utils/camera_utils.py

#
# Copyright (C) 2023, Inria
# GRAPHDECO research group, https://team.inria.fr/graphdeco
# All rights reserved.
#
# This software is free for non-commercial, research and evaluation use 
# under the terms of the LICENSE.md file.
#
# For inquiries contact  george.drettakis@inria.fr
#

from scene.cameras import Camera
import numpy as np
from utils.general_utils import PILtoTorch
from utils.graphics_utils import fov2focal

WARNED = False

def loadCam(args, id, cam_info, resolution_scale):

    image_rgb = PILtoTorch(cam_info.image).type("torch.ByteTensor")
    background = PILtoTorch(cam_info.background)[:3, ...].type("torch.ByteTensor")

    gt_image = image_rgb[:3, ...]
    loaded_mask = None

    return Camera(colmap_id=cam_info.uid, R=cam_info.R, T=cam_info.T, 
                  FoVx=cam_info.FovX, FoVy=cam_info.FovY, 
                  image=gt_image, gt_alpha_mask=loaded_mask, background=background, talking_dict=cam_info.talking_dict,
                  image_name=cam_info.image_name, uid=id, data_device=args.data_device)

def cameraList_from_camInfos(cam_infos, resolution_scale, args):
    camera_list = []

    for id, c in enumerate(cam_infos):
        camera_list.append(loadCam(args, id, c, resolution_scale))

    return camera_list

def camera_to_JSON(id, camera : Camera):
    Rt = np.zeros((4, 4))
    Rt[:3, :3] = camera.R.transpose()
    Rt[:3, 3] = camera.T
    Rt[3, 3] = 1.0

    W2C = np.linalg.inv(Rt)
    pos = W2C[:3, 3]
    rot = W2C[:3, :3]
    serializable_array_2d = [x.tolist() for x in rot]
    camera_entry = {
        'id' : id,
        'img_name' : camera.image_name,
        'width' : camera.width,
        'height' : camera.height,
        'position': pos.tolist(),
        'rotation': serializable_array_2d,
        'fy' : fov2focal(camera.FovY, camera.height),
        'fx' : fov2focal(camera.FovX, camera.width)
    }
    return camera_entry