/Imputing-spatial-PM2.5-values-via-image-inpainting-A-case-study-in-Taiwan

透過圖像修補技術補空間中的 PM2.5 值：以台灣 為例

Primary LanguagePython

Imputing-spatial-PM2.5-values-via-image-inpainting-A-case-study-in-Taiwan

透過圖像修補技術補空間中的 PM2.5 值：以台灣為例

How to run RFR

  • python run.py --model_path YOUR_MODEL_PATH_TO_LOAD --num_iters EPOCH_NUM --txt RESULT_FILE_NAME --batch_size YOUR_VATCH_SIZE --result_save_path IMAGE_SAVE_PATH --model_save_path WHERE_TO_SAVE_YOUR_MODEL
  • ADD -test IF YOU WANT RUN TEST DIRECTLY

OTHER SETTING

switch the dataset in run.py to change the random setting

  • ar -> random in all time
  • sr -> random in start
  • switch the RFRnet in model.py to change the batchnorm setting
  • bnoff -> cancel batchnorm

Data

data link

點資訊(M,D,H,Lon,Lat,PM2.5)

  • a_202101.csv
  • E_202101.csv

圖資訊

  • a_by_loc -> airbox 站點圖
  • by_loc_main -> EPA 站點圖
  • e+a_mask_gt_main -> EPA+airbox kriging結果
  • mask -> 台灣地區遮罩