直方图反向投影
1、先把RGB色彩空间转换为HSV色彩空间
2、计算样本的2D直方图
3、对样本直方图作归一化
4、对目标图像作反向投影
import cv2 as cv
import numpy as np
from matplotlib import pyplot as plt
def back_project_demo():
sample = cv.imread('C:/Users/Y/Pictures/Saved Pictures/lena.png')
target = cv.imread('C:/Users/Y/Pictures/Saved Pictures/lenanoise.png')
roi_hsv = cv.cvtColor(sample, cv.COLOR_BGR2HSV)
target_hsv = cv.cvtColor(target, cv.COLOR_BGR2HSV)
# show iamges
cv.imshow('sample', sample)
cv.imshow('target', target)
# 计算样本的2D直方图，[0, 1]表示2个通道，[180，256]分别表示两个通道的范围，[0,180,0,256]是固定的值，不用改
# 是HSV两个通道的取值范围
# [180,256]这两个值的数值越大，对像素的划分越精细，图像看起来越不准确，可适当将数值改小
roiHist = cv.calcHist([roi_hsv], [0, 1], None, [180, 256], [0, 180, 0, 256])
# 根据MINMAX作归一化
cv.normalize(roiHist, roiHist, 0, 255, cv.NORM_MINMAX)
# 反向投影
dst = cv.calcBackProject([target_hsv], [0, 1], roiHist, [0, 180, 0, 256], 1)
cv.imshow('backProjectionDemo', dst)
# 建立2D直方图
def hist2d_demo(image):
hsv = cv.cvtColor(image, cv.COLOR_BGR2HSV)
# mask是None，两个通道的histSize分别是180和256，H范围是0-180，S范围是0-256,
hist = cv.calcHist([image], [0, 1], None, [32, 32], [0, 180, 0, 256])
# cv.imshow('hist2d', hist)
plt.imshow(hist, interpolation='nearest')
plt.title('2D Histogram')
plt.show()
src = cv.imread('C:/Users/Y/Pictures/Saved Pictures/demo.png')
cv.namedWindow('input image', cv.WINDOW_AUTOSIZE)
cv.imshow('input image', src)
hist2d_demo(src)
back_project_demo(src)
cv.waitKey(0)
cv.destroyAllWindows()
原图像 原图相对应的2D直方图
sample图像 target图像 对target图像做反向投影