Article
Computer Science, Information Systems
Yunbai Qin, Jialiang Li, Pinqun Jiang, F. Jiang
Summary: The paper introduces a novel stitching method that locates overlapped regions and records feature points, guides image warping with a local warp model and arc function weight model, showing better performance than other state-of-the-art methods. The approach effectively improves matching efficiency, eliminates color difference and ghosting, and maintains accuracy and stability in complex scenes.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Geochemistry & Geophysics
Donghai Guo, Jun Chen, Linbo Luo, Wenping Gong, Longsheng Wei
Summary: Our UAV image stitching strategy effectively improves alignment accuracy and achieves natural-looking results through the proposed global alignment strategy, shape-preserving warp, and LPM matching method.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Computer Science, Information Systems
Xin Chen, Mei Yu, Yang Song
Summary: This study proposes an optimized seam-driven image stitching method considering depth, color, and texture information. By introducing depth information, the method can find the seam that adapts to the depth of the scene and effectively avoid passing through salient objects, resulting in high-quality stitching results.
Article
Environmental Sciences
Wanli Xue, Zhe Zhang, Shengyong Chen
Summary: This study proposes a novel image stitching method that can identify and eliminate ghosts through multi-component collaboration without object distortion, segmentation, or repetition, and validates the effectiveness of the proposed method through experiments.
Article
Construction & Building Technology
Jiepeng Liu, Tong Jiao, Shuai Li, Zhou Wu, Y. Frank Chen
Summary: This paper proposes a vision-based model using a deep learning network and symbol-patching method for planning welding trajectories of welding robots. Experimental results show that the proposed method achieves a high recall rate under different lighting conditions and demonstrates robustness and accuracy in detecting and localizing weld seams.
AUTOMATION IN CONSTRUCTION
(2022)
Article
Physics, Multidisciplinary
Yizhi Cong, Yan Wang, Wenju Hou, Wei Pang
Summary: Feature detection and correct matching are crucial for image stitching. Existing methods often suffer from insufficient matching points in low-textured or repetitively-textured areas. This paper proposes a novel approach that increases feature correspondences and optimizes hybrid terms to eliminate misalignment. The method also mitigates projection distortion and perspective distortion, achieving accurate alignment and reduced distortion.
Article
Engineering, Electrical & Electronic
Wanli Xue, Weilun Xie, Yao Zhang, Shengyong Chen
Summary: Parallax tolerance is a challenging issue in image stitching, and addressing spatial variations, seam cutting, and consistency of homographies is crucial. To tackle these challenges, a stable image stitching framework is proposed in this study, involving a uniform linear structure model and a stable hybrid actor-critic algorithm to reduce ghosting and preserve structure.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2022)
Article
Computer Science, Information Systems
Shaoping Wen, Xiaolei Wang, Weichao Zhang, Guanjun Wang, Mengxing Huang, Benguo Yu
Summary: In this paper, a new hybrid warping model based on multi-homography and structure preservation is proposed for accurate alignment of regions at different depths while preserving image structures. Experimental results show that the proposed algorithm presents more accurate stitching results for images with large parallax and preserves salient image structures, outperforming existing methods both qualitatively and quantitatively.
Article
Engineering, Electrical & Electronic
Su Yan, Mingxi Ma, Jinliang Wang, Jun Zhang
Summary: In this paper, a parallax image stitching algorithm combining improved feature optimization with an innovative iterative seam estimation is proposed. Point features and line features of input images are detected and optimized. The mesh warp is optimized by minimizing a sparse quadratic total energy function to achieve accurate alignment in the overlapping regions. An iterative seam estimation method using an improved quality evaluation strategy is also proposed. Experimental results demonstrate that the proposed method achieves higher matching accuracy and visually better image stitching performance than other methods.
JOURNAL OF ELECTRONIC IMAGING
(2023)
Article
Engineering, Electrical & Electronic
Chang-Hsun Chiang, Jih-Gau Juang
Summary: This study utilizes unmanned aerial vehicles (UAVs) for riverbank inspection, capturing images and identifying garbage with algorithms. The system can replace traditional inspection labor and improve efficiency.
Article
Computer Science, Information Systems
Taeha Kim, Seongyeop Yang, Byeongkeun Kang, Heekyung Lee, Jeongil Seo, Yeejin Lee
Summary: A novel segmentation-based seam cutting algorithm was proposed to efficiently generate visually plausible high-resolution 360-degree videos, addressing issues such as distortion, temporal inconsistency, and time constraints. Experimental results demonstrated that the algorithm can accurately find seam boundaries, avoiding objects and improving temporary consistency.
Article
Computer Science, Software Engineering
Jiongli Gao, Jun Wu, Xuemei Zhao, Gang Xu
Summary: This paper presents a novel approach for natural stitching multiple images by integrating thin-plate spline (TPS), cylindrical projection, and plumb-line constraint. The proposed approach effectively transforms each image under plumb-line constraint and applies cylindrical projection to accurately align the transformed images. Experiments show that it outperforms other existing methods in achieving natural stitching.
Article
Green & Sustainable Science & Technology
Runmin Zhao, Yi Huang, Haoyuan Luo, Xiaoming Huang, Yangzezhi Zheng
Summary: This paper investigates the use of unmanned aerial vehicles (UAVs) for road damage detection and provides a theoretical and technical basis for UAV road damage detection by determining the recommended flight and camera parameters. It also proposes a UAV control framework for road damage detection, which includes automatic route planning and control, continuous photography control, and image stitching and smoothing tasks. Experimental results show that the road stitched long image obtained by the proposed method is applicable for machine learning pavement disease identification.
Article
Environmental Sciences
Liyao Song, Haiwei Li, Tieqiao Chen, Junyu Chen, Song Liu, Jiancun Fan, Quan Wang
Summary: The study proposed an integrated solution for acquiring and correcting UAV-borne hyperspectral images, including radiation and geometric correction, image registration, and multi-strip stitching. By introducing an improved method, the accuracy of radiation correction was enhanced, and the error of large-scale image stitching was controlled.
Article
Geochemistry & Geophysics
Zongyi Peng, Yong Ma, Xiaoguang Mei, Jun Huang, Fan Fan
Summary: A novel HSI stitching strategy was proposed using optimal seamline detection approach, selecting the optimal seamline with structural similarity (SSIM) index, which outperformed traditional methods in both spatial and spectral performances.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)