東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁 到查詢結果 [ null ]

切換: 標籤 | MARC模式 | ISBD

Extending the Utility of Photogramme...

Eldefrawy, Mahmoud,

FindBook

Google Book

Amazon

博客來

Extending the Utility of Photogrammetry (Structure From Motion) With AI in the Context of Agriculture Applications /

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Extending the Utility of Photogrammetry (Structure From Motion) With AI in the Context of Agriculture Applications // Mahmoud Eldefrawy.
作者:	Eldefrawy, Mahmoud,
面頁冊數:	1 electronic resource (126 pages)
附註:	Source: Dissertations Abstracts International, Volume: 86-12, Section: B.
Contained By:	Dissertations Abstracts International86-12B.
標題:	Computer science. -
電子資源:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31939610
ISBN:	9798286441761

Extending the Utility of Photogrammetry (Structure From Motion) With AI in the Context of Agriculture Applications /
Eldefrawy, Mahmoud,

Extending the Utility of Photogrammetry (Structure From Motion) With AI in the Context of Agriculture Applications /Mahmoud Eldefrawy. - 1 electronic resource (126 pages)

Source: Dissertations Abstracts International, Volume: 86-12, Section: B.

Structure from motion and multi-view stereo (SfM/MVS) provide an inexpensive and accessible way to generate detailed 3D models from RGB images. These techniques are widely applied across various scales-from digitizing small heritage artifacts to monitoring large agricultural fields. Their popularity stems from the ability to create high-resolution, spatially accurate reconstructions using relatively inexpensive equipment such as cameras. However, a fundamental assumption of SfM/MVS is that the scene remains static during image capture. When this assumption is violated, such as in the presence of moving objects, reconstruction quality can significantly degrade or even fail altogether due to incorrect feature matching and triangulation. SfM/MVS is particularly valuable in environmental monitoring and change detection applications, where comparing 3D reconstructions over time provides insights into dynamic processes. In agriculture, for example, uncrewed aerial systems (UAS) and SfM/MVS allow for large-scale, repeated data collection across growing seasons. These systems can also provide geo-referenced outputs by linking reconstructions to global coordinate systems. However, aligning these reconstructions accurately over time remains a key challenge, which requires careful execution, especially in environments that change rapidly. Misalignment can result in unreliable measurements and hinder the ability to monitor growth or detect change effectively. In addition to spatial accuracy, SfM/MVS outputs often contain rich semantic information-such as individual plant structures or field patterns-that is rarely leveraged. This becomes a limitation in precision agriculture, where actionable insights are needed for timely decisions. For instance, cotton stand count is a crucial indicator during the early germination phase, typically seven days after planting helping to make replanting decisions.While SfM/MVS can produce accurate representations of the field, it does not inherently extract this type of semantic data. To address these limitations, our work introduces improvements at three key stages of the SfM/MVS pipeline. First, during pre-processing, we incorporate object detection and object tracking based on semantic information to identify and separate moving elements in the scene. This helps isolate different motion systems and preserves the integrity of static scene components, significantly improving reconstruction quality in dynamic environments. Second, we evaluate several reconstruction and alignment strategies specifically suited for agricultural fields. Through extensive testing, we find that four-dimensional SfM (4D SfM) offers consistent and accurate plant height measurements across time with minimal manual intervention-an important step toward full automation. Finally, we introduce a post-processing step to extract semantic insights by developing a cotton stand count model. The model is based on the YOLOv7 architecture and is trained on a synthetic dataset to improve generalization. It demonstrates strong performance on unseen real-world data, making it a practical tool for early crop monitoring. Together, these enhancements create a more robust, automated, and semantically enriched SfM/MVS pipeline tailored for dynamic agricultural applications.

English

ISBN: 9798286441761Subjects--Topical Terms:

523869
Computer science.
Subjects--Index Terms:

Structure from motion

Extending the Utility of Photogrammetry (Structure From Motion) With AI in the Context of Agriculture Applications /
LDR:04769nmm a22004093i 4500 001 2435818
005 20251029060358.5
006 m o d
007 cr|nu||||||||
008 261011s2025 miu||||||m |||||||eng d
020 $a 9798286441761
035 $a (MiAaPQD)AAI31939610
035 $a AAI31939610
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Eldefrawy, Mahmoud, $e author. $3 3826263
245 1 0 $a Extending the Utility of Photogrammetry (Structure From Motion) With AI in the Context of Agriculture Applications / $c Mahmoud Eldefrawy.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2025
300 $a 1 electronic resource (126 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 86-12, Section: B.
500 $a Advisors: King, Scott Committee members: Bhandari, Mahendra; Medrano, Antonio; Starek, Michael; Garza, Susan.
502 $b Ph.D. $c Texas A&M University - Corpus Christi $d 2025.
520 $a Structure from motion and multi-view stereo (SfM/MVS) provide an inexpensive and accessible way to generate detailed 3D models from RGB images. These techniques are widely applied across various scales-from digitizing small heritage artifacts to monitoring large agricultural fields. Their popularity stems from the ability to create high-resolution, spatially accurate reconstructions using relatively inexpensive equipment such as cameras. However, a fundamental assumption of SfM/MVS is that the scene remains static during image capture. When this assumption is violated, such as in the presence of moving objects, reconstruction quality can significantly degrade or even fail altogether due to incorrect feature matching and triangulation. SfM/MVS is particularly valuable in environmental monitoring and change detection applications, where comparing 3D reconstructions over time provides insights into dynamic processes. In agriculture, for example, uncrewed aerial systems (UAS) and SfM/MVS allow for large-scale, repeated data collection across growing seasons. These systems can also provide geo-referenced outputs by linking reconstructions to global coordinate systems. However, aligning these reconstructions accurately over time remains a key challenge, which requires careful execution, especially in environments that change rapidly. Misalignment can result in unreliable measurements and hinder the ability to monitor growth or detect change effectively. In addition to spatial accuracy, SfM/MVS outputs often contain rich semantic information-such as individual plant structures or field patterns-that is rarely leveraged. This becomes a limitation in precision agriculture, where actionable insights are needed for timely decisions. For instance, cotton stand count is a crucial indicator during the early germination phase, typically seven days after planting helping to make replanting decisions.While SfM/MVS can produce accurate representations of the field, it does not inherently extract this type of semantic data. To address these limitations, our work introduces improvements at three key stages of the SfM/MVS pipeline. First, during pre-processing, we incorporate object detection and object tracking based on semantic information to identify and separate moving elements in the scene. This helps isolate different motion systems and preserves the integrity of static scene components, significantly improving reconstruction quality in dynamic environments. Second, we evaluate several reconstruction and alignment strategies specifically suited for agricultural fields. Through extensive testing, we find that four-dimensional SfM (4D SfM) offers consistent and accurate plant height measurements across time with minimal manual intervention-an important step toward full automation. Finally, we introduce a post-processing step to extract semantic insights by developing a cotton stand count model. The model is based on the YOLOv7 architecture and is trained on a synthetic dataset to improve generalization. It demonstrates strong performance on unseen real-world data, making it a practical tool for early crop monitoring. Together, these enhancements create a more robust, automated, and semantically enriched SfM/MVS pipeline tailored for dynamic agricultural applications.
546 $a English
590 $a School code: 1417
650 4 $a Computer science. $3 523869
650 4 $a Agricultural engineering. $3 3168406
653 $a Structure from motion
653 $a Multi-view stereo
653 $a Uncrewed aerial systems
690 $a 0984
690 $a 0800
690 $a 0539
710 2 $a Texas A&M University - Corpus Christi. $b Geospatial Computing Sciences. $e degree granting institution. $3 3826264
720 1 $a King, Scott $e degree supervisor.
773 0 $t Dissertations Abstracts International $g 86-12B.
790 $a 1417
791 $a Ph.D.
792 $a 2025
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31939610