Radiometric calibration, validation and correction of multispectral photogrammetric imagery

Lauri Markelin

    Research output: ThesisDoctoral ThesisCollection of Articles

    Abstract

    Vast amounts of remote sensing data are acquire daily all over the globe from satellites, from manned or unmanned airborne platforms, and from the ground. Airborne photogrammetry provides a flexible method for acquiring high-resolution imagery in a timely manner over large areas. Aerial images are increasingly being used in a more automatic and quantitative way for applications such as land cover classification and environmental monitoring. Apart from the high geometric quality of photogrammetric sensors, also their radiometric properties are important. Different objects reflect solar irradiance according to their individual spectral and directional properties, and radiometric analysis can be used to identify such objects and changes in them. The perquisite for quantitative radiometry is the absolute radiometric calibration of the sensor, which links the recorded digital numbers to physical units. The major benefit of a radiometrically calibrated sensor is the possibility to radiometrically correct images form atmospheric effects to surface reflectance. Radiometric correction becomes a necessity, when imagery from different dates and sensors are used for quantitative image analysis. The objectives of this study were, first, to develop a vicarious method for the radiometric calibration and validation (Cal/Val) of a photogrammetric sensor in a test field. Second, three radiometric correction methods suitable for reflectance image product generation from photogrammetric images were evaluated. Finally, the influence of the solar elevation angle in the radiometric performance of multispectral photogrammetry was evaluated. The Cal/Val method developed in this study utilizes field measured nadir reflectance factors of the reference targets to match the reflectance factors measured at a laboratory in an exact imaging geometry to the current weather conditions. When evaluating the radiometric correction methods, a reflectance accuracy level of 5 % was achievable with all of the evaluated methods when using well-defined isotropic reference targets. For other targets, reflectance accuracies of between 5 and 20 % were possible. The results showed that a low solar elevation of 25° did not cause the general performance of the photogrammetric processes and 3D point cloud generation to deteriorate. The radiometric Cal/Val method presented in this study presents a step towards developing traceable processes for photogrammetric sensors. The results also confirmed the high radiometric quality of photogrammetric sensors and proved the suitability of the photogrammetric imagery for radiometric correction. This makes possible the rigorous radiometric processing of photogrammetric images and improves the quality and accuracy of automatic image interpretation and classification tasks.
    Translated title of the contributionFotogrammetristen ilmakuvien radiometrinen kalibrointi, arviointi ja korjaus
    Original languageEnglish
    QualificationDoctor's degree
    Awarding Institution
    • Aalto University
    Supervisors/Advisors
    • Haggren, Henrik, Supervising Professor
    • Honkavaara, Eija, Thesis Advisor, External person
    Print ISBNs978-951-711-294-9
    Electronic ISBNs978-951-711-295-6
    Publication statusPublished - 2013
    MoE publication typeG5 Doctoral dissertation (article)

    Keywords

    • photogrammetry
    • radiometric calibration
    • radiometric correction
    • airborne imaging
    • multispectral
    • radiative transfer modelling

    Fingerprint Dive into the research topics of 'Radiometric calibration, validation and correction of multispectral photogrammetric imagery'. Together they form a unique fingerprint.

    Cite this