Mining discriminative spatial cues for aerial image quality assessment towards big data

• Published in: Signal processing. Image communication Vol. 80; p. 115646
• Main Authors: Zhang, Xiaoci, Gu, Naijie, Chang, Jie, Ye, Hong, Lin, Chuanwen, Shen, Jie
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.02.2020; Elsevier BV
• Subjects: Airport construction; Airports; Algorithms; Artificial intelligent; Big data; Computer vision; Cues; Data mining; Image quality; Image quality assessment; Playgrounds; Quality assessment; Regression models; Satellite imagery; Support vector machines; Image quality assessment; Big data; Artificial intelligent; Data mining
• ISSN: 0923-5965; 1879-2677
• DOI: 10.1016/j.image.2019.115646

Evaluating massive-scale aerial/satellite images quality is useful in computer vision and intelligent applications. Traditional local features-based algorithms have achieved impressive performance. However, spatial cues, i.e., geometric property and topological structure, have not been exploited effectively and explicitly. Thus, in this paper, we propose a novel method for image quality assessment towards aerial/satellite images, where discriminative spatial cues are well encoded. More specifically, in order to mine inherent spatial structure of aerial images, each image is segmented into several basic components such as buildings, airport and playground. Afterwards, a weighted region adjacency graph (RAG) is built based on the basic components to represent the spatial feature of each aerial image. We integrate the spatial feature with other transform domain features, and train a support vector regression model to achieve image quality assessment. Experiments demonstrate that our method shows competitive or even better performance compared with several state-of-the-art algorithms. •Spatial characteristics are important feature of aerial images that are affectable with distortions.•Weighted region adjacent graph represents both the topologic structure and object interaction relationship within an image.•Integrating spatial features with transform domain features in image quality assessment improves the model performance.