magistrsko delo magistrskega študijskega programa II. stopnje Strojništvo
Blaž Menegatti (Author), Drago Bračun (Mentor), Igor Petrović (Co-mentor)

Abstract

V magistrski nalogi obravnavamo problematiko prepoznavanja in štetja rastlin koruze z metodami globokega učenja za potrebe natančnega kmetijstva. Na osnovi slik koruznega polja, zajetih z brezpilotnim letalnikom, je ustvarjena podatkovna baza z označenimi listi in rastlinami, ki služi za učenje globokega modela BranchedERFNet. Model, ki omogoča sočasno semantično in instančno segmentacijo listov in rastlin, je naučen in testiran na rastlinah v različnih fenoloških fazah in pod različnimi pogoji zajema slik. Naučen model kljub slabši segmentaciji uspešno zazna in prešteje koruzo v 82 \% primerov, pri čemer je boljša uspešnost na koruzi v zgodnejših fenoloških fazah. Iz rezultatov smo ugotovili, da model dosega dobre rezultate pri različnih višinah zajemanja slik, slabše pa v primeru gosto prepletenih listov sosednjih rastlin koruze. Natančnost segmentacije je bila pri različnih fenoloških fazah rastlin 25-odstotna in je ključen dejavnik za manj uspešno natančno določanje centrov rastlin koruze.

Keywords

magistrske naloge;natančno kmetijstvo;štetje rastlin koruze;daljinsko zaznavanje;strojni vid;globok model učenja;konvolucijska nevronska mreža;semantična segmentacija;instančna segmentacija;

Data

Language: Slovenian
Year of publishing:
Typology: 2.09 - Master's Thesis
Organization: UL FS - Faculty of Mechanical Engineering
Publisher: [B. Menegatti]
UDC: 004.85:004.932.72:633.15(043.2)
COBISS: 231775235 Link will open in a new window
Views: 133
Downloads: 39
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Corn crop detection using deep learning methods
Secondary abstract: In the master's thesis, we address the problem of recognizing and counting corn plants using deep learning methods for precision agriculture. Based on images of cornfields captured by a drone, a dataset with labeled leaves and plants was created to train the deep learning model BranchedERFNet. This model, which enables simultaneous semantic and instance segmentation of leaves and plants, was trained and tested across different phenological stages and image capture conditions. The trained model successfully detects and counts corn plants with an accuracy of 82\%, despite some limitations in segmentation quality. The performance is higher for corn plants in earlier phenological stages. The results indicate that the model performs well at different image capture heights but struggles when leaves of neighboring plants are densely intertwined. The accuracy of segmentation at different phenological stages of the plants was 25\% and is a key factor in the less successful precise determination of the centers of corn plants.
Secondary keywords: master thesis;precision agriculture;corn crop plant counting;remote sensing;computer vision;deep learning model;convolutional neural network;semantic segmentation;instance segmentation;
Type (COBISS): Master's thesis/paper
Study programme: 0
Thesis comment: Univ. Ljubljana, Fak. za strojništvo
Pages: XXII, 55 str.
ID: 26131460