Geri Dön

Damage detection of twospotted spider mites (Tetranychus urticae koch) in strawberry plants using unmanned aerial vehicle multispectral images

Başlık çevirisi mevcut değil.

PDF İndir

  1. Tez No: 771890
  2. Yazar: MURAT ÇAKIR
  3. Danışmanlar: Belirtilmemiş.
  4. Tez Türü: Yüksek Lisans
  5. Konular: Biyoloji, Botanik, Ormancılık ve Orman Mühendisliği, Biology, Botany, Forestry and Forest Engineering
  6. Anahtar Kelimeler: Belirtilmemiş.
  7. Yıl: 2016
  8. Dil: İngilizce
  9. Üniversite: University of Florida
  10. Enstitü: Yurtdışı Enstitü
  11. Ana Bilim Dalı: Belirtilmemiş.
  12. Bilim Dalı: Belirtilmemiş.
  13. Sayfa Sayısı: 60

Özet

Recent studies have indicated that remote sensing by UAVs has great potential for precision agriculture. The twospotted spider mite (TSSM) is one of the main pests affecting strawberry production in Florida. This study investigates the potential to detect spider mite counts and damage assessment through multispectral images consisted of five bands in Blue, Green, Red, Red Edge and NIR area of the spectrum. Fifty strawberry plants with varying levels of twospotted spider mite infestation were sampled from a commercial strawberry field located in Plant City, FL. Images were captured weekly for 4 weeks from March 16th to April 5th of 2016. Linear regression models of the log-transformed TSSM mite counts and visually assessed damage classification were developed using individual band values, Normalized Difference Vegetation Indices (NDVI) and Simple Ratio (SR) as independent variables. As second independent variables, predator mite (Phytoseilus Persimilis) presence or absence was noted on each leaf and the standard deviation of the bands or index values within the canopy were tested in multiple regression analysis. A Leave-One-Out Cross validation (LOOCV) was used to assess the results. The Random Forest (RF) regression algorithm was also applied using the logtransformed mite counts and damage class as dependent variables and each of the band value, NDVI index, and SR index data groups as well as their standard deviation values as independent variables. In implementing the RF algorithm, 70% of the samples were set aside for validation and RMSE values were computed using the predicted values of this dataset. The results indicate that mite counts can be modeled using individual band values and vegetation indices in early infestation stages. In later infestation stages, no mite count models were statistically significant (p > 0.05). However, visually assessed damage class models were significant (p < 0.05) in later infestation stage promoting the severity of the mite-induced damage affecting the plant spectral signature across the observed spectrum. The results also showed that the visible bands can be used to model mite counts at early infestation stage, while almost all bands used in vegetation indices can be used to model the damage class in the later infestation stage. The RF regression algorithm produced superior results (Table 3-6 and Table 3- 9) compared with linear regressions as shown by the LOOCV validation dataset (Table 3-1, Table 3-2 and Table 3-3). The findings of the thesis demonstrate the utility of both the linear regression and random forest methods in estimating spider mite counts and damage classes. They are capable methods to predict spider mites and damage classes. However, studying only one variety with a relatively small sample size and the spatial resolution of the images can be considered as limitations. A more comprehensive study using different strawberry varieties and spatial resolution extending along the growing season is recommended as an extension to this study.

Özet (Çeviri)

Özet çevirisi mevcut değil.

Benzer Tezler

  1. Tez No

    850381

    Derin öğrenme teknikleri ile bazı bağ zararlılarının oluşturduğu hasarın belirlenmesi

    Determination of the damage caused by some vineyard pests with deep learning techniques

    TAHSİN UYGUN

    Doktora

    Türkçe

    Türkçe

    2024

    ZiraatTokat Gaziosmanpaşa Üniversitesi

    Biyosistem Mühendisliği Ana Bilim Dalı

    DOÇ. DR. MEHMET METİN ÖZGÜVEN

    PROF. DR. DÜRDANE YANAR

  2. Tez No

    252050

    Sonlu eleman modeli güncellemesi tekniğinde benzetilmiş tavlama algoritması kullanılarak mekanik sistemlerde hasar tespiti

    Damage detection of mechanical systems in finite element model updating method using simulated annealing algorithm

    YILMAZ AVCI

    Yüksek Lisans

    Türkçe

    Türkçe

    2008

    İnşaat Mühendisliğiİstanbul Teknik Üniversitesi

    İnşaat Mühendisliği Ana Bilim Dalı

    DOÇ. DR. PELİN GÜNDEŞ BAKIR

  3. Tez No

    761677

    Damage detection of laminated composite structures using inverse finite element method

    Lamine kompozit yapılarda ters sonlu elemanlar yöntemi ile hasar tespiti

    FARAZ GANJDOUST

    Yüksek Lisans

    İngilizce

    İngilizce

    2022

    Mekatronik MühendisliğiSabancı Üniversitesi

    Mekatronik Mühendisliği Ana Bilim Dalı

    DR. ÖĞR. ÜYESİ ADNAN KEFAL

  4. Tez No

    546217

    Tarihi yığma köprü hasarlarının analitik model ve deneysel yöntemlerle belirlenmesi

    Damage detection of historical masonry bridges with analytical model and experimental techniques

    EMRE ALPASLAN

    Doktora

    Türkçe

    Türkçe

    2019

    İnşaat MühendisliğiOndokuz Mayıs Üniversitesi

    İnşaat Mühendisliği Ana Bilim Dalı

    PROF. DR. ZEKİ KARACA

  5. Tez No

    789915

    Savunma sanayi için imal edilen gemilerde iletilebilirlik yaklaşımı ile desteklenen hasar tespit metodolojisinin DH36 çelik yapılar üzerinde uygulanması

    Application of damage assessment methodology on DH36 steel structures supported by the transmissibility approach on ships manufactured for the defense industry

    MERVE KARA

    Yüksek Lisans

    Türkçe

    Türkçe

    2023

    Makine MühendisliğiKocaeli Üniversitesi

    Makine Mühendisliği Ana Bilim Dalı

    PROF. DR. SEDAT KARABAY

    DR. HAKAN UÇAR

Geri Dön