So, i had this very weird bug while treating my dem data lately. I used SAGA to create flows channel networks on a dem i acquired a little while ago. Preprocessed it using sink route detection and filling the sinks, then used the "channel network and drainage basins" tool. Everything worked perfectly, i got the data i needed on the area i'm working on.
Then 2 weeks ago, i get some new data covering the same area but with a higher resolution, went from pixels of 25x25 meters to pixels of 5 by 5. The data here was way bigger, but i used the same tools and preprocess. It was long but i got my preprocessed data, and after that everytime i tried using the channel network tool SAGA crashes at the "edge detection" part. Tried using the fill sinks xxl for preprocess, didn't work. Used tiles to divide my data and then applied the process to the samples, worked but what i want is doing this on the whole area in order to have only one vector file with everything i need and the right Strahler orders. Filling in the gaps manually, considering the size of the data and the area afflicted by the discontinuities caused by the tiles, is clearly out of the question.
The original data is over 1 go, i think that might be the issue and what bothers me is that i can't find a fix anywhere, did anyone run into the same issue? (From what i gathered the crash might be related to the ram SAGA is trying to use for the process, attributing to it more ressources than what the computer has.)
EDIT: I did find a way to make the data a little easier to complete and fill by making sure the gaps were as little as possible. For this i used the dem with the 25x25 pixels and extracted the watershed basins from it, then i used thoses to divide my high resolution dem into big basins that i treated on their own afterwards. Stacked the whole vector datas i obtained, added a few lines here and there to make sure that the main lines don't have any gaps and done. It is a lot better than before but obtaining the right Strahler orders and the real basins seems still a little tricky to obtain with thoses. That and using this process actualy erased some of the smaller lines on every edge of thetheorical basins i created.