全景图 与 6面图转换

目录

全景图转6面图：

6面图转全景图

---

全景图转6面图：

https://github.com/springcheese/panoramic_to_cubemap_generation

# Necessary Imports
import math
import argparse
import numpy as np
from PIL import Image# Dictionary for CUBEMAP FACES
CUBEMAP_FACES = {'back': 0, 'left': 1, 'front': 2, 'right': 3, 'top' : 4, 'bottom' : 5 }# Dimensions of the input image
INPUT_WIDTH = 2048
INPUT_HEIGHT = 1024def output_image_to_xyz(i, j, cube_face_id, edge_length):"""Get x,y,z coordinates from the output image pixel coordinates (i,j) taking into account different faces of the cube map:param i: Image row Pixel Coordinate:param j: Image column Pixel Coordinate:param cube_face_id: The face Id of the Cube Map:param edge_length: The size of the face:return: Tuple (x,y,z)"""a = 2.0 * float(i) / edge_lengthb = 2.0 * float(j) / edge_lengthif cube_face_id == CUBEMAP_FACES['back']:(x,y,z) = (-1.0, 1.0 - a, 1.0 - b)elif cube_face_id == CUBEMAP_FACES['left']:(x,y,z) = (a - 1.0, -1.0, 1.0 - b)elif cube_face_id == CUBEMAP_FACES['front']:(x,y,z) = (1.0, a - 1.0, 1.0 - b)elif cube_face_id == CUBEMAP_FACES['right']:(x,y,z) = (1.0 - a, 1.0, 1.0 - b)elif cube_face_id == CUBEMAP_FACES['top']:(x,y,z) = (b - 1.0, a - 1.0, 1.0)elif cube_face_id == CUBEMAP_FACES['bottom']:(x,y,z) = (1.0 - b, a - 1.0, -1.0)return (x, y, z)# -------------------------------------------------------
def generate_cubemap_face(input_image, output_image, cube_face_id):"""Generate a face of the Cube Map from the Input Image:param input_image: Input Image:param output_image: Output Image Handle as the blank image:param cube_face_id: Face ID to construct:return:Implicit in the Output Image"""input_size = input_image.sizeoutput_size = output_image.sizeinput_pixels = input_image.load()output_pixels = output_image.load()edge_length = output_size[0]for x_out in range(edge_length):for y_out in range(edge_length):(x,y,z) = output_image_to_xyz(x_out, y_out, cube_face_id, edge_length)theta = math.atan2(y,x)r = math.hypot(x,y)phi = math.atan2(z,r)# Source Image Coordinatesuf = 0.5 * input_size[0] * (theta + math.pi) / math.pivf = 0.5 * input_size[0] * (math.pi/2 - phi) / math.pi# Bilinear Interpolation amongst the four surrounding pixelsui = math.floor(uf)  # bottom leftvi = math.floor(vf)u2 = ui+1            # top rightv2 = vi+1mu = uf-uinu = vf-vi# Pixel values of four cornersA = input_pixels[ui % input_size[0], np.clip(vi, 0, input_size[1]-1)]B = input_pixels[u2 % input_size[0], np.clip(vi, 0, input_size[1]-1)]C = input_pixels[ui % input_size[0], np.clip(v2, 0, input_size[1]-1)]D = input_pixels[u2 % input_size[0], np.clip(v2, 0, input_size[1]-1)]# interpolate(r,g,b) = (A[0]*(1-mu)*(1-nu) + B[0]*(mu)*(1-nu) + C[0]*(1-mu)*nu+D[0]*mu*nu,A[1]*(1-mu)*(1-nu) + B[1]*(mu)*(1-nu) + C[1]*(1-mu)*nu+D[1]*mu*nu,A[2]*(1-mu)*(1-nu) + B[2]*(mu)*(1-nu) + C[2]*(1-mu)*nu+D[2]*mu*nu )output_pixels[x_out, y_out] = (int(round(r)), int(round(g)), int(round(b)))# -------------------------------------------------------
def generate_cubemap_outputs(input_image_path):"""Generate the cubemap faces output from the input image path(Directly save the output in the same path with suffixes for cubemap faces):param input_image_path: Input Image:return:Implicit in the Output Image Saving Process"""print ('---------------------------------------------------------------------------')print ('Generating Cubemap Faces  ------------ ' + input_image_path)# Read the input imageinput_image = Image.open(input_image_path)input_image = input_image.resize((int(INPUT_WIDTH), int(INPUT_HEIGHT)))input_size = input_image.size# Define the edge length - This will be the width and height of each cubemap faceedge_length = input_size[0] // 4# Generate cubemap for the left faceout_image = Image.new("RGB", (edge_length, edge_length), "black")generate_cubemap_face(input_image, out_image, CUBEMAP_FACES['left'])out_image.save(input_image_path[0:-4] + '_cubemap_left.png')print ('LEFT face done')# Generate cubemap for the front faceout_image = Image.new("RGB", (edge_length, edge_length), "black")generate_cubemap_face(input_image, out_image, CUBEMAP_FACES['front'])out_image.save(input_image_path[0:-4] + '_cubemap_front.png')print ('FRONT face done')# Generate cubemap for the right faceout_image = Image.new("RGB", (edge_length, edge_length), "black")generate_cubemap_face(input_image, out_image, CUBEMAP_FACES['right'])out_image.save(input_image_path[0:-4] + '_cubemap_right.png')print ('RIGHT face done')# Generate cubemap for the back faceout_image = Image.new("RGB", (edge_length, edge_length), "black")generate_cubemap_face(input_image, out_image, CUBEMAP_FACES['back'])out_image.save(input_image_path[0:-4] + '_cubemap_back.png')print ('BACK face done')# Generate cubemap for the top faceout_image = Image.new("RGB", (edge_length, edge_length), "black")generate_cubemap_face(input_image, out_image, CUBEMAP_FACES['top'])out_image.save(input_image_path[0:-4] + '_cubemap_top.png')print ('TOP face done')# Generate cubemap for the bottom faceout_image = Image.new("RGB", (edge_length, edge_length), "black")generate_cubemap_face(input_image, out_image, CUBEMAP_FACES['bottom'])out_image.save(input_image_path[0:-4] + '_cubemap_bottom.png')print ('BOTTOM face done')print ('---------------------------------------------------------------------------')def main():parser = argparse.ArgumentParser()parser.add_argument("--input", help = "Enter the path of the input photosphere image (Resizes to INPUT_WIDTH x INPUT_HEIGHT for speed)",type=str, default='')args = parser.parse_args()input_image_path = r"D:\360极速浏览器X下载\517c9510-c456-481d-aaa8-78e6a7c2098b.png"input_image_path = r"E:\project\jijia\angle_360\demo\ComfyUI_00041_.png"input_image_path = r"E:\project\jijia\angle_360\demo\quan\ComfyUI_00043_.png"generate_cubemap_outputs(input_image_path)if __name__ == "__main__":main()

6面图转全景图

GitHub - eddyteddy3/CubeMap-To-Panorama: This code will transform 6-cubemap images to panorama using Py

import os
import numpy as np
from PIL import Image
from scipy.ndimage import map_coordinates# ---------- Image Utility ----------"""-> 6-images"""def image_list(images):assert len(images) == 6image_np_array = []for image in images:image_np_array.append(np.array(image))concatenated_image = np.concatenate(image_np_array, axis=1)# horizontal to diceassert concatenated_image.shape[0] * 6 == concatenated_image.shape[1]w = concatenated_image.shape[0]cube_dice = np.zeros((w * 3, w * 4, concatenated_image.shape[2]), dtype=concatenated_image.dtype)# cube_list = cube_h2list(concatenated_image)cube_list = np.split(concatenated_image, 6, axis=1)# Order: F R B L U Dsxy = [(1, 1), (2, 1), (3, 1), (0, 1), (1, 0), (1, 2)]for i, (sx, sy) in enumerate(sxy):face = cube_list[i]if i in [1, 2]:face = np.flip(face, axis=1)if i == 4:face = np.flip(face, axis=0)cube_dice[sy*w:(sy+1)*w, sx*w:(sx+1)*w] = face# dice to horizontalw = cube_dice.shape[0] // 3assert cube_dice.shape[0] == w * 3 and cube_dice.shape[1] == w * 4cube_h = np.zeros((w, w * 6, cube_dice.shape[2]), dtype=cube_dice.dtype)# Order: F R B L U Dsxy = [(1, 1), (2, 1), (3, 1), (0, 1), (1, 0), (1, 2)]for i, (sx, sy) in enumerate(sxy):face = cube_dice[sy*w:(sy+1)*w, sx*w:(sx+1)*w]# if i in [1, 2]:#     face = np.flip(face, axis=1)# if i == 4:#     face = np.flip(face, axis=0)cube_h[:, i*w:(i+1)*w] = facefixed_horizontal_images = cube_hreturn fixed_horizontal_images# -------------- Main Logic --------------"""generates the grid of u,v coordinates from 
u = -pi to pi 
v = pi/2 to -pi/2Mathematical notation. You can relate it with Longitude and Latitudeu ∈ [−π,π] ,v ∈ [ π/2, - π/2]"""def equirect_uvgrid(h, w):u = np.linspace(-np.pi, np.pi, num=w, dtype=np.float32)v = np.linspace(np.pi, -np.pi, num=h, dtype=np.float32) / 2return np.stack(np.meshgrid(u, v), axis=-1)"""creates a mapping for cube imagessuch that it represents which cube map maps to which equirectangle projection.
also determines the boundaries"""def equirect_facetype(h, w):'''0F 1R 2B 3L 4U 5D'''tp = np.roll(np.arange(4).repeat(w // 4)[None, :].repeat(h, 0), 3 * w // 8, 1)# Prepare ceil maskmask = np.zeros((h, w // 4), bool)idx = np.linspace(-np.pi, np.pi, w // 4) / 4idx = h // 2 - np.round(np.arctan(np.cos(idx)) * h / np.pi).astype(int)for i, j in enumerate(idx):mask[:j, i] = 1"""{x, y, z}-X        +x{-------|-------}"""mask = np.roll(np.concatenate([mask] * 4, 1), 3 * w // 8, 1)tp[mask] = 4tp[np.flip(mask, 0)] = 5return tp.astype(np.int32)"""applies the transformation to spherical coordinates, we acquired previously.
The padding is to handle the edges."""def sample_cubefaces(cube_faces, tp, coor_y, coor_x, order):cube_faces = cube_faces.copy()cube_faces[1] = np.flip(cube_faces[1], 1)cube_faces[2] = np.flip(cube_faces[2], 1)cube_faces[4] = np.flip(cube_faces[4], 0)# Pad up downpad_ud = np.zeros((6, 2, cube_faces.shape[2]))pad_ud[0, 0] = cube_faces[5, 0, :]pad_ud[0, 1] = cube_faces[4, -1, :]pad_ud[1, 0] = cube_faces[5, :, -1]pad_ud[1, 1] = cube_faces[4, ::-1, -1]pad_ud[2, 0] = cube_faces[5, -1, ::-1]pad_ud[2, 1] = cube_faces[4, 0, ::-1]pad_ud[3, 0] = cube_faces[5, ::-1, 0]pad_ud[3, 1] = cube_faces[4, :, 0]pad_ud[4, 0] = cube_faces[0, 0, :]pad_ud[4, 1] = cube_faces[2, 0, ::-1]pad_ud[5, 0] = cube_faces[2, -1, ::-1]pad_ud[5, 1] = cube_faces[0, -1, :]cube_faces = np.concatenate([cube_faces, pad_ud], 1)# Pad left rightpad_lr = np.zeros((6, cube_faces.shape[1], 2))pad_lr[0, :, 0] = cube_faces[1, :, 0]pad_lr[0, :, 1] = cube_faces[3, :, -1]pad_lr[1, :, 0] = cube_faces[2, :, 0]pad_lr[1, :, 1] = cube_faces[0, :, -1]pad_lr[2, :, 0] = cube_faces[3, :, 0]pad_lr[2, :, 1] = cube_faces[1, :, -1]pad_lr[3, :, 0] = cube_faces[0, :, 0]pad_lr[3, :, 1] = cube_faces[2, :, -1]pad_lr[4, 1:-1, 0] = cube_faces[1, 0, ::-1]pad_lr[4, 1:-1, 1] = cube_faces[3, 0, :]pad_lr[5, 1:-1, 0] = cube_faces[1, -2, :]pad_lr[5, 1:-1, 1] = cube_faces[3, -2, ::-1]cube_faces = np.concatenate([cube_faces, pad_lr], 2)return map_coordinates(cube_faces, [tp, coor_y, coor_x], order=order, mode='wrap')"""Main Function to convert the cube maps to equirectangle image- Calculates the U,V with helper function
- Creates Mapping
- Normalize the coordinates
- Reconstruction of image"""def cube_to_equirectanlge(cubemap, h, w, mode='bilinear'):if mode == 'bilinear':order = 1elif mode == 'nearest':order = 0else:raise NotImplementedError('unknown mode')assert len(cubemap.shape) == 3assert cubemap.shape[0] * 6 == cubemap.shape[1]assert w % 8 == 0face_w = cubemap.shape[0]uv = equirect_uvgrid(h, w)u, v = np.split(uv, 2, axis=-1)u = u[..., 0]v = v[..., 0]cube_faces = np.stack(np.split(cubemap, 6, 1), 0)# Get face id to each pixel: 0F 1R 2B 3L 4U 5Dtp = equirect_facetype(h, w)coor_x = np.zeros((h, w))coor_y = np.zeros((h, w))for i in range(4):mask = (tp == i)coor_x[mask] = 0.5 * np.tan(u[mask] - np.pi * i / 2)coor_y[mask] = -0.5 * np.tan(v[mask]) / np.cos(u[mask] - np.pi * i / 2)mask = (tp == 4)c = 0.5 * np.tan(np.pi / 2 - v[mask])coor_x[mask] = c * np.sin(u[mask])coor_y[mask] = c * np.cos(u[mask])mask = (tp == 5)c = 0.5 * np.tan(np.pi / 2 - np.abs(v[mask]))coor_x[mask] = c * np.sin(u[mask])coor_y[mask] = -c * np.cos(u[mask])# Final renormalizecoor_x = (np.clip(coor_x, -0.5, 0.5) + 0.5) * face_wcoor_y = (np.clip(coor_y, -0.5, 0.5) + 0.5) * face_wequirec = np.stack([sample_cubefaces(cube_faces[..., i], tp, coor_y, coor_x, order=order)for i in range(cube_faces.shape[3])], axis=-1)return equirec# ------- Load Images ------fixed_image = image_list([Image.open('front.jpeg'),  # FImage.open('right.jpeg'),  # RImage.open('back.jpeg'),   # BImage.open('left.jpeg'),   # LImage.open('up.jpeg'),     # UImage.open('down.jpeg')    # D]
)pano_image = cube_to_equirectanlge(fixed_image, 1024, 1280, mode='nearest')# img = Image.fromarray(pano_image.astype(np.uint8), 'RGB')# img = Image.fromarray(fixed_image, 'RGB')
# img.show()def generate_folder_info(root_folder):folder_info = []for subdir, _, files in os.walk(root_folder):image_files = [f for f in files if f.lower().endswith('jpg')]if len(image_files) == 6:relative_path = os.path.relpath(subdir, root_folder)parts = relative_path.split(os.sep)# print("parts ", parts)if len(parts) > 2:level1 = parts[0]level2 = parts[1]level3 = parts[2]folder_info.append({"level1": level1,"level2": level2,"level3": level3})elif len(parts) == 2:level1 = parts[0]level2 = parts[1]folder_info.append({"level1": level1, "level2": level2})return folder_info# Step 1: Read all the directories {Level1: {id of classified}, level2: 'roomName'}
folder_infos = generate_folder_info("image_set")#Step 
for folder_info in folder_infos:level1 = folder_info['level1']level2 = folder_info['level2']level_heirarchy = f"{level1}/{level2}"pathToImageFolder = f"image_set/{level_heirarchy}"if 'level3' in folder_info:level3 = folder_info['level3']print("more than 3 levels", f"image_set/{level_heirarchy}/{level3}")pathToImageFolder = f"image_set/{level_heirarchy}/{level3}""""4 x 2pano_width = 1024 * 4 {pano_f.width}pano_height = 1024 * 2 {pano_f.height}actual_size = 4096 x 2048"""print("path: ", pathToImageFolder)temp_image = np.array(Image.open(f"{pathToImageFolder}/pano_f.jpg"))fixed_image = image_list([Image.open(f"{pathToImageFolder}/pano_f.jpg"), #Image.open('front.jpeg'),  # FImage.open(f"{pathToImageFolder}/pano_r.jpg"), #Image.open('right.jpeg'),  # RImage.open(f"{pathToImageFolder}/pano_b.jpg"), #Image.open('back.jpeg'),   # BImage.open(f"{pathToImageFolder}/pano_l.jpg"), #Image.open('left.jpeg'),   # LImage.open(f"{pathToImageFolder}/pano_u.jpg"), #Image.open('up.jpeg'),     # UImage.open(f"{pathToImageFolder}/pano_d.jpg") #Image.open('down.jpeg')    # D])fixed_width = temp_image.shape[1] * 4fixed_height = temp_image.shape[0] * 2pano_image = cube_to_equirectanlge(fixed_image, fixed_height, fixed_width, mode='nearest')img = Image.fromarray(pano_image.astype(np.uint8), 'RGB')print("sucking it bitch!")target = f"panorama_images/{level1}/"if not os.path.exists(target):os.makedirs(target)img.save(f"{target}/{level2}.jpg")print("sucked it bitch!")"""{x, y} STITCHED IMAGE -> Sphere {U,V} {Longitude & Latitude}folder_info -> [list folders] -> [6 images]image_set/{id}/room{something}/nameOfImage.jpg"""""""6 - cube image to Pano: cubemap-to-pano[6 cube maps] -> image_list -> one_stiched_image (NOT PANO)
one_stiched_image -> cube_to_equirectanlge -> Panorama Imagescript to scrap images: scrap-to-CubeMap
(some-very-smart-function) -> [6-cube-map-image]Proposal 1:
scrap-to-CubeMap -> [6-set-images] -> cubemap-to-pano -> Panorama ImageProposal 2: (STATIC & LOCAL) (WE ARE MOVING FORWRAD WITH THIS)
Aleady downloaded pictures
[DIRECTORY] -> [List of folders] -> [nested list of each classifieds] -> [nested list of each room] -> [6-cube-map] -> cubemap-to-pano -> Panorama ImageProposal 3: (Scrapping Dynamic) (BACKEND SIDE)
scrap-to-CubeMap -> [6-set-images] -> cubemap-to-pano -> Panorama Image"""