Case 3: Construction of ecological corridor in NDVI restoration site based on regression prediction

Whether returning farmland to forest or the damaged environment needs to be restored, the original vegetation community will be restored. Therefore, we should consider whether the relationship between the influence area and its surrounding environment is established in the restoration process. Since the corresponding value can be predicted according to the explanatory variables provided by regression prediction, can the unaffected environment around the restored land be used as the explanation variables? How about the prediction after the regression model is established by using the data from the regions with better environment? This case exploration is such a experience.

Technology

1 Prediction data processing

Green line: surrounding area of the restoration site; Orange Line: restoration area image (after basic processing of landsat8 Oli)

2 Training data processing

3 Training model and preservation

4 Loading model and prediction

The establishment of the model is decided by the training data set. This did not delete the ecological perfect area, but directly used the remote sensing image data of Yunnan Guizhou region. In the follow-up exploration, we will further explore how to determine the training data, and try to give the evaluation method according to the landscape index.