Getting Started With Introduction To Remote Sensing Fourth Edition
I picked up this book when I was first trying to figure out how to actually make sense of satellite imagery instead of just clicking through a few algorithms blindly. The 4th edition by James A. Ridd goes pretty hard on the fundamentals. You know, the stuff people skip when they're in a rush to jump into deep learning segmentation pipelines. It covers electromagnetic radiation, the physics behind how sensors work, and then moves into image processing and classification. Not everything you need is in there, but it's a solid foundation if you stick with it. The book works best when you read it in order early on. The chapters build on each other, especially the sections on radiometric calibration and atmospheric correction. I learned the hard way that skipping the radiometry chapter meant I kept getting garbage results when I tried to do change detection on Landsat scenes from different years. Same sensor, different acquisition dates, and my outputs were completely miscalibrated because I didn't understand what the numbers in the reflectance files actually meant. There's a practical problem that comes up with this book. It assumes you have some math background and doesn't hold your hand through the Fourier transforms or the stats behind classification. If you're coming from a programming background and want to dive in fast, you might find yourself pausing at the spectral signature chapter just to look up what the underlying equations are doing. I spent maybe two weeks working through the math while reading alongside it, but it was worth it. The classification chapters made way more sense after that.
One thing the book gets right that a lot of other introductory texts gloss over is the difference between spectral resolution, spatial resolution, temporal resolution, and radiometric resolution. You see people conflate those all the time in forums. They'll say their NDVI isn't working because their spatial resolution is too low, when really the issue is the spectral bands don't capture the red and near-infrared channels correctly for the vegetation index to mean anything. That distinction matters more than people realize. Another thing that trips people up: the book presents classification as this neat step where you pick a method and get results. In practice, supervised classification is a huge pain. I was working on a land cover mapping project using Sentinel-2 data and spent about a day and a half just picking training sites. The real problem was that the classes I thought were distinct spectrally weren't. Bare soil and built-up areas overlap significantly in several band combinations. I had to go back and restructure my class definitions based on field knowledge, not just spectral clustering. If you're working with the examples in the book, expect to supplement them. The datasets they use are simplified and idealized. Real satellite imagery has clouds, cloud shadows, topographic effects, and sensor artifacts that the textbook examples don't really address in depth. For atmospheric correction, the book mentions techniques but doesn't walk you through ENVI or SNAP in detail. I ended up using Sen2Cor for the Sentinel data and just cross-referencing with the book's explanations to understand what the processor was doing under the hood.
There are limitations to this textbook as a standalone resource. It was published before some of the recent shifts in open-source remote sensing tooling became mainstream. It doesn't cover Google Earth Engine, Python-based libraries like rasterio or xarray, or the workflow changes that have happened in the last few years. You should treat it as a theory and methodology reference, not a software manual. For people just starting out, I'd suggest reading the first five chapters thoroughly, then using the later chapters as reference material as you encounter specific problems in your own work. Don't try to finish the whole thing cover to cover before doing anything practical. Pick a small dataset, follow along with what the book explains, and fill in the gaps with whatever documentation or tutorials you need at the time. The book is available through standard academic publishers and resellers. Taylor & Francis publishes it. You can find it on major book retail sites and through institutional library access if you have that. It's not free, but it's reasonably priced for what it covers.
