Related Experiment Video
Updated: Jun 8, 2026
10:59
Reconstruction of 3-Dimensional Histology Volume and its Application to Study Mouse Mammary Glands
Published on: July 26, 2014
HAIRIS: a method for automatic image registration through histogram-based image segmentation
Hernâni Gonçalves1, José Alberto Gonçalves, Luís Corte-Real
1Departamento de Geociências, Ambiente e Ordenamento do Território, Faculdade de Ciências, Universidade do Porto, 4169-007 Porto, Portugal. hernani.goncalves@fc.up.pt
Summary
A new method called HAIRIS uses histogram-based image segmentation for automatic image registration. This approach achieves subpixel accuracy for remote sensing images, overcoming challenges in rotation and translation registration.
Area of Science:
- Computer Vision
- Remote Sensing
- Image Processing
Background:
- Automatic image registration remains a significant challenge across various scientific fields.
- Existing methods have limited adoption in applications like remote sensing.
- Accurate image registration is crucial for multi-temporal and multi-sensor data analysis.
Purpose of the Study:
- To propose a novel method for automatic image registration using histogram-based segmentation.
- To address the limitations of current registration techniques in remote sensing applications.
- To achieve high accuracy in registering images with rotational and translational differences.
Main Methods:
- Developed the Histogram-based Image Registration (HAIRIS) method.
- Combined multiple image segmentations based on histogram mode delineation with a relaxation parameter.
- Characterized segmented objects using area, ellipse axis ratio, perimeter, and fractal dimension.
- Employed a robust statistical procedure for object matching.
Main Results:
- Achieved accuracy below 1° for rotation and subpixel accuracy for translation in simulated datasets.
- Demonstrated effectiveness on satellite imagery with varying spectral content and simulated translation.
- Successfully applied to real-world remote sensing data from different sensors, dates, and viewing angles.
- Validated HAIRIS's capability to register multi-temporal and multi-sensor images with spectral variations.
Conclusions:
- HAIRIS offers a robust solution for automatic image registration, particularly for remote sensing.
- The method achieves high accuracy for images with rotational and translational differences.
- HAIRIS demonstrates potential for broad application in analyzing diverse remote sensing datasets.
