| DC 欄位 |
值 |
語言 |
| DC.contributor | 土木工程學系 | zh_TW |
| DC.creator | 錢冠宇 | zh_TW |
| DC.creator | Kuan-Yu Chien | en_US |
| dc.date.accessioned | 2023-1-18T07:39:07Z | |
| dc.date.available | 2023-1-18T07:39:07Z | |
| dc.date.issued | 2023 | |
| dc.identifier.uri | http://ir.lib.ncu.edu.tw:88/thesis/view_etd.asp?URN=110325003 | |
| dc.contributor.department | 土木工程學系 | zh_TW |
| DC.description | 國立中央大學 | zh_TW |
| DC.description | National Central University | en_US |
| dc.description.abstract | 土石方之管理需投注大量資源以及人力。每年臺灣有超過 3 億立方公尺的剩餘土石方，相關業者亟需採行有效率之經營以進行妥善管理。本研究旨在發展智慧型計算增強深度傳感檢測，應用於以 U-Net 輔助之土石方管理。回顧之過去文獻主題包含土石方計算、應用無人機技術檢測之可行性、物件偵測、以及 U-Net 之設定。使用無人機於土石方堆置場進行範圍 43,939 平方公尺之實地飛行深度感測。飛行高度為海拔 60 公尺、速度每秒 4.7 公尺， 10 分 51 秒之總飛行時間中總共擷取 276 張高解析度航拍影像，其航程重疊率為 80%，照片重疊率為 70%。經由 U-Net 分析，土石方堆置場檢測之平均準確率達 92.5%。比對 2021 年 10 月至 2022 年 8 月，10 個月之土石進場申報量，本研究建置之模型檢測到該場域實際增加 13.84%體積的土方量，此結果大致符合土方場之實際運作狀況。本研究之成果亦可提供相關業者更具效率之土石方量計算方法。
| zh_TW |
| dc.description.abstract | Earthwork management requires massive resources and manpower to deal with over 30 million cube meters of earth annually in Taiwan. An efficient solution is desired for practitioners to operate earthwork management in the field. The research goal is to develop a computational ntelligence-enhanced depth sensing detection on earthwork management
using U-Net. The literature review suggests earthwork calculation, Unmanned Aerial Vehicle (UAV) detection feasibility, object detection, and settings for U-Net. Field aerial depth sensing is performed using UAV at a pre-determined earthwork site covering for a 43,939 square meter area. The total UAV airborne time is 10 minutes and 51 seconds at steady 60 meters above sea leave with 4.7m/sec for speed. A total of 276 high-resolution photos are btained with a course overlap rate at 80% and photo overlap rate at 70%. The average accuracy rate for detection on earthwork site yielded from U-Net is 92.5%. With the comparison to the actual earth imported to the field during 10 months 2021/10-
2022/08), the proposed model detects +13.84% incremental volume to the field. The finding that matches empirical operation for an earthwork field can also provide practitioners with an efficient management way for earthwork calculation.
| en_US |
| DC.subject | 土石方管理 | zh_TW |
| DC.subject | 無人機 | zh_TW |
| DC.subject | 物件偵測 | zh_TW |
| DC.subject | 辨識 | zh_TW |
| DC.subject | 智慧型計算 | zh_TW |
| DC.subject | Earthwork management | en_US |
| DC.subject | Unmanned Aerial Vehicle (UAV) | en_US |
| DC.subject | object detection | en_US |
| DC.subject | recognition | en_US |
| DC.subject | computational intelligence | en_US |
| DC.title | 智慧型計算結合深度攝影影像應用於土方 空拍測量之研究 | zh_TW |
| dc.language.iso | zh-TW | zh-TW |
| DC.title | Developing computational intelligence-enhanced depth sensing detection on earthwork management | en_US |
| DC.type | 博碩士論文 | zh_TW |
| DC.type | thesis | en_US |
| DC.publisher | National Central University | en_US |