Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
A review of three machine learning models using UAV derived photogrammetry for detecting and measuring tree stumps
Thesis (MSc)--Stellenbosch University, 2023.
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | en_ZA |
| Published: |
Stellenbosch : Stellenbosch University
2023
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867614087898202112 |
|---|---|
| access_status_str | Open Access |
| author | Hubbard, Alan John |
| author2 | Talbot, Bruce |
| author_browse | Hubbard, Alan John Talbot, Bruce |
| author_facet | Talbot, Bruce Hubbard, Alan John |
| author_sort | Hubbard, Alan John |
| collection | Thesis |
| dc_rights_str_mv | Stellenbosch University |
| description | Thesis (MSc)--Stellenbosch University, 2023. |
| format | Thesis |
| id | oai:scholar.sun.ac.za:10019.1/128399 |
| institution | Stellenbosch University (South Africa) |
| language | en_ZA |
| last_indexed | 2026-06-10T12:46:28.519Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from SUNScholar — Stellenbosch University Repository |
| publishDate | 2023 |
| publishDateRange | 2023 |
| publishDateSort | 2023 |
| publisher | Stellenbosch : Stellenbosch University |
| publisherStr | Stellenbosch : Stellenbosch University |
| record_format | dspace |
| source_str | SUNScholar — Stellenbosch University Repository |
| spelling | oai:scholar.sun.ac.za:10019.1/128399 A review of three machine learning models using UAV derived photogrammetry for detecting and measuring tree stumps Hubbard, Alan John Talbot, Bruce Ackerman, Simon Stellenbosch University. Faculty of AgriSciences. Dept. of Forest and Wood Science. Forests and forestry -- Remote sensing Sustainable forestry -- Information technology Forest management -- Information technology Forests and forestry -- Mathematical models Artificial intelligence UCTD Thesis (MSc)--Stellenbosch University, 2023. ENGLISH SUMMARY: Accurate biophysical data from stumps left on post-harvested sites is required to ascertain volume loss from inefficient harvesting techniques or volume gain from biomass utilisation. Recently advances in digital aerial photogrammetry data from unmanned aerial vehicles (UAVs) and machine learning, and object detection algorithms, have led to the increased use of this technology in forestry for remote sensing. Stumps, being mostly uniform in distribution and shape, are ideal objects for machine detection on digital orthomosaics. Resultant data from machine learning algorithms enables the possibility of estimating stump diameter and heights from virtual sources. In this study we trained three different machine learning model types, namely, Faster Region-based Convolutional Neural Network (R-CNN), Single Shot Multibox Detector (SSD) and You-Only-Look-Once (YOLO). We assessed the detection rates of each model and compared metrics by using similarly annotated images. The resultant bounding boxes that encapsulated detected stumps were used to calculate diameters and compared to actual. Stump heights were determined using multiple methods which were also compared to actual height values. We found that visible stumps in post-harvested sites could be detected with high rates of accuracy, with almost perfect precision from some object detection models, albeit at low levels of recall. Overall, all three model types had an F1-score of above 73% with the best model attaining an F1-score of 89%. Diameters, although successfully calculated, produced an overestimation from actual in most cases. Similarly, calculated stump heights were underestimated in most cases. The objectives of this study were met in that insights into using machine learning algorithms for stump detection were broadened. The ability to process photogrammetry data quickly and accurately, with good estimations of diameter and height values, provides a useful tool to industry for estimating biomass volume from stumps left on post-harvested sites. Future development in technology will certainly improve accuracy and turn-around time of available data. AFRIKAANSE OPSOMMING: Akkurate data van afgesaagde bome wat op geoeste plantasies agtergelaat word, word benodig om volumeverlies van ondoeltreffende oestegnieke, asook die volume van agtergeblewe biomassa, vas te stel. Onlangse ontwikkeling in digitale lugfotogrammetrie data van onbemande lugvoertuie (hommeltuie) en masjienleer, en voorwerpopsporingsalgoritmes, het gelei tot die toenemende gebruik van hierdie tegnologie in bosbou vir afstandswaarneming. Stompe, wat meestal eenvormig in plasing en vorm is, is ideale voorwerpe vir masjienopsporing op digitale ortomosaieke. Die data van masjienleeralgoritmes gee die moontlikheid om stompdeursnitte en hoogtes vanaf virtuele bronne te bepaal. In hierdie studie het ons drie verskillende tipes masjienleermodelle opgelei, naamlik R-CNN, SSD en YOLO. Ons het die opsporingssyfers van elke model met mekaar vergelyk. Die gevolglike omsluitende reghoeke wat om opgespoorde stompe getrek word, is gebruik om deursnitte te bereken en met werklike maates te vergelyk. Stomphoogtes is met behulp van verskeie metodes bepaal en ook met werklike hoogtes vergelyk. Ons het gevind dat sigbare stompe in geoesde plantasies met hoe akkuraatheid opgespoor kon word, met byna perfekte presisie (“precision”) van sommige voorwerpopsporingsmodelle, alhoewel teen lae vlakke van herroeping (“recall”). In die algemeen het al drie modeltipes 'n F1-telling van meer as 73% gehad, met die beste model wat 'n F1-telling van 89% behaal het. Deursnitte, alhoewel suksesvol bereken, het in die meeste gevalle 'n oorskatting van die werklike opgelewer. Op soortgelyke wyse is berekende stomphoogtes in die meeste gevalle onderskat. Die doelwit van hierdie studie is bereik deurdat insigte oor die gebruik van masjienleeralgoritmes vir stompopsporing verbreed is. Die vermoe om fotogrammetrie data vinnig en akkuraat te verwerk, met goeie skattings van deursnit- en hoogtewaardes, bied 'n nuttige hulpmiddel aan die industrie om biomassavolume te skat vanaf stompe wat op geoeste plantasies gelaat word. Toekomstige ontwikkeling in tegnologie sal beslis die akkuraatheid en spoed van berekening van beskikbare data verbeter. Masters 2023-03-05T20:05:06Z 2023-08-30T13:12:59Z 2023-03 2023-08-31T09:18:35Z 2023-03-05T20:05:06Z 2023-08-31T09:18:35Z 2023-03-05T20:05:06Z 2023-03 Thesis https://scholar.sun.ac.za/handle/10019.1/128399 en_ZA Stellenbosch University application/pdf 100 pages : illustrations, maps application/pdf Stellenbosch : Stellenbosch University |
| spellingShingle | Forests and forestry -- Remote sensing Sustainable forestry -- Information technology Forest management -- Information technology Forests and forestry -- Mathematical models Artificial intelligence UCTD Hubbard, Alan John A review of three machine learning models using UAV derived photogrammetry for detecting and measuring tree stumps |
| title | A review of three machine learning models using UAV derived photogrammetry for detecting and measuring tree stumps |
| title_full | A review of three machine learning models using UAV derived photogrammetry for detecting and measuring tree stumps |
| title_fullStr | A review of three machine learning models using UAV derived photogrammetry for detecting and measuring tree stumps |
| title_full_unstemmed | A review of three machine learning models using UAV derived photogrammetry for detecting and measuring tree stumps |
| title_short | A review of three machine learning models using UAV derived photogrammetry for detecting and measuring tree stumps |
| title_sort | review of three machine learning models using uav derived photogrammetry for detecting and measuring tree stumps |
| topic | Forests and forestry -- Remote sensing Sustainable forestry -- Information technology Forest management -- Information technology Forests and forestry -- Mathematical models Artificial intelligence UCTD |
| url | https://scholar.sun.ac.za/handle/10019.1/128399 |
| work_keys_str_mv | AT hubbardalanjohn areviewofthreemachinelearningmodelsusinguavderivedphotogrammetryfordetectingandmeasuringtreestumps AT hubbardalanjohn reviewofthreemachinelearningmodelsusinguavderivedphotogrammetryfordetectingandmeasuringtreestumps |