Ground Control Points Selection and Collection for Image Orthorectification and for Orthoimages Accuracy Validation
ISBN 978-85-88783-11-9
Authors
1Coelho, A.; 2Melo, M.; 3Pina, W.; 4Silveira, J.
1IBGE Email: aline.lopes@ibge.gov.br
2IBGE Email: marilea.melo@ibge.gov.br
3IBGE Email: walter.pina@ibge.gov.br
4IBGE Email: juliane.silveira@ibge.gov.br
Abstract
Ground control points are used in the orthorectification of remote sensing imagery and as check points for validation and quality assessment. Fieldwork is usually the most expensive step in the process of validation and orthorectification of imagery. For this reason, it is crucial to carefully choose, in the images, points that can be correctly identified in the field, as well as to accurately measure and process its coordinates. Poor GCP selection, and also problems in locating the points on the ground or in processing their coordinates, can severely compromise the image rectification and quality assessment procedures for orthoimages generation. During the planning phase, it is necessary to define the number and distribution of GCPs, the survey method, the GPS equipment and the crew to be used. The number of GCPs must be carefully defined, taking into account the characteristics of both the sensor and the area of interest. With respect to image rectification, the number of GCPs will also depend on the expected accuracy for the orthoimage to be generated as well as the tools and mathematical model available with the software used. When it comes to quality assessment of orthoimages, the number of GCPs also depends on the maximum acceptable error for the orthoimages. Regarding the sensor characteristics , the higher its spatial resolution the more options are available for identifying good targets on the image - and easier it is to identify them on the field too. Recent acquisition dates provide a positive contribution to the fieldwork. The smaller the time difference between image acquisition and fieldwork, the greater the possibility of finding the planned GCPs exactly as they are seen in the image. On the other hand, the better the target discrimination capability of the sensor, the more accurate the GCP coordinates must be. This will determine the survey method, which is also influenced by the type of GPS receiver available. Regarding the GCPs distribution, the distance between them should be calculated taking into account either the survey specifications or the project quality assessment. In practice, this distance cannot be fixed, and circular regions within which the GCPs should be chosen are selected. The identification and selection of visible points in the image that appear to be good candidates for control points or field validation,when made at the office, has been more valuable than selecting the point directly while onsite, as it facilitates planning the optimal access route and decreases the time spent in the field, thereby improving productivity and reducing costs. These points should be chosen over targets perfectly identifiable both in the image and on the ground. Moreover, priority should be given to targets that are easy to reach, with careful attention to possible obstacles that could impact the ability of the field crew to reach the selected spot. Targets that could have changed between the date of acquisition and the date in which the survey takes place should also be discarded. This article presents suggestions on targets that usually work well and targets that should be avoided. In order to ensure that the surveyed GCPs can be unequivocally identified in the image, and in that way, correctly used in the rectification or validation processes, it is necessary to document the point, collecting information that can help the team in the office - such as location sketches and photos of the GCP spot that include the survey equipment being used, among others.
Keywords
Ground control points; Image Accuracy Validation; Quality Assessment