Image spectral/radiometry assessment for Land Parcel Identification System (LPIS) applications

Background

With the ongoing prevalence of digital sensors, both airborne and space-borne imaging techniques allow efficient identification of land cover. Land cover identification and its quantification is an important issue for the European Common Agriculture Policy (CAP), as the aids for European farmers are area-based. The payments are channeled through implementation of the Land Parcel Identification System (LPIS). In order to demonstrate and improve the quality of the LPIS system, JRC has implemented LPIS Quality Assurance (LPISQA) framework in 2010. Within this framework, orthoimages are used as the main basis of the inspection procedures (including the land cover identification step). The land cover identification step in LPISQA is performed by visual photointerpretation and the assessment of object significance in LPISQA is represented through parcel delineation and eligible land attribution. The outcome of this process depends highly on photo interpretation correctness and on the orthoimage quality.

State of play

Analysis from the first two years of the LPISQA implementation revealed sub-optimal quality of some orthoimage data. Determination of the orthoimage radiometric aspects is crucial for the LPISQA data observation methodology, as data objectivity and their comparison potential depend, among others, on the orthoimage radiometric quality. In contrast to the geometric orthoimage quality, required for LPISQA applications, not much is known about the radiometric orthoimage quality that is needed. Therefore, the study focuses on orthoimage radiometric quality attributes, such as orthoimage brightness, contrast, color composite, noise, histogram peak and distribution, shadows on the orthoimage and file formats.

Objectives

Taking into consideration: changing requirements resulting from the future post-2013 CAP; trends towards more preventive updates and controls; assumption that images taken from airborne or satellite platforms share the same technical properties and considering that these can be specified qualitatively and quantitatively with a common set of elements, the objectives of this research are:

• to identify issues and problems related to the image quality encountered in image-based parcel inspections;
• to design metrics and thresholds for the orthoimage radiometric quality in the context of the LPIS;
• to perform the test of the designed quality metrics against the commonly available datasets in a case study;
• to validate the usability of the metrics.
  

Goal

Successful achievement of the project’s objectives will produce harmonized, standardized package of quality metrics and thresholds for assessing the radiometric (understood as non-geometric) aspects of an orthoimage, applicable for correct land cover feature capturing through human photointerpretation.