A Geographical Analysis of Image Interpretation Methods

Abstract

Image interpretation methods, procedures for relating image pattern to ground conditions, are essential to our use of remote sensing imagery. These methods can be analyzed in respect to the role of ancillary information in the image interpretation process. In general, those procedures that are comparatively independent of ancillary information can be applied in varied geographic settings. Because almost all interpretation procedures depend to some extent upon ancillary information, a detailed and integrated knowledge of the cultural and physical landscape is a prerequisite for both manual and automated interpretation procedures.

Climate change promotes hybridisation between deeply divergent species

“…The two images were imported into ArcGIS 10.3.1 (ESRI©), with a resolution of 4.5 m per pixel, and were georeferenced using administrative boundaries as reference points (RMS error = 6.23 m for the 2014 image; RMS error = 4.48 m for the 2006 image). The images were interpreted using direct recognition (Campbell, 1978), considering five discrete land-use classes that hold a clear ecological importance for both toad species: agriculture, forests, natural clearings, natural vegetation (other than forests), and artificial areas. For both images, a vector layer (format shapefile, ESRI©) was produced at a 1:25,000 scale.…”

Integrating Remote Sensing and Geographic Information Systems

“…Photo interpretation demands the application of reasoning and logic, based on use of multiple data sources and formalized using aids such as interpretation keys (Campbell 1978, Estes et al 1983. In recent decades, there have been many efforts to implement, and improve upon, strategies and methods for image analysis developed during the pre-digital era.…”

“…Hallmarks of image interpretation logic include (1) use of a systematic strategy that proceeds from 'knowns' to 'unknowns,' and (2) use of inference and 'convergence of evidence' exploiting observed relationships between multiple data types (image and ancillary data). The process of image interpretation often involves use of heuristics and/or 'rules' based on expert knowledge and observation (e.g., rules about biogeographic relationships between vegetation zonation, elevation, slope, and aspect in mountainous areas) (Campbell 1978, Estes et al 1983). Here, we examine several approaches to integrating GIS with remote sensing to improve extraction of information from digital images.…”