Abstract
With finer spatial scale, high-resolution images provide complex, spatial, and massive information on the earth’s surface, which brings new challenges to remote sensing segmentation methods. In view of these challenges, finding a more effective segmentation model and parallel processing method is crucial to improve the segmentation accuracy and process efficiency of large-scale high-resolution images. To this end, this study proposed a minimum spanning tree (MST) model integrated into a regional-based parallel segmentation method. First, an image was decomposed into several blocks by regular tessellation. The corresponding homogeneous regions were obtained using the minimum heterogeneity rule (MHR) partitioning technique in a multicore parallel processing mode, and the initial segmentation results were obtained by the parallel block merging method. On this basis, a regionalized fuzzy c-means (FCM) method based on master-slave parallel mode was proposed to achieve fast and optimal segmentation. The proposed segmentation approach was tested on high-resolution images. The results from the qualitative assessment, quantitative evaluation, and parallel analysis verified the feasibility and validity of the proposed method.
Highlights
Remote sensing image segmentation is a process of partitioning image domain into several meaningful regions and serves as a bridge between remote sensing image and high-level information processing [1,2]
For a WorldView-3 Pan-Sharpened multiband remote sensing image with 8192 × 8192 pixels and 0.5-m spatial resolution, the running time and speedup curve for the parallel minimum heterogeneity rule (MHR) partitioning, parallel regional hidden Markov random field (RHMRF)-fuzzy c-means (FCM), and the proposed method were examined using four sockets server with 72 Cores
This study proposed an efficient parallel high-resolution image segmentation method based on the minimum spanning tree (MST) and RHMRF-FCM model
Summary
Remote sensing image segmentation is a process of partitioning image domain into several meaningful regions and serves as a bridge between remote sensing image and high-level information processing [1,2]. Remote sensing image is a type of data that records spectral information of the earth’s surface from a spaceborne sensor. With a finer spatial scale, high-resolution images provide more detailed information through abundant geometric, textural, and spectral features, resulting in new challenges to the traditional remote sensing image segmentation methods [6,7]. These challenges include geometric noise caused by peripheral objects, the phenomena of different objects having the same spectral signature, and similar object types shown as having varying spectral signatures in high-resolution images [8]. The abundant spatial and geometric information, which determines the spatial and geometric models, must be taken into account in building the segmentation model
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have