Interactive imagesThe land use map on the ground
The land use map on the ground
The land use map classified from the time series or the embeddings, as you choose: the probability of each class, the strongest class with and without smoothing, and a divider to compare the map with the image of the window and, on the embeddings, with the embeddings too.
Choose the time series or the embeddings, step through the probability of each class, switch smoothing on and off, show the pixels whose label changed, and swipe between the image and the map, or, on the embeddings, the embeddings too.
The map and its probabilities come from the input chosen here.
Loading rgb-months.png (1.4 MB).
Image: the window in May 2017, with SWIR16, NIR08 and RED as red, green and blue. Map: the land use map after smoothing, in the colours of the classes. Drag the divider, or focus it and use the arrow keys, to compare the two pictures.
The legend gives the share of the labelled pixels that each class holds, counted from this map. The shares add up to 100 %.
What to try
- One class at a time. Press One class and choose a class: white is a probability of 0, black of 1. Click a pixel, or use the arrow keys on the window, to read all its probabilities as bars.
- Switch smoothing on. The picture becomes the smoothed score of the class, and the marked pixels are those whose label smoothing changed. Where do they gather?
- The strongest class. Press Strongest class and switch smoothing off and on: the map from the raw probabilities against the map from the smoothed ones. Show the changed pixels to see where they differ.
- Follow a boundary on the ground. Press On the ground, choose Image | Map and move the divider slowly across a field: where does the map’s edge follow the field in the image, and where does it round a corner?
- From embeddings to classes. This one needs the embeddings: choose Embeddings above the figure, then Embeddings | Map, and drag the divider: where do the map’s classes follow the regions of the embeddings’ colours, and where do they not?