多期多尺度影像結合深度學習於邊坡地貌變異判識之初探(2/2)-影像處理方法及公路邊坡影像類型適用性探討

本計畫為2年期計畫之第2年期，主要研究成果計有：(1)蒐集相關文獻並探討深度學習應用於邊坡地貌變異判識之方法與可行性；(2)分析多尺度之衛星、航測、UAV…等遙測載具及空拍影像處理方法；(3)訪談實務應用單位瞭解公路邊坡維管制度及需求性；(4)說明深度學習神經網路模型對於邊坡地貌判識任務之適用性。

This is the second year of a two-year research project, with the following main research achievements: (1) Collected relevant literature and discussed the methods and feasibility of applying deep learning to slope geomorphic variation identification; (2) Analyzed multi-scale remote sensing platforms including satellites, aerial survey, UAV, etc., as well as aerial image processing methods; (3) Interviewed practical application units to understand the highway slope maintenance and management systems and their corresponding demands; (4) Explained the applicability of deep learning neural network models to slope geomorphic identification tasks.

搜集汇总

背景与挑战

背景概述

该数据集为一项2年期研究计划的第二期，重点探讨了多尺度影像处理方法及其在公路边坡地貌变异识别中的适用性。研究内容包括文献综述、多源遥感影像分析、实地需求访谈，并评估了深度学习模型在此任务中的应用效果。