Validating a Sample Collection
Your sample collection will now appear on the Dataset Details page. You can then add a new validation session based on the sample collection you created. Click on Add a validation session to get started. This will take you to the Create your validation exercise screen (Figure 9).
Enter a name for the validation session along with a description, and then choose a validation method. The following validation methods are currently supported:
Blind: The user has no information about the classification of the sample and therefore needs to provide a new interpretation based on a pre-defined list of classes.
Plausibility: The user has information about the classification of the sample and evaluates the plausibility of the thematic classification with yes or no.
Enhanced Plausibility: The user has information about the classification of the sample, can answer yes and no, and if no (i.e., incorrect), they can correct the classification based on a pre-defined list of classes.
You can also choose your validation platform. There are two supported validation platforms available:
Web: The validation process will be carried out inside a web browser using a number of base imagery data sets.
Mobile: The validation process will be carried out using the LACO-Wiki mobile smartphone app on the ground. The mobile app can be downloaded from the Google Playstore and the Apple AppStore.
After choosing web or mobile, you can select some additional validation settings shown in Figure 10 such as the option to add a comment field, geo-tagged photographs from Flick and NDVI (normalized difference vegetation indices) profiles to the validation session.
Finally, you can set the minimum and maximum zoom level, the sample navigation (feature extent or fixed zoom level) and define the base layers and overlays (Figure 11).
When you have selected all the features, click on the Create Validation Session button to start validating your sample collection (Figure 12).
The process of validating your samples can be paused at any time and then continued when required. See the example in Figure 12 where the session was paused.
LACO-Wiki also allows you to split the validation workload by sharing the current validation session with other users. These users can then contribute to validating your data set.
The actual validation is done by classifying the object (point, polygon or raster grid cell) using one of the legend items. Figure 13 shows an example of blind validation of a point object. In this case you would classify the object by choosing one of the legend items on the right.
Last updated
