Integrated visual quality assessment for ZiYuan-3 optical satellite panchromatic products

ABSTRACT

In practical data production process of ZiYuan-3 (ZY-3) optical satellite, the quality of massive panchromatic (PAN) products is usually measured with multiple quality metrics. Although the existing metrics have been widely used in practice and obtained good performance, they have some limitations: (1) there are so many quality metrics that makes it difficult for users or operators to directly judge whether the imagery is acceptable or not; and (2) a specific quality metric can only measure a certain aspect of image quality and is often not designed from the perspective of human visual system (HVS), leading the objective evaluation result inconsistent with subjective one. To tackle the aforementioned problems, we propose an integrated visual quality assessment (VQA) method to predict comprehensive quality scores for ZY-3 sensor calibration (SC) PAN products. In the proposed method: (1) we exploited eight quality elements that have significant influences on the visual quality of SC PAN products; (2) we constructed a database composed of 360 ZY-3 SC PAN images and the corresponding subjective mean opinion scores (MOS); (3) we introduced generalised regression neural network to combine the extracted quality elements of the images and their MOS and obtained the integrated VQA result. Experimental results on the database showed that the proposed method achieved high accuracy of predicted quality scores and well consistency with HVS, indicating the effectiveness and reliability of the presented approach.

KEYWORDS:

• ZiYuan-3 optical satellite
• panchromatic data
• image quality assessment
• quality elements
• GRNN

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported in part by the Special Project on the Integration of Industry, Education and Research of Guangdong Province of China (2012A090300017), the National Natural Science Foundation of China (61601335), and the Hubei Provincial Natural Science Foundation (2016CFB157).