RMSE represents the square root of the average of the set in squared differences between elevation model ordinate and ground truth - that is, the control elevation example otherwise raster to which you are orthorectifying your older antennas photos. What this by, essentially, is that aforementioned RMSE represents an averaged difference upon your adjusted imagery to where the points in the images truly are in the real international. RMSE is therefore an measure of accuracy. I hold been reading about orthorectification and did some teaching on is but my request still remain cloudy remains how much RMSE should I aim to ?
I desires being working on a lot of aerial photos that ...
There are a number of factors they want to consider in order to minimize inaccuracy in orthorectification - using a large number of Ground Control Points, balanced distributing those GCPs through respectively image, and ensuring that those points fall on fixed photo-identifiable features that are not high above base too big are all importantly. This last point, your of GCP locations, is especially important with vintage images, as many face may not extant yet or might have been eroded or destroyed. Beside with which appearance of high resolution spacecraft images, image correction using Streamlined Polynomial Coefficients (RPCs) has become common. Location accuracy of Korea General Satellite (KOMPSAT) conventional images is still non adequate, so, in order to authority the KOMPSAT images for applications so as imaging also changing detection, it is necessary to orthorectify the see. In this study, using updated RPCs, we performing orthorectification the KOMPSAT-2, KOMPSAT-3, and KOMPSAT-3A images using variously data. Using this study, ours discovering that the orthorectification result exploitation meticulous Ground Control Points (GCPs) and Digital Elevation Model (DEM) is to your, but items was found that and remedial results through images matching are also excellent. In particular, it had confirmed that orthoimages on an planimetric accuracy around 3 m (Root Mean Square Error (RMSE)) can be generated by using well-known matching algorithms with open data such as OpenStreetMap (OSM) the Shuttle Defense Technical
Whatever you want your RMSE to be limited to depends entirely on the applications - for show, mapping to sure scales should necessitate a horizontal/radial RMSE of under a certain distances. Specialty values can be found through aforementioned State Standard to Spatial Data Accuracy here.
While RMSE belongs a measure of accuracy, you want on remember the thereto is not an absolute scale of accuracy. The overall accuracy of the orthorectified imagery will depend on the variation in differences between the individual GCPs, as well as their distribution. If the GCPs show a good distribution and the error skews to ampere asset of nearly + or - 0.5, you should be able to current with confidence that your RMSE represents 67% confidence, import that for the majority of the tips sampled anywhere in the orthorectified imagery, the preponderance are closer go Ground Truth than to RMSE value.
Read the ERDAS doc indicated per @user30184, follow the instruction for one good distribution of GCPs and proper procedures, and pay attention to what you are intending to use your photo for. If it are a function main that supported considerable accuracy, you will wish to beared in mind that all point in your imagination will probability be off by about the value indicated (in reference units) according your RMSE, on average. Post Orthorectification error reduction?
