Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

A hybrid classifier for remote sensing applications

G S Ruppert1, M Schardt, G Balzuweit

  • 1Institute of Image Processing, Joanneum Research, Graz, Austria. georg.ruppert@joanneum.ac.at

International Journal of Neural Systems
|February 1, 1997
PubMed
Summary

This study introduces a hybrid classifier for remote sensing land use classification. It offers a user-friendly, effective alternative to traditional methods, especially for complex data.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Remote versus in-clinic cochlear implant programming during rehabilitation: clinical outcomes and patient perspectives.

Disability and rehabilitation. Assistive technology·2026
Same author

Corrigendum to "Positive impact of early-probiotic administration on performance parameters, intestinal health and microbiota populations in broiler chickens" [Poultry Science, Volume 103/issue 12,104401].

Poultry science·2025
Same author

Analyzing optimal control techniques in a nonlinear fractional Rubella model with the Atangana-Baleanu derivative.

Computers in biology and medicine·2025
Same author

Real World Multi-centre UK Review of Nivolumab Monotherapy in Metastatic Endometrial Cancer With Mismatch Repair Deficiency During COVID-19.

Clinical oncology (Royal College of Radiologists (Great Britain))·2025
Same author

Ipsilateral motor pathways related to proximal upper limb function in tetraplegia.

Clinical neurophysiology : official journal of the International Federation of Clinical Neurophysiology·2025
Same author

Positive impact of early-probiotic administration on performance parameters, intestinal health and microbiota populations in broiler chickens.

Poultry science·2024

Area of Science:

  • Remote Sensing
  • Machine Learning
  • Image Classification

Background:

  • Accurate land use classification is crucial for environmental monitoring and resource management.
  • Traditional classifiers often require expert knowledge and struggle with complex, non-normally distributed data.
  • Existing methods like Maximum Likelihood and Multilayer Perceptions have limitations in terms of user interaction and data distribution handling.

Purpose of the Study:

  • To develop and evaluate a hybrid unsupervised and supervised classifier for land use classification of remote sensing images.
  • To assess the performance of the proposed classifier against established methods like Maximum Likelihood and Multilayer Perceptions.
  • To highlight the advantages of the hybrid classifier in terms of reduced user interaction and suitability for complex datasets.

Related Experiment Videos

Main Methods:

  • A hybrid approach combining unsupervised Neural Gas (NG) quantization with supervised majority voting for codebook labeling.
  • Unsupervised quantization of the entire satellite image using the Neural Gas algorithm.
  • Supervised labeling of the resulting codebook using ground truth data via majority voting.

Main Results:

  • The hybrid classifier demonstrates performance comparable to Maximum Likelihood (ML) and slightly below Multilayer Perceptions (MLP).
  • The proposed method requires no expert knowledge for training and classification post-ground truth collection.
  • It shows superior performance compared to ML for complex, non-normally distributed classes.
  • Significantly less user interaction is needed compared to ML and other classifiers.

Conclusions:

  • The hybrid Neural Gas classifier provides an efficient and user-friendly solution for land use classification in remote sensing.
  • It effectively handles complex, non-normally distributed classes, outperforming Maximum Likelihood in such scenarios.
  • The method's reduced need for expert intervention makes it a practical choice for large-scale remote sensing image analysis.