gaozhangyang/Precipitation_Estimation

Infrared Precipitation Estimation

1. Framework

2. Dataset

region	scope
RegC	30N-45N, 90W-105W
RegW	35N-40N, 110W-115W
RegE	35N-40N, 80W-85W

training set	evaluating set	testing set
RegC, June to July 2012	RegC, August,2012
		RegC,June to August, 2014 RegC,December 2012 to February 2013
		RegW and RegE, June to August 2012

3. Usage

1. train

   python autotrain.py

2. evaluate

   cd Tools
   python generate_mp4.py

4. Performence

supplementary materials of iden experiments

Name(iden,val)	Acc0	Acc1
001(NR/R=200000/50000)	0.8898	0.8950
002(200000/100000)	0.8894	0.9022
003(340000/340000)	0.8635	0.9297
004(400000/200000)	0.8887	0.9027
005(250000/50000)	0.9335	0.8044
006(300000/50000)	0.9389	0.7840
007(500000/100000)	0.9320	0.8092

supplementary materials of esti experiments

Name(esti,val)	CC	BIAS	MSE	Acc0	Acc1
001(huber 2.5)	0.3579	0.5256	0.5683	0.9336	0.8044
002(huber 5)	0.3627	0.7709	0.5934	0.9336	0.8044
003(huber 7.5)	0.3634	0.8883	0.6126	0.9337	0.8041
004(huber 10)	0.3625	0.9633	0.6277	0.9339	0.8032
005(huber 12.5)	0.3621	1.027	0.6420	0.9339	0.8028
006(huber 15)	0.3621	1.060	0.6501	0.9341	0.8008
007(huber 20)	0.3616	1.0858	0.6604	0.9344	0.7994
008(huber 25)	0.3590	1.1206	0.6732	0.9352	0.7920
009(huber 12.5+KL)	0.3043	0.2022	0.7038	0.9508	0.6528
010(huber20+KL)	0.2928	0.2763	0.7715	0.9464	0.6994
011(huber25+KL)	0.2241	0.6564	1.091	0.9463	0.6875
012(huber 2.5+KL)	0.2684	0.2143	0.8625	0.9494	0.6877
013(huber 5+KL)	0.2919	0.2316	0.6972	0.9429	0.7326
014(huber 7.5+KL)	0.2447	0.3118	0.8018	0.9494	0.6736
015(huber 10+KL)	0.2974	0.2311	0.7527	0.9458	0.7113

5. citation

{
    %0 Journal Article
    %T Infrared Precipitation Estimation Using Convolutional Neural Network
    %P 1-14
    %U https://ieeexplore.ieee.org/document/9085928/
    %G en
    %J IEEE Transactions on Geoscience and Remote Sensing
    %A Wang, Cunguang
    %A Xu, Jing
    %A Tang, Guoqiang
    %A Yang, Yi
    %A Hong, Yang
    %D 2020
}