Neural networks in Remote Sensing

Neural networks are a type of algorithm used in image classification and other areas of machine learning. They are based on the structure and function of the human brain, and are made up of layers of interconnected nodes, called neurons. The neurons in the input layer receive input data, and the neurons in the output layer produce the predicted class label. The neurons in the hidden layers process the input data and transmit it to the output layer.

In the context of image classification, the input data is the image pixels and the output is the class label. The neurons in the input layer receive the image pixels, and the neurons in the output layer produce the predicted class label. The neurons in the hidden layers process the image pixels and transmit it to the output layer.

Neural networks are known for their ability to learn from data and improve their performance over time. They are considered as a supervised learning algorithm, it requires a labeled dataset to train the network on. The labeled dataset contains the image and its corresponding class label. The algorithm uses the labeled dataset to learn the relationships between the image pixels and the class labels, and then uses this knowledge to classify new images.

One of the main advantages of neural networks is that they can handle a large amount of data and can be used to solve a wide range of image classification problems. They can also be used in combination with other methods such as decision trees and support vector machines to improve the performance of the classification algorithm.

Overall, neural networks are a powerful method for image classification, they are known for their ability to learn from data and improve their performance over time, can handle a large amount of data and can be used to solve a wide range of image classification problems and also they can be used in combination with other methods to improve the performance of the classification algorithm

Comments

Accuracy Assessment

Accuracy assessment is the process of checking how correct your classified satellite image is . 👉 After supervised classification, the satellite image is divided into classes like: Water Forest Agriculture Built-up land Barren land But classification is done using computer algorithms, so some areas may be wrongly classified . 👉 Accuracy assessment helps to answer this question: ✔ "How much of my classified map is correct compared to real ground conditions?" Goal The main goal is to: Measure reliability of classified maps Identify classification errors Improve classification results Provide scientific validity to research 👉 Without accuracy assessment, a classified map is not considered scientifically reliable . Reference Data (Ground Truth Data) Reference data is real-world information used to check classification accuracy. It can be collected from: ✔ Field survey using GPS ✔ High-resolution satellite images (Google Earth etc.) ✔ Existing maps or survey reports 🧭 Exampl...

REMOTE SENSING INDICES

Remote sensing indices are band ratios designed to highlight specific surface features (vegetation, soil, water, urban areas, snow, burned areas, etc.) using the spectral reflectance properties of the Earth's surface. They improve classification accuracy and environmental monitoring. 1. Vegetation Indices NDVI – Normalized Difference Vegetation Index Formula: (NIR – RED) / (NIR + RED) Concept: Vegetation reflects strongly in NIR and absorbs in RED due to chlorophyll. Measures: Vegetation greenness & health Uses: Agriculture, drought monitoring, biomass estimation EVI – Enhanced Vegetation Index Formula: G × (NIR – RED) / (NIR + C1×RED – C2×BLUE + L) Concept: Corrects for soil and atmospheric noise. Measures: Vegetation vigor in dense canopies Uses: Tropical rainforest mapping, high biomass regions GNDVI – Green Normalized Difference Vegetation Index Formula: (NIR – GREEN) / (NIR + GREEN) Concept: Uses Green instead of Red ...

Landsat 8 Band designation and Band Combination.

Landsat 8 Band designation and Band Combination. Landsat 8-9 Operational Land Imager (OLI) and Thermal Infrared Sensor (TIRS) Bands Wavelength (micrometers) Resolution (meters) Band 1 - Coastal aerosol 0.43-0.45 30 Band 2 - Blue 0.45-0.51 30 Band 3 - Green 0.53-0.59 30 Band 4 - Red 0.64-0.67 30 Band 5 - Near Infrared (NIR) 0.85-0.88 30 Band 6 - SWIR 1 1.57-1.65 30 Band 7 - SWIR 2 2.11-2.29 30 Band 8 - Panchromatic 0.50-0.68 15 Band 9 - Cirrus 1.36-1.38 30 Band 10 - Thermal Infrared (TIRS) 1 10.6-11.19 100 Band 11 - Thermal Infrared (TIRS) 2 11.50-12.51 100 Vineesh V Assistant Professor of Geography, Directorate of Education, Government of Kerala. https://www.facebook.com/Applied.Geography http://geogisgeo.blogspot.com

Energy Interaction with Atmosphere and Earth Surface

In Remote Sensing , satellites record electromagnetic radiation (EMR) that is reflected or emitted from the Earth. Before reaching the sensor, radiation interacts with: The Atmosphere The Earth's Surface These interactions control how satellite images look and how we interpret them. I. Interaction of EMR with the Atmosphere When solar radiation travels from the Sun to the Earth, four main processes occur: 1. Absorption Definition: Absorption occurs when atmospheric gases absorb radiation at specific wavelengths and convert it into heat. Main absorbing gases: Ozone (O₃) → absorbs Ultraviolet (UV) Carbon dioxide (CO₂) → absorbs Thermal Infrared Water vapour (H₂O) → absorbs Infrared Concept: Atmospheric Windows These are wavelength regions where absorption is very low, allowing radiation to pass through the atmosphere. Remote sensing depends on these windows. For example, satellites like Landsat 8 use visible, near-infrared, and thermal bands located in atmospheric windows. 2. Trans...

Change Detection

Change detection is the process of finding differences on the Earth's surface over time by comparing satellite images of the same area taken on different dates . After supervised classification , two classified maps (e.g., Year-1 and Year-2) are compared to identify land use / land cover changes . Goal To detect where , what , and how much change has occurred To monitor urban growth, deforestation, floods, agriculture, etc. Basic Concept Forest → Forest = No change Forest → Urban = Change detected Key Terminologies Multi-temporal images : Images of the same area at different times Post-classification comparison : Comparing two classified maps Change matrix : Table showing class-to-class change Change / No-change : Whether land cover remains same or different Main Methods Post-classification comparison – Most common and easy Image differencing – Subtract pixel values Image ratioing – Divide pixel values Deep learning methods – Advanced AI-based detection Examples Agricult...

Vineesh V, Geography

Search This Blog