Remote sensing
No Thumbnail Available
Date
2026
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Oum El Bouaghi
Abstract
This handbook presents the scientific foundations of remote sensing and their application to urban techniques management, for second-year Licence students with an introductory background in geography, physics and cartography. It defines remote sensing as obtaining information about a surface without physical contact, traces its history from balloon photography to Landsat-1 in 1972 and cloud platforms, and sets out the seven-element acquisition chain from energy source to application. That chain rests on physical principles: electromagnetic radiation, the spectrum, interactions with matter and the atmosphere, and the black-body laws. These lead to the spectral signature, which makes surfaces distinguishable.
Building on these principles, the handbook turns to instruments and data. Platforms (ground, aerial, drone, satellite), orbits (geostationary, sun-synchronous) and passive or active sensors, including radar and LiDAR, determine what can be observed. Each sensor is described by four resolutions, spatial, spectral, temporal and radiometric, which cannot all be maximised at once, so selection starts from the question to be answered. Chapter 5 compares the main systems: the free Landsat and Sentinel missions, the very-high-resolution commercial SPOT, IKONOS and Pléiades, and the daily-coverage MODIS. It recommends a two-tier strategy combining free medium-resolution data with very-high-resolution data for detail.
The third part forms the digital core of the course. An image is a raster of pixels holding digital numbers, calibrated to reflectance and displayed as true- or false-colour composites. It is then processed through geometric correction, georeferencing and radiometric and atmospheric correction, improved with filters, and distilled into spectral indices: NDVI for vegetation, NDBI for built-up areas and NDWI for water. The image becomes a thematic map through unsupervised or supervised classification (Maximum Likelihood, Random Forest, SVM) or object-based analysis (OBIA), and its accuracy is always assessed with the confusion matrix, overall accuracy and the kappa coefficient.
Finally, these tools are applied to urban questions: land cover, urban sprawl, green-space inventory, heat islands, change detection and hazards such as floods and ground deformation. Their value is fully realised through GIS integration in a six-step workflow (acquire, pre-process, analyse, validate, integrate, produce), applied in the Constantine case study using two Landsat scenes from 2000 and 2020 and free software such as QGIS and SNAP. The handbook closes with current trends (open data, Google Earth Engine, deep learning, drones) and with the need for honesty about uncertainty, methodological transparency and respect for privacy: reliability depends on the whole chain, from photon to policy.