/tensorflow-calculate-SVF

Using PSPNet-Keras-tensorflow to calculate SVF(sky view factor) and TVF(tree view factor).

Primary LanguagePythonMIT LicenseMIT

tensorflow-calculate-SVF

Using PSPNet-Keras-tensorflow to calculate SVF(sky view factor) and TVF(tree view factor).

Usage:

  1. Install dependencies: pip install -r requirements.txt --upgrade

  2. download pspnet50_ade20k.h5, pspnet50_ade20k.json, this is the weight I used in this project. (for more information please refer to this link.)

  3. Go to the project directory, create a folder 'imgs'(or whatever you like) and put your panorama pictures in it.

  4. Run python autoCallpspnet.py imgs, the program identifies sky/tree/building elements in the pictures and colors them differently. The recognized pictures are automaticlly saved in 'imgs-recognized'.

origin
recognized

  1. Run python PanoramicToFisheye.py imgs-recognized. This will transfer the panorama pictures to fisheye pictures in order to calculate SVF and TVF. The fisheye pictures are automaticlly saved in 'imgs-fisheye'.

recognized
fisheye

  1. Run python calculateResult.py imgs-fisheye. This will calculate SVF and TVF results.