#include <iostream>
#include <opencv2/imgcodecs.hpp>
#include <opencv2/highgui.hpp>
#include <opencv2/imgproc.hpp>
#include <vector>



void showImg(const std::string& windowName,const cv::Mat& img){
	cv::imshow(windowName,img);
}


void getBackground(const cv::Mat& source,cv::Mat& dst) {


	cv::dilate(source,dst,cv::Mat::ones(3,3,CV_8U)); //Kernel 3x3
} 

void getForeground(const cv::Mat& source,cv::Mat& dst) {

	cv::distanceTransform(source,dst,cv::DIST_L2,3,CV_32F);
	cv::normalize(dst, dst, 0, 1, cv::NORM_MINMAX);
}

void findMarker(const cv::Mat& sureBg,cv::Mat& markers,
							std::vector<std::vector<cv::Point>>& contours)
{
	cv::findContours(sureBg,contours,cv::RETR_EXTERNAL,
															cv::CHAIN_APPROX_SIMPLE);

	// Draw the foreground markers
	for (size_t i = 0,size = contours.size(); i < size; i++)
			drawContours(markers, contours, static_cast<int>(i),
								cv::Scalar(static_cast<int>(i)+1), -1);
}


void getRandomColor(std::vector<cv::Vec3b>& colors,size_t size)
{
	for (int i = 0; i < size ; ++i)
	{
			int b = cv::theRNG().uniform(0, 256);
			int g = cv::theRNG().uniform(0, 256);
			int r = cv::theRNG().uniform(0, 256);
			colors.emplace_back(cv::Vec3b((uchar)b, (uchar)g, (uchar)r));
	}
}


int main (int argc,char** argv) {
// 讀取圖片
	if(argc < 2)
	{
		std::cerr << "Error\n";
		std::cerr << "Provide Input Image:\n<program> <inputimage>\n";
		return -1;
	}
	cv::Mat original_img = cv::imread(argv[1]);

	if(original_img.empty())
	{
		std::cerr << "Error\n";
		std::cerr << "Cannot Read Image\n";
		return -1;
	}
// 去除圖像中的噪聲， 均值偏移模糊（MeanShift）是一種圖像邊緣保留濾波算法，常用于圖像分水嶺分割前的去噪，可顯著提高分水嶺分割效果。
	cv::Mat shifted;
	cv::pyrMeanShiftFiltering(original_img,shifted,21,51);
	showImg("Mean Shifted",shifted);
// 將原始圖像轉(zhuǎn)換為灰度和二值圖像
	cv::Mat gray_img;
	cv::cvtColor(original_img,gray_img,cv::COLOR_BGR2GRAY);
	showImg("GrayIMg",gray_img);

	cv::Mat bin_img;
	cv::threshold(gray_img,bin_img,0,255,cv::THRESH_BINARY | cv::THRESH_OTSU);
	showImg("thres img",bin_img);
// 尋找確定的背景圖像， 在這一步中，我們找到圖像中的背景區(qū)域。
	cv::Mat sure_bg;
	getBackground(bin_img,sure_bg);
	showImg("Sure Background",sure_bg);
// 找到確定前景的圖像， 對于圖像的前景，我們采用距離變換算法
	cv::Mat sure_fg;
	getForeground(bin_img,sure_fg);
	showImg("Sure ForeGround",sure_fg);
// 找到標記，在應(yīng)用分水嶺算法之前，我們需要標記。為此，我們將使用opencv中提供的findContour()函數(shù)來查找圖像中的標記。
	cv::Mat markers = cv::Mat::zeros(sure_bg.size(),CV_32S);
	std::vector<std::vector<cv::Point>> contours;
	findMarker(sure_bg,markers,contours);
	cv::circle(markers, cv::Point(5, 5), 3, cv::Scalar(255), -1); //Drawing Circle around the marker
// 應(yīng)用分水嶺算法
	cv::watershed(original_img,markers);

	cv::Mat mark;
	markers.convertTo(mark, CV_8U);
	cv::bitwise_not(mark, mark); //黑變白，白變黑
	showImg("MARKER",mark);
//高亮顯示圖像中的標記
	std::vector<cv::Vec3b> colors;
	getRandomColor(colors,contours.size());

	//構(gòu)建結(jié)果圖像
	cv::Mat dst = cv::Mat::zeros(markers.size(), CV_8UC3);
	// 用隨機的顏色填充已標記的物體
	for (int i = 0; i < markers.rows; i++)
	{
			for (int j = 0; j < markers.cols; j++)
			{
					int index = markers.at<int>(i,j);
					if (index > 0 && index <= static_cast<int>(contours.size()))
							dst.at<cv::Vec3b>(i,j) = colors[index-1];
			}
	}

	showImg("Final Result",dst);

	cv::waitKey(0);
	return 0;
}

import cv2 as cv
import numpy as np
import argparse
import random as rng
rng.seed(12345)
parser = argparse.ArgumentParser(description='Code for Image Segmentation with Distance Transform and Watershed Algorithm.\
    Sample code showing how to segment overlapping objects using Laplacian filtering, \
    in addition to Watershed and Distance Transformation')
parser.add_argument('--input', help='Path to input image.', default='HFOUG.jpg')
args = parser.parse_args()
src = cv.imread(cv.samples.findFile(args.input))
if src is None:
    print('Could not open or find the image:', args.input)
    exit(0)
# Show source image
cv.imshow('Source Image', src)
cv.waitKey()

gray = cv.cvtColor(src, cv.COLOR_BGR2GRAY)
ret, thresh = cv.threshold(gray, 0, 255, cv.THRESH_BINARY + cv.THRESH_OTSU)
# noise removal
kernel = np.ones((5, 5), np.uint8)
opening = cv.morphologyEx(thresh, cv.MORPH_OPEN, kernel, iterations=2)

# 獲取背景圖
sure_bg = opening.copy()  # 背景
# Show output image
cv.imshow('Black Background Image', sure_bg)  # 黑色是背景
cv.waitKey()

# 獲取前景圖
dist = cv.distanceTransform(opening, cv.DIST_L2, 3)
# Normalize the distance image for range = {0.0, 1.0}
# so we can visualize and threshold it
cv.normalize(dist, dist, 0, 1.0, cv.NORM_MINMAX)
cv.imshow('Distance Transform Image', dist)
_, dist = cv.threshold(dist, 0.2, 1.0, cv.THRESH_BINARY)
# Dilate a bit the dist image
kernel1 = np.ones((3, 3), dtype=np.uint8)
dist = cv.dilate(dist, kernel1)
cv.imshow('Peaks', dist)

# 構(gòu)建初始markers
dist_8u = dist.astype('uint8')
# Find total markers
contours, _ = cv.findContours(dist_8u, cv.RETR_EXTERNAL, cv.CHAIN_APPROX_SIMPLE)
# 創(chuàng)建即將應(yīng)用分水嶺算法的標記圖像
markers = np.zeros(dist.shape, dtype=np.int32)
# 標記前景
for i in range(len(contours)):
    cv.drawContours(markers, contours, i, (i + 1), -1)  # 輪廓標記從1開始
# 標記背景
cv.circle(markers, (5, 5), 3, 255, -1)  # 此處背景標記為255
print("before watershed: ", np.unique(markers))  # 0表示不確定標記區(qū)域
# 可視化markers
markers_8u = (markers * 10).astype('uint8')
cv.imshow('Markers', markers_8u)
cv.waitKey()

# 應(yīng)用分水嶺分割算法
markers = cv.watershed(src, markers)
print("after watershed: ", np.unique(markers))  # -1表示邊界

# mark = np.zeros(markers.shape, dtype=np.uint8)
mark = markers.astype('uint8')
mark = cv.bitwise_not(mark)
# uncomment this if you want to see how the mark
# image looks like at that point
# cv.imshow('Markers_v2', mark)
# Generate random colors
colors = []
for contour in contours:
    colors.append((rng.randint(0, 256), rng.randint(0, 256), rng.randint(0, 256)))
# Create the result image
dst = np.zeros((markers.shape[0], markers.shape[1], 3), dtype=np.uint8)
# Fill labeled objects with random colors
for i in range(markers.shape[0]):
    for j in range(markers.shape[1]):
        index = markers[i, j]
        if index > 0 and index <= len(contours):  # -1表示邊界， 255表示背景
            dst[i, j, :] = colors[index - 1]
# Visualize the final image
cv.imshow('Final Result', dst)
cv.waitKey()

import cv2 as cv
import numpy as np
import argparse
import random as rng
def process_img2(img):
    img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    img_gray = cv2.GaussianBlur(img_gray, (5, 5), 0.1)
    img_gray = cv2.medianBlur(img_gray, 5)
    _, image_binary = cv2.threshold(img_gray, 0, 255, cv2.THRESH_OTSU + cv2.THRESH_BINARY)

    kernel = np.ones((7, 7), np.uint8)
    # sure_bg = cv.morphologyEx(image_binary, cv.MORPH_CLOSE, kernel, iterations=3)
    sure_bg = cv.dilate(image_binary, kernel, iterations=2)
    sure_bg = cv.bitwise_not(sure_bg)

    element = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (3, 3))
    image_binary = cv2.morphologyEx(image_binary, cv2.MORPH_OPEN, element)

    imageSC = cv2.distanceTransform(image_binary, cv2.DIST_L2, 5)
    imageSC = imageSC.astype(np.uint8)
    imageSC = cv2.normalize(imageSC, 0, 255, cv2.NORM_MINMAX)
    _, imageSC = cv2.threshold(imageSC, 0, 255, cv2.THRESH_OTSU + cv2.THRESH_BINARY)
    return imageSC, sure_bg

rng.seed(12345)
imgPath = "HFOUG.jpg"
src = cv.imread(imgPath)
shifted = cv2.pyrMeanShiftFiltering(src, 7, 15)
if src is None:
    print('Could not open or find the image:')
    # print('Could not open or find the image:', args.input)
    exit(0)
# Show source image
cv.imshow('Source Image', src)
cv.waitKey()
opening, sure_bg = process_img2(shifted)
# Show output image
cv.imshow('Background Image', sure_bg)  # 背景
cv.waitKey()

# 獲取前景圖
dist = cv.distanceTransform(opening, cv.DIST_L2, 3)
# Normalize the distance image for range = {0.0, 1.0}
# so we can visualize and threshold it
cv.normalize(dist, dist, 0, 1.0, cv.NORM_MINMAX)
cv.imshow('Distance Transform Image', dist)
_, dist = cv.threshold(dist, 0.3, 1.0, cv.THRESH_BINARY)
# Dilate a bit the dist image
kernel1 = np.ones((3, 3), dtype=np.uint8)
dist = cv.dilate(dist, kernel1)
cv.imshow('Peaks', dist)

# 構(gòu)建初始markers
dist_8u = dist.astype('uint8')
# Find total markers
contours, _ = cv.findContours(dist_8u, cv.RETR_EXTERNAL, cv.CHAIN_APPROX_SIMPLE)
# 創(chuàng)建即將應(yīng)用分水嶺算法的標記圖像
# markers = np.zeros(dist.shape, dtype=np.int32)
markers = sure_bg.copy().astype(np.int32)
# 標記前景
for i in range(len(contours)):
    cv.drawContours(markers, contours, i, (i + 1), -1)  # 輪廓標記從1開始
# 標記背景
# cv.circle(markers, (5, 5), 3, 255, -1)  # 此處背景標記為255
# 可視化markers

print("before watershed: ", np.unique(markers))  # 0表示不確定標記區(qū)域
markers_8u = (markers * 10).astype('uint8')
cv.imshow('Markers', markers_8u)

# 應(yīng)用分水嶺分割算法
markers = cv.watershed(src, markers)

print("after watershed: ", np.unique(markers))  # -1表示邊界

# mark = np.zeros(markers.shape, dtype=np.uint8)
mark = markers.astype('uint8')
mark = cv.bitwise_not(mark)

cv.imshow('Markers_v2', mark)
# Generate random colors
colors = []
for contour in contours:
    colors.append((rng.randint(0, 256), rng.randint(0, 256), rng.randint(0, 256)))
# Create the result image
dst = np.zeros((markers.shape[0], markers.shape[1], 3), dtype=np.uint8)
# Fill labeled objects with random colors
for i in range(markers.shape[0]):
    for j in range(markers.shape[1]):
        index = markers[i, j]
        if index > 0 and index <= len(contours):  # -1表示邊界， 255表示背景
            dst[i, j, :] = colors[index - 1]
# Visualize the final image
cv.imshow('Final Result', dst)
cv.waitKey(0)
cv2.destroyAllWindows()