remote sensing image segmentation based on deep learning method, data fusion, multi modle

AI remote sensing image semantic segmentation requires scientific research project experience, multiple remote sensing data fusion semantic segmentation, identification of different forest types, the data set already exists, the label is also available, but the number of labels is not many, the establishment of a deep learning algorithm (photographic machine or CNN non- Supervised recognition) 1. Identify different types of forests, visualize the recognition results, and calculate the area of the output type. 2. According to the classification of the forest type, calculate the corresponding type of carbon sink. The traditional method is relatively mature. If you want to use deep learning to do it, you will use the prediction Training model, etc.


Have experience in image segmentation and CNN algorithm development for image roll machine Note: 1. mIOU-0.70, the data has no error, and the data after recognition should have no overlap and no contact. 2. Provide source code, operating environment, remote environment deployment. 3. Later assistance needs to be adjusted and modified