fruit parameters will be measured and feasibility of sensor operation will be tested. Mosaic structure of soil and plants will be determined using the collected data.

Data management will be achieved with geographic information system (GIS). Management zones and irrigation maps considering LAI, yield, and fruit data will be determined and the DSS concept will be verified.

Results of 3D-Mosaic will be communicated by appropriate media to stakeholders from agri business, scientific community (technical papers, conference presentations), politics (presentations in committees), and general public (homepage, booth on agricultural fair). A workshop will be held to demonstrate the concept and realization. Fruit sensors as well as specific GIS can be commercialized by the participating SME.

Collaborate supervision of young scientists; establishment of links with the research community and industry through ICT-AGRI will enhance the intensity of networks in the field of precision horticulture.

Methoden
(Englisch)

2.1 Leaf Area Index (LAI)

calculations out of the tree silhouette for leaf area index (LAI) with LiDAR in side view arrangement For the 2D photogrammetric methods the algorithms used to measure soil cover will be adapted.

2.2 Fruit Development

measures for each tree row from both sides of the row. The fruit count method use a 2D and 3D photogrammetry system with side view from both sides of each tree row. To improve the detection rate, the fusion of three-with two-dimensional detection will be established. In contrast to the stereo camera two-dimensional cameras allow much higher resolutions and color fidelity needed for pattern and color detection. The information is used to overlay the 3D photogrammetric image.

The 3D photogrammetric system will be a camera like Bumblebee XB3. The ZHAW will adapt the necessary software for the 3D fruit recognition. The 2D texture overlay with the fruit separation for the3D image will be produced from the University of Kassel. To test and calibrate the system, field trials will be done in Switzerland at ART.

Projektziele
(Englisch)

WP 2. Vision System (ZHAW, Switzerland; UniKas, Germany; ART, Switzerland)

Improving the understanding of the influence of heterogeneity on tree growth and fruit yield, a spatial addressing of trees and their fruits is necessary. A new combination of Global Navigation Satellite System (GNSS), laser scanner (LiDAR) and 2D & 3D photogrammetric survey are used to address leaf area index and single fruits. The number of detected fruits with their attributed size is the basis for the yield estimation.  The procedures that segment the 3D-Data will be extended for multispectral data analysis. The segmentation procedure processes acquired data separately and in the end step superposes the results to receive more reliable results. The superposing procedure of the segmented objects is the main issue in this WP. All data will be elaborated and stored specifically for each estimated tree.  Tasks in the WP2 are:

2.1 Leaf Area Index (LAI) (ZHAW, ART, UniKas)

An indicator of the leaf area index will be calculated out of the tree silhouette. Two different methods for leaf area index (LAI) calculation will be competitively tested, one with LiDAR in side view arrangement (ZHAW) and one with 2D photogrammetry (UniKas). The scanner is already in use on the autonomous platform. For the 2D photogrammetric methods the algorithms used to measure soil cover will be adapted for this goal too (UniKas, ZHAW). As the localization of each tree is known, the measured tree profiles can be attributed to each single tree.

Deliverables

D2.1.1: Definition of interfaces and coordination of the autonomous platform (month 4)

D2.1.2: Semi-automatedsystem for the first field trial (month 9)

D2.1.3: Platform with all spatial detection units with fixed spatial arrangement (month 18)

D2.1.4: Software for automated spatial coordinate calculation (month 9)

2.2 Fruit Development (ZHAW, ART, UniKas)

Task 2.2 aims at automated yield estimation in different growth stages. These parameters will be measured for each tree row from both sides of the row.

The fruit count method use a 2D and 3D photogrammetry system with side view from both sides of each tree row.To improve the detection rate, the fusion of three-with two-dimensional detection will be established. In contrast to the stereo camera two-dimensional cameras allow much higher resolutions and color fidelity needed for pattern and color detection. The information is used to overlay the 3D photogrammetric image.

The 3D photogrammetric system will be a camera like Bumblebee XB3. The ZHAW will adapt the necessary software for the 3D fruit recognition. The 2D texture overlay with the fruit separation for the3D image will be produced from the University of Kassel. To test and calibrate the system, field trials will be done in Switzerland at ART.

Deliverables

D2.2.1: Field calibration for fruit detection (month 6-9)

D2.2.2: Development of algorithms for fruit detection (month 18)

Publikationen / Ergebnisse
(Englisch)
Zugehörige Dokumente