Automatic Geometric Processing for Very High Resolution Optical Satellite Data Based on Vector Roads and Orthophotos

Peter Pehani,Aleš Marsetič,Klemen Čotar,Janez Zaletelj,Krištof Oštir

doi:10.3390/rs8040343

Abstract

In response to the increasing need for fast satellite image processing SPACE-SI developed STORM—a fully automatic image processing chain that performs all processing steps from the input optical images to web-delivered map-ready products for various sensors. This paper focuses on the automatic geometric corrections module and its adaptation to very high resolution (VHR) multispectral images. In the automatic ground control points (GCPs) extraction sub-module a two-step algorithm that utilizes vector roads as a reference layer and delivers GCPs for high resolution RapidEye images with near pixel accuracy was initially implemented. Super-fine positioning of individual GCPs onto an aerial orthophoto was introduced for VHR images. The enhanced algorithm is capable of achieving accuracy of approximately 1.5 pixels on WorldView-2 data. In the case of RapidEye images the accuracies of the physical sensor model reach sub-pixel values at independent check points. When compared to the reference national aerial orthophoto the accuracies of WorldView-2 orthoimages automatically produced with the rational function model reach near-pixel values. On a heterogeneous set of 41 RapidEye images the rate of automatic processing reached 97.6%. Image processing times remained under one hour for standard-size images of both sensor types.

Highlights

The ever increasing quantity of satellite data from large and small Earth observation satellites offers the potential for new and innovative applications and boosts the need for automatic and fast data processing
The largest image (“20.04.2011”), Sens. 2016, 8, 343 which is the size of three standard images, was processed in 1 h 42 min, which is still acceptable for extraordinary large large image.image
With the two-step algorithm delivering a sufficient number of ground control points (GCPs) with an average accuracy of 1.18 px on heterogeneous high resolution (HR) RapidEye images, we consider that we are getting close to the limitations of a fully automatic GCP extraction method

Summary

Introduction

The ever increasing quantity of satellite data from large and small Earth observation satellites offers the potential for new and innovative applications and boosts the need for automatic and fast data processing. Complex and slow data processing remains the main obstacle that prevents most end users from using satellite data. The image processing steps are usually known, they are neither automatic nor performed in real-time. Automatic and generic (i.e., adapted to different sensors) systems for processing high resolution (HR; from 10 m to 2 m) and very high resolution 2 m and below) optical images are predominantly demanded by the Earth observation community. All optical image processing—either automatic or manual—includes several pre-processing and product generation steps. The obligatory pre-processing steps are composed of geometric and radiometric corrections in which the pixels’ locations and values are corrected.

Results

Discussion

Conclusion