Introduction to remote sensing image registration

Moigne, Jacqueline Le

doi:10.1109/igarss.2017.8127519

Cited by 17 publications

(10 citation statements)

References 19 publications

(24 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The applications of image registration are crucial in a gamut of applications such as remote sensing, [1][2][3] infrared physics, 4 biomedical images, 5,6 farming and cultivation, 7 and image mosaic. 8 Furthermore, the growing number of devices for data acquisition and the pervasiveness of such devices necessitate the usage of image registration.…”

Section: Introductionmentioning

confidence: 99%

Multistage Forward Path Regenerative Genetic Algorithm for Brain Magnetic Resonant Imaging Registration

et al. 2022

View full text Add to dashboard Cite

In image registration, the search space used to compute the optimal transformation between the images depends on the group of pixels in the vicinity. Favorable results can be achieved by significantly increasing the number of neighboring pixels in the search space; however, this strategy increases the computational load, thus making it challenging to realize the most desirable solution in a reasonable amount of time. To address the mentioned problem, the genetic algorithm is used to find the optimum solution and the solution lies in finding the best chromosomes. In rigid image registration problem, chromosomes contain a set of three parameters, x-translation, y-translation, and rotation. The genetic algorithm iteratively improves chromosomes from generation to generation and selects the best one having the best fittest value. Chromosomes with high fitness value are the ones with an optimal solution where the template image best aligns reference image. Fitness function in the genetic algorithm for image registration problem uses similarity measure index measure to find the amount of similarity between two images. The best fittest value is the one with a high similarity measure that shows the best-aligned template and reference image. Here we used the structural similarity index measure in fitness function that helps in evaluating the best chromosome, even for the compressed images with low quality, intensity nonuniformity (INU), and noise degradation. Building on the genetic algorithm, we propose a novel approach called multistage forward path regenerative genetic algorithm (MFRGA), abbreviated as MFRGA, with reducing search space at each stage. Compared with the single stage of genetic algorithm, our approach proved to be more reliable and accurate in terms of finding true rigid image transformation for alignment. At each increasing stage of MFRGA, results are computed with decreasing search space and increasing precision levels. Moreover, to prove the robustness of our algorithm, we utilized compressed images of brain magnetic resonant imaging that vary in compression qualities ranging from 10 to 100. Furthermore, we added noise levels of 1%, 3%, 5%, 7%, and 9% with an INU of 20% and 40%, respectively, provided by the online BrainWeb simulator. We achieved the monomodal rigid image registration that proves to be successful using MFRGA, even when the noise is critical, the compression quality is the least, and the intensity is nonuniform.

show abstract

Section: Introductionmentioning

confidence: 99%

Multistage Forward Path Regenerative Genetic Algorithm for Brain Magnetic Resonant Imaging Registration

et al. 2022

View full text Add to dashboard Cite

show abstract

“…These images are acquired from different times and viewpoints by different sensors [1]. Thus, image registration is an essential preprocess step in many remote sensing applications because the main process, which includes change detection, image fusion, image mosaic, environment monitoring, and map updating can be drastically influenced by these differences [1,2]. Many types of image registration techniques have been developed in the areas of remote sensing over the past few decades.…”

Section: Introductionmentioning

confidence: 99%

A CNN-Based High-Accuracy Registration for Remote Sensing Images

Lee

Sim

2021

Remote Sensing

View full text Add to dashboard Cite

In this paper, a convolutional neural network-based registration framework is proposed for remote sensing to improve the registration accuracy between two remote-sensed images acquired from different times and viewpoints. The proposed framework consists of four stages. In the first stage, key-points are extracted from two input images—a reference and a sensed image. Then, a patch is constructed at each key-point. The second stage consists of three processes for patch matching—candidate patch pair list generation, one-to-one matched label selection, and geometric distortion compensation. One-to-one matched patch pairs between two images are found, and the exact matching is found by compensating for geometric distortions in the matched patch pairs. A global geometric affine parameter set is computed using the random sample consensus algorithm (RANSAC) algorithm in the third stage. Finally, a registered image is generated after warping the input sensed image using the affine parameter set. The proposed high-accuracy registration framework is evaluated using the KOMPSAT-3 dataset by comparing the conventional frameworks based on machine learning and deep-learning-based frameworks. The proposed framework obtains the least root mean square error value of 34.922 based on all control points and achieves a 68.4% increase in the matching accuracy compared with the conventional registration framework.

show abstract

“…The process of image registration is to find the pixel space mapping relationship between the sensed and reference images, thereby, transforming the sensed image into the geometric coordinate system of the reference image. The sensed and reference images are usually the same scene taken by different times, sensors, or viewpoints [1][2][3]. Registration is a significant task in the application of remote sensing images, and feature-based methods are often recommended to achieve it due to their effectiveness [4].…”

Section: Introductionmentioning

confidence: 99%

An Instance Segmentation Based Framework for Large-Sized High-Resolution Remote Sensing Images Registration

Jia

et al. 2021

Remote Sensing

View full text Add to dashboard Cite

Feature-based remote sensing image registration methods have achieved great accomplishments. However, they have faced some limitations of applicability, automation, accuracy, efficiency, and robustness for large high-resolution remote sensing image registration. To address the above issues, we propose a novel instance segmentation based registration framework specifically for large-sized high-resolution remote sensing images. First, we design an instance segmentation model based on a convolutional neural network (CNN), which can efficiently extract fine-grained instances as the deep features for local area matching. Then, a feature-based method combined with the instance segmentation results is adopted to acquire more accurate local feature matching. Finally, multi-constraints based on the instance segmentation results are introduced to work on the outlier removal. In the experiments of high-resolution remote sensing image registration, the proposal effectively copes with the circumstance of the sensed image with poor positioning accuracy. In addition, the method achieves superior accuracy and competitive robustness compared with state-of-the-art feature-based methods, while being rather efficient.

show abstract

Introduction to remote sensing image registration

Cited by 17 publications

References 19 publications

Multistage Forward Path Regenerative Genetic Algorithm for Brain Magnetic Resonant Imaging Registration

Multistage Forward Path Regenerative Genetic Algorithm for Brain Magnetic Resonant Imaging Registration

A CNN-Based High-Accuracy Registration for Remote Sensing Images

An Instance Segmentation Based Framework for Large-Sized High-Resolution Remote Sensing Images Registration

Contact Info

Product

Resources

About