A Review on integrating Photogrammetry and Deep Learning for Semantic Segmentation Appliations

Abstract

The monitoring of large areas is essential to assess their conditions and patterns of change and support decision-making from which different ways are look for ways and optimize these problems and techniques such as photogrammetry are used to assess changes in large areas and, in conjunction with semantic segmentation, are important to extract the necessary information and to provide classification analysis at the pixel level for an accurate assessment of irregular shapes The reference ground sample distance (GSD) method translates pixels into a unit of measurement, facilitating accurate image scale calculations and precise measurements. This method explores these strategies using deep learning, potentially improving accuracy and simplifying segmentation. It also looks at different works that talk about how segmentation and photogrammetry can be used in different areas, including their methods and results, to find the best ways to do things, adapt, and lead to new developments in the field.