使用 python 从图像中提取与给定中心成特定角度的特定大小的补丁

Question

对于我的项目，我想使用 python 从尺寸为 1024x720 的图像中以特定角度提取尺寸为 224x224 的补丁。图像上补丁中心的像素位置已给出，补丁所成的角度也已给出。我知道如何使用阵列切片以 0（度）角提取补丁，但我想以一定角度切片图像。任何有关相同的帮助将不胜感激。谢谢！

Answer 1

from scipy import ndimage
import numpy as np
import math as m
import cv2   
 def patchmaker(img,height,width,center_y,center_x,angle):
        theta = angle/180*3.14
        img_shape = np.shape(img)
        print(img_shape)
        x = [[i for i in range(0,img_shape[1])] for y in range(img_shape[0])]
        y = [[j for i in range(img_shape[1])] for j in range(0,img_shape[0])]
        x = np.asarray(x)
        y = np.asarray(y)
        rotatex = x[center_y-m.floor(height/2):center_y+m.floor(height/2),center_x-m.floor(width/2):center_x+m.floor(width/2)]
        rotatey = y[center_y-m.floor(height/2):center_y+m.floor(height/2),center_x-m.floor(width/2):center_x+m.floor(width/2)]
        coords   = [rotatex.reshape((1,height*width))-center_x,rotatey.reshape((1,height*width))-center_y]
        coords = np.asarray(coords)
        coords = coords.reshape(2,height*width)
        roatemat = [[m.cos(theta),m.sin(theta)],[-m.sin(theta),m.cos(theta)]]
        rotatedcoords = np.matmul(roatemat,coords)
        patch = ndimage.map_coordinates(img,[rotatedcoords[1]+center_y,rotatedcoords[0]+center_x], order=1, mode='nearest').reshape(height,width)
        return patch

使用 python 从图像中提取与给定中心成特定角度的特定大小的补丁

extracting patch of a certain size at certain angle with given center from an image using python

python

opencv

image

angle

slice