update README and data

.gitignore

@@ -1,17 +1,9 @@
 .idea/
-
 __pycache__/
 
-
+vflood/
 logs/
-
-records/cp_WaterNet.pth.tar
-output/
-output2/
-overlay/
 records/
+MeshTransformer/
 
-video_module/logs/
-image_module/WaterSegModels/
-
-env/
+output/

README.md

@@ -2,17 +2,24 @@
 
 This is an official PyTorch implementation for paper "V-FloodNet: A Video Segmentation System for Urban Flood Detection and Quantification".
 
-## Environments
+## 1 Environments
+
+### 1.1 Code and packages
 We developed and tested the source code under Ubuntu 18.04 and PyTorch framework. 
 The following packages are required to run the code.
 
-First, a python virtual environment is recommended. 
+First, git clone this repository
+```bash
+git clone https://github.com/xmlyqing00/V-FloodNet.git
+```
+
+Second, a python virtual environment is recommended. 
 I use `pip` to create a virtual environment named `env` and activate it.
 Then, recursively pull the submodules code.
 
 ```shell
-python3 -m venv env
-source env/bin/activate
+python3 -m venv vflood
+source vflood/bin/activate
 git submodule update --init --recursive
 ```
 
@@ -23,7 +30,7 @@ In the virtual environment, install the following required packages from their o
 - [Detectron2](https://github.com/facebookresearch/detectron2) for reference objects segmentation.
 - [MeshTransformer](https://github.com/microsoft/MeshTransformer) for human detection and 3D mesh alignment.
 
-We provide the corresponding installation command here
+We provide the corresponding installation command here, you can replace the version number that fit your environment.
 
 ```shell
 pip install torch==1.8.2+cu111 torchvision==0.9.2+cu111 torchaudio==0.8.2 -f https://download.pytorch.org/whl/lts/1.8/torch_lts.html
@@ -39,27 +46,56 @@ Then install the rest packages indicated in `requirements.txt`
 pip install -r requirements.txt
 ```
 
-## Usage
+### 1.2 Pretrained Models
 
-Download and extract the pretrained weights, and put them in the folder `./records/`. Weights and groundtruths are stored in [Google Drive](https://drive.google.com/file/d/1r0YmT24t4uMwi4xtSLXD5jyaMIuMzorS/view?usp=sharing).
+First, run the following script to download the pretrained models of MeshTransformer
+```bash
+sh scripts/download_MeshTransformer_models.sh
+```
 
-### Water Image Segmentation
-Put the testing images in `image_folder`, then
+Second, download SMPL model `mpips_smplify_public_v2.zip` from the official website [SMPLify](http://smplify.is.tue.mpg.de/). Extract it and place the model file `basicModel_neutral_lbs_10_207_0_v1.0.0.pkl` at `./MeshTransformer/metro/modeling/data`.
+<!-- - Download `MANO_RIGHT.pkl` from [MANO](https://mano.is.tue.mpg.de/), and place it at `${REPO_DIR}/metro/modeling/data`. -->
+<!-- 
+```
+${REPO_DIR}  
+|-- metro  
+|   |-- modeling
+|   |   |-- data
+|   |   |   |-- basicModel_neutral_lbs_10_207_0_v1.0.0.pkl
+|   |   |   |-- MANO_RIGHT.pkl
+|-- models
+|-- datasets
+|-- predictions
+|-- README.md 
+|-- ... 
+|-- ... 
+``` -->
+<!-- Please check [/metro/modeling/data/README.md](../metro/modeling/data/README.md) for further details. -->
+
+Third, download the archives from [Google Drive](https://drive.google.com/drive/folders/1DURwcb_qhBeWYznTrpJ-7yGJTHxm7pxC?usp=sharing).
+Extract the pretrained models for water segmentation `records.zip` and put them in the folder `./records/`. 
+Extract the water dataset `WaterDataset` in any path, which includes the training images and testing videos.
+
+
+## 2 Usage
+
+### 2.1 Water Image Segmentation
+Put the testing images in a folder then
 ```shell
 python test_image_seg.py \
     --test_path=/path/to/image_folder --test_name=<test_name>
 ```
 The default output folder is `output/segs/`
 
-### Water Video Segmentation
+### 2.2 Water Video Segmentation
 If your input is a video, we provide a script `scripts/cvt_video_to_imgs.py` to extract frames of the video.
-Put the extracted frames in `frame_folder`, then
+Put the extracted frames in a folder then
 ```shell
 python test_video_seg.py \
     --test-path=/path/to/frame_folder --test-name=<test_name>
 ```
 
-### Water Depth Estimation
+### 2.3 Water Depth Estimation
 
 We provide three options `stopsign`, `people`, and `ref` for `--opt` to specify three types reference objects.
 ```shell
@@ -71,5 +107,5 @@ For input video, to compare the estimated water level with the groundtruths in `
 python cmp_hydrograph.py --test-name=<test_name>
 ```
 
-## Copyright
+## 3 Copyright
 This paper is submitted to Elsevier Journal Computers, Environment and Urban Systems under review. The corresponding author is Xin Li (Xin Li <[email protected]>).  All rights are reserved.

est_waterlevel.py

@@ -20,7 +20,7 @@ def get_parser():
     parser.add_argument('--out-dir', default='output/waterlevel',
                         help='A file or directory to save output results.')
     parser.add_argument('--opt', type=str,
-                        help='Estimation options.')
+                        help='Estimation options. "people", "stopsign", or "ref"')
 
     return parser.parse_args()
 
@@ -33,7 +33,7 @@ def main(args):
     out_dir = os.path.join(args.out_dir, f'{args.test_name}_{args.opt}')
     os.makedirs(out_dir, exist_ok=True)
 
-    if args.opt in ['skeleton', 'stopsign']:
+    if args.opt in ['people', 'stopsign']:
         est_by_obj_detection(img_list, water_mask_list, out_dir, args.opt)
     elif args.opt == 'ref':
         est_by_reference(img_list, water_mask_list, out_dir, args.test_name)