MOŽNOST UPORABE METODE ANALIZE SLIKE ZA ZGODNJO NAPOVED PRIDELKA JABOLK V SLOVENIJI

Tomše, Edvard

| | SLO | ENG | Cookies and privacy

Bigger font | Smaller font

First page > Show document

Show document

Title:	MOŽNOST UPORABE METODE ANALIZE SLIKE ZA ZGODNJO NAPOVED PRIDELKA JABOLK V SLOVENIJI
Authors:	ID Tomše, Edvard (Author) ID Stajnko, Denis (Mentor) More about this mentor...
Files:	UNI_Tomse_Edvard_2009.pdf (5,54 MB) MD5: ADE4C1D409438F3EA59AF7EDE769C03B PID: 20.500.12556/dkum/a515da06-aaaf-427a-b38d-ebb9f9d67bce
Language:	Slovenian
Work type:	Undergraduate thesis
Organization:	FKBV - Faculty of Agriculture and Life Sciences
Abstract:	V diplomskem delu smo leta 2008 izvajali poskus zgodnje napovedi pridelka jabolk s pomočjo analize slike v okviru napovedovanja pridelka jabolk za evropski program Prognosfruit. Poskus je potekal na območju Prekmurja, Slovenskih goric, okolice Maribora s Slovensko Bistrico, okolice Celja ter v Savski dolini z okolico Krškega. V poskusu so bile vključene sorte Braeburn, Elstar, Fuji, Gala, Idared, Jonagold ter Zlati delišes. Namen poskusa je bil napovedati pridelek jabolk za Slovenijo v letu 2008 ter proučiti natančnost napovedovanja pridelka jabolk z uporabo metode analize slike. Ob tem pa s statističnim proučevanjem ugotoviti morebitne vplive pridelovalca, pridelovalnega območja ter sorte na natančnost napovedi pridelka jabolk, hkrati pa iskanje možnosti izboljšanja napovedi pridelka v prihodnje. V začetku julija 2008 smo fotografirali izbrana drevesa v sadovnjakih, nato pa s pomočjo računalniške analize slike napovedali pričakovan pridelek. Statistična analiza za pridelovalce je pokazala, da se napovedan in tehtan pridelek statistično značilno najbolj razlikujeta pri pridelovalcu številka 5, najbolj točno napoved pa smo podali pri pridelovalcu številka 9, kjer je bila razlika med napovedanim in dejanskim pridelkom 99 kg/ha. Po pridelovalnih območjih je bila največja razlika v Prekmurju (Io = 2,04), najmanjša razlika med napovedanim in dejanskim pridelkom pa je bila v pridelovalnem območju Maribor (Io = 1,00) oziroma 146 kg/ha razlike Najmanjše odstopanje pri sorti Gala (Io = 1,02) oziroma 531 kg/ha, največje odstopanje pa je bilo pri sorti Fuji in sicer za 8.760 kg/ha. Rezultati ugotovitev potrjujejo hipotezo, da metoda omogoča natančno napovedovanje pridelkov, vendar obstaja še precej faktorjev, ki bi se jih dalo izboljšati.
Keywords:	jabolka / število plodov / vizualizacija / obdelava fotografije / napoved pridelka / Slovenija
Place of publishing:	Maribor
Year of publishing:	2009
PID:	20.500.12556/DKUM-10535
NUK URN:	URN:SI:UM:DK:S4QFMYWY
Publication date in DKUM:	04.06.2009
Views:	3721
Downloads:	222
Metadata:
Categories:	FKBV
:	Copy citation

Average score:	(0 votes)
Your score:	Voting is allowed only for logged in users.
Share:

Hover the mouse pointer over a document title to show the abstract or click on the title to get all document metadata.

Secondary language

Language:	English
Title:	THE POSSIBLE APPLICATION OF THE IMAGE ANALYSIS METHOD FOR THE EARLY FORECAST OF APPLE YIELD IN SLOVENIA
Abstract:	In the diploma work a test of the early forecast of apple yield with the help of a image analysis within an apple yield prediction for the European programme Prognosfruit was performed in 2008. The test was performed in Prekmurje, Slovenske gorice, on the outskirts of Maribor with Slovenska Bistrica, on the outskirts of Celje and in Sava Valley, together with the outskirts of Krško. The test included Braeburn, Elstar, Fuji, Gala, Idared, Jonagold and Golden Delicious apple varieties. The purpose of the resaerch was to predict the apple yield in Slovenia for 2008 and to examine the precision of the apple yield prediction with the image analysis method. Furthermore, the possible influences of the producer, production area and the sort on the precision of the apple yield prediction was statistically examined. At the same time the possibility to improve the future yield predictions was looking for. In July 2008 the selected trees in the orchard were captured by a digital camera. then the expected yield was computerized by the help of image analysis. The statistical analysis showed the biggest difference between the predicted and weighted yield by the producer number 5, and the most accurate forecast by the producer number 9 (99 kg/ha). The biggest difference was detected in Prekmurje (Io = 2,04), and the smallest in Maribor region (Io = 1,00) or 146 kg/ha between the predicted and harvested yield. The smallest deviation was measured by the Gala variety (Io = 1,02) or 531 kg/ha and the biggest by Fuji variety (8.760 kg/ha). The results of the experiment confirmed the hypothesis that our method enables a quite precise yield prediction, but there are still factors, which might be improved.
Keywords:	apple / fruit number / vizualization / image processing / yield forcast / Slowenien

Comments

Leave comment

You must log in to leave a comment.

Comments (0)

0 - 0 / 0

There are no comments!

Back