Reading List

Contents

1. Atmosphric Sciences

2. Hydrology

3. Remote Sensing

4. Computer Vision

4.1 Machine Learning Basics

4.2 Object Detection

4.3 ML/DL for Weather Prediction

4.4 ML/DL for Rainfall Estimation

4.5 Generative Adversaril Network

5. Numerical Models

5.1 Optical Flow

5.2 Semi-Lagrangian Scheme

6. Statistics

7. Updates

Atmospheric Sciences

1. Atmospheric Science by JONH M. WALLACE and PETER V. HOBBS
2. Python Gallery for meterology
3. A Review of Global Precipitation Datasets: Data Sources, Estimation, and Intercomparisons

Remote Sensing

1. Hydrologic Remote Sensing: Capacity Building for Sustainability and Resilience

Hydrology (notes)

1. 23 Unsolved problems in hydrology UPH a community perspective

Computer Vision

Machine Learning Basics

1. 神经网络与深度学习
2. Standford CS231: Convolutional Neural Networks for Visual Recognition
3. Learning kernels for CV 3.1 Orientation Filter: Gabor Filter Interesting intro; opencv
4. Collection of Pytorch Lists

Object Detection/Tracking Generalized

1. Gaussian Mixture Models (GMM)

ML/DL for weather prediction (notes)

1. Generating Videos with Scene Dynamics/GAN
2. A Predictive Neural Network for Learning Higher-Order Non-Stationarity from Spatiotemporal Dynamics
3. Deep Learning for Precipitation Nowcasting: A benchmark and A new model
4. Temporal Convolutional Neural Network for the Classification of Satellite Image Time Series
5. NCAR DL for atmospheric sciences summer school playlist

ML/DL for rainfall estimation (notes)

1. Video-based rainfall removal

   1.1 Video Rain Streak Removal By Multiscale ConvolutionalSparse Coding

   1.2 Is it Raining Outside?Detection of Rainfall using General-Purpose Surveillance Cameras | code

2. Image-based rainfall removal

   2.1 JORDEN: Deep Joint Rain Detection from a Single Image

   2.2 Rain Streaks Detection and Removal in Image based on Entropy Maximization and Background Estimation

   2.3 Spatial Attentive Single-Image Deraining with a High Quality Real Rain dataset

   2.4 Dynamic Routing Residue Recurrent Network for Video Removal

   2.5 Progressive Image Deraining Networks: A Better and Simpler Baseline

Generative Adversaril Network(GAN) (notes)

1. Pytorch-GAN-zoo
2. InfoGAN
3. tempoGAN
4. WGAN
5. WGAN-improved | github
6. DehazeGAN | code
7. Glow | Blog | github

Numerical Methods

Optical Flow

Methods:

Lucas-Kanade method: it tracks the corner with Shi-Tomasi algorithm and calculate (u,v) by solving 9 equations and then estimate the 3x3 patch movement;

cv2.calcOpticalFlowPyrLK

Gunner Farneback's algorithm (Dense): It computes the optical flow for all the points in the frame;

cv2.calcOpticalFlowFarneback

1. Optical flow models as an open benchmark for radar-based precipitation nowcasting

   Codes available on github: optical flow

2. Novel Video Prediction for Large-scale Scene using Optical Flow
3. Two Frame Motion Estimation Based on Polynomial Expansion
4. my radar project demo
5. FlowNet: Learning Optical Flow with Convolutional Networks

Semi-Lagrangian Scheme

1. My radar project demo

Statistics

1. Probabilistic Programming and Bayesian Methods for Hackers

Updates

• [x] Migrate to library
• [x] Update hydrology (2019.6.16)
• [x] Update GAN collections and rainfall removal category(2019.6.2)
• [x] FlowNet: Learning Optical Flow with Convolutional Networks(2019.5.23)
• [x] 神经网络与深度学习 (2019.5.22)
• [x] optical flow models as an open benchmark for radar-based precipitation nowcasting (2019.5.16)
• [x] Novel Video Prediction for Large-scale Scene using Optical Flow (2019.5.16)
• [x] Generating Videos with Scene Dynamics/GAN (2019.5.16)
• [ ] Two Frame Motion Estimation Based on Polynomial Expansion