AVHYAS: A Python Based QGIS Plugin for Advanced Hyperspectral Image Analysis



Original Source Here

Advanced Hyperspectral Data Analysis Software(AVHYAS) plugin is a Python-3 based Quantum-GIS (QGIS) plugin designed to process and analyse hyperspectral (Hx) images. Starting with version 1.0, AVHYAS serves as a free and open-source platform for sharing and distributing Hx data analysis methods among research scholars, scientists and potential end-users. It is developed to guarantee full usage of present and future Hx airborne or space-borne sensors and provides access to advanced algorithms for Hx data processing. The software is freely available and offers a range of basic and advanced tools such as atmospheric correction (for airborne AVIRIS-NG image), standard processing tools as well as powerful machine learning and Deep Learning interfaces for Hx data analysis.

The Overarching aim of AVHYAS is to provide free-of-charge and user-friendly access to advanced approaches for Hx data processing for both beginners and advanced users. The AVHYAS is integrated into the QGIS classic menu to extend its range of available applications. It can also be used with any Mx imagery for specific applications. A standard workflow was adopted for effectively integrating machine learning approaches to the QGIS environment. This way, functionalities were implemented that include the standard methods available in other proprietary software (e.g., ENVI, ArcGIS) or non-commercial/open-source software (e.g., EnMAP-Box) and the advanced algorithms (e.g. DL based classification) which are not available in the standard software (technical documentation is available at the website: https://sites.google.com/view/avhyas-sac-isro/home)

Key Functionalities

  1. Atmospheric Correction Module for AVIRIS-NG Data
  2. Basic Tools Module: It contains sub-modules such as sensor-utility, data-subset, spectral plot, scatter plot, scaling, and Region of Interest (or Class) Separability Analysis.
  3. Pre-processing Module: It contains sub-modules such as Dimensionality Reduction (DR), General Purpose Utility (e.g. cloud removal),Feature Extraction, Data Transformation and Savitsky-Golay-Filtering.
  4. Data Quality Analysis Modules: This module can be used for analysing the quality of the Hx images in terms of spectral and spatial characteristics.
  5. Un-mixing Module: The Un-mixing module performs end-member extraction and the abundance estimation on Hx images. Sub-modules consist of Material-Count, end-member extraction, abundance estimation, sparse based un-mixing, interactive scatter plot visualization, visualization of un-mixing results, and un-mixing error analysis.
  6. Classification Module: Supervised, Un-Supervised classification and Segmentation
  7. Deep Learning Module: DL module consists of state-of-the-art DL algorithms for the classification of Hx images. The inference sub-module of DL provides an interface for performing prediction based on a trained deep learning model. The user has to provide a valid H5 file which contains the weight and biases of the model being trained (using AVHYAS). The design of the UI for classification workflow has been adopted from the EnMap-Box toolbox. Model performance is printed as html reports and the assessment report will pop up on the default browser.
  8. Regression Module: This module can be applied for the estimation of bio-physical properties of the target material
  9. Fusion Module: It contains AROSICS module for co-registration and fusion module for Hx-Mx fusion.
  10. Geo-Physical Applications and Spectral Indices Modules: PROSAIL simulation module (for the simulation of vegetation spectra) is equipped with the spectral response function of different satellite payloads (includes Indian satellites).
Key Functionalities

Major Applications

Sub-pixel Mineral Classification, Agricultural Crop Classification, Crop-Forest health monitoring, Forest Classification, Hyperspectral Data quality check, Hx-Mx co-registration, Target bio-physical property (Soil property, biomass, canopy chlorophyll content, water quality, etc.) estimation, Target identification (mineral, camouflage, etc. ), Hx-mx data fusion etc.

AVHYAS is not limited to HRS data obtained from the airborne or space-borne sensors, it can also be used for analyzing the data acquired by the handheld/tabletop Hx cameras. Moreover, the chance for the AVHYAS to become an evolving plugin with a constantly growing set of applications is high, given its flexibility of integrating new algorithms for different Hx sensors and new powerful ML/DL libraries.

Conclusion

AVHYAS integrates powerful machine-learning and deep-learning algorithms to perform various data analysis tasks for extracting information from Hx data. Basic and advanced algorithms were incorporated in the AVHYAS toolbox aiming at the extension of the user community (in the field of hyperspectral remote sensing) in India and providing the most powerful algorithms for the analysis of the present and the future Hx-imaging sensor data. The AVHYAS plugin development was thus driven by the idea of familiarising advanced Hx image analysis algorithms with the multidisciplinary community. It was used for training the academia and research community to provide hands-on experience on Hx image analysis and always received positive responses. The AVHYAS plugin is developed at Space Applications Centre (SAC), Indian Space Research Organisation (ISRO), Ahmedabad, Gujarat.

Lyngdoh, R. B., Anand S. S., Ahmad, T., Rathore, P. S., Mishra, M., Gupta, P. K., & Misra, A. (2021). AVHYAS: A Free and Open Source QGIS Plugin for Advanced Hyperspectral Image Analysis. arXiv preprint arXiv:2106.12776.

Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the official policy or position of any agency of the Indian government.

Author: Anand S Sahadevan, Ph. D [ Data Scientist, Hyperspectral Remote Sensing ]

AI/ML

Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot

%d bloggers like this: