程序员经验分享-hwhale

如何修改蒙版使其成为完美的圆形

2023-12-25

我有这样的不完美圆形面具。如何使用 opencv 轮廓函数（或任何其他方式）去除右上角的伪影？

这是数据：

mask = np.array([
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0,0,0,0,255,255,255],
[0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0,0,255,255,255],
[0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,255,255,255,255],
[0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,255,255,255,255],
[0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,255,255,255,255,255],
[0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,255,255,255,255,255],
[0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255],
[0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255],
[0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255],
[0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255],
[0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255],
[0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255],
[0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,255,255],
[0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0],
[0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0],
[0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0],
[0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0],
[0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0],
[0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0],
[0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0],
[0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0],
[0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0],
[0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0],
[0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0],
[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]], dtype=np.uint8)

一种方法是大津的门槛 https://opencv-python-tutroals.readthedocs.io/en/latest/py_tutorials/py_imgproc/py_thresholding/py_thresholding.html#otsus-binarization图像以获得二值图像。从这里开始，我们执行形态开放 https://opencv-python-tutroals.readthedocs.io/en/latest/py_tutorials/py_imgproc/py_morphological_ops/py_morphological_ops.html#opening with a 椭圆形核 https://opencv-python-tutroals.readthedocs.io/en/latest/py_tutorials/py_imgproc/py_morphological_ops/py_morphological_ops.html#structuring-element。此步骤将有效地消除额外的伪影，但会稍微扭曲圆圈。为了修复圆，我们找到轮廓并使用cv2.minEnclosingCircle() https://docs.opencv.org/2.4/modules/imgproc/doc/structural_analysis_and_shape_descriptors.html?highlight=minenclosingcircle#minenclosingcircle然后将其绘制到蒙版上以获得完美的圆形。

以下是每个步骤的可视化：

我截取了你的图像，没有网格线。输入图像：

获得二值图像的大津阈值

具有椭圆形内核的变形开口

结果来自cv2.minEnclosingCircle()并将所得轮廓绘制到掩模上

Code

import cv2
import numpy as np

# Load image, convert to grayscale, then Otsu's threshold
image = cv2.imread('1.png')
mask = np.zeros(image.shape, dtype=np.uint8)
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
thresh = cv2.threshold(gray, 0, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)[1]

# Morph open with a elliptical shaped kernel
kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (75,75))
opening = cv2.morphologyEx(thresh, cv2.MORPH_OPEN, kernel, iterations=4)

# Find contours and create perfect circle on mask
cnts = cv2.findContours(opening, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
cnts = cnts[0] if len(cnts) == 2 else cnts[1]
for c in cnts:
    ((x, y), r) = cv2.minEnclosingCircle(c)
    cv2.circle(image, (int(x), int(y)), int(r), (36, 255, 12), 3)
    cv2.circle(mask, (int(x), int(y)), int(r), (255, 255, 255), -1)

cv2.imshow('thresh', thresh)
cv2.imshow('opening', opening)
cv2.imshow('image', image)
cv2.imshow('mask', mask)
cv2.waitKey()

如果您没有图像，而是有一个np.array，过程保持不变，但您可以跳过阈值步骤。另外，根据图像的大小，您可能需要调整内核大小。例如，将其更改为(75, 75) to say (10, 10)。您还可以尝试执行变形打开的迭代次数。这是一个示例，说明如何执行此操作（如果您有）np.array形成图像的点

输入图像->变形开放-> Result

Code

import cv2
import numpy as np

mask = np.array([ [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255], [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255], [0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0,0,0,0,255,255,255], [0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0,0,255,255,255], [0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,255,255,255,255], [0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,255,255,255,255], [0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,255,255,255,255,255], [0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,255,255,255,255,255], [0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255], [0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255], [0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255], [0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255], [0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255], [0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255], [0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,255,255], [0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0], [0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0], [0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0], [0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0], [0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0], [0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0], [0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0], [0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0], [0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0], [0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0], [0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0], [0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0,0], [0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0,0,0], [0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0,0,0,0], [0,0,0,0,0,0,0,0,0,0,0,0,0,0,255,255,255,255,255,255,255,255,255,0,0,0,0,0,0,0,0,0,0,0,0], [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0], [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0], [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]], dtype=np.uint8)

# Create blank image with the same size as mask
image = np.zeros(mask.shape, dtype=np.uint8)

# Morph open with a elliptical shaped kernel
kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (10,10))
opening = cv2.morphologyEx(mask, cv2.MORPH_OPEN, kernel, iterations=2)

# Find contours and create perfect circle on mask
cnts = cv2.findContours(opening, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
cnts = cnts[0] if len(cnts) == 2 else cnts[1]
for c in cnts:
    ((x, y), r) = cv2.minEnclosingCircle(c)
    cv2.circle(image, (int(x), int(y)), int(r), (255, 255, 255), -1)

cv2.imshow('opening', opening)
cv2.imshow('image', image)
cv2.imshow('mask', mask)
cv2.waitKey()
本文内容由网友自发贡献，版权归原作者所有，本站不承担相应法律责任。如您发现有涉嫌抄袭侵权的内容，请联系:hwhale#tublm.com(使用前将#替换为@)

python

image

opencv

imageprocessing

ComputerVision

如何修改蒙版使其成为完美的圆形 的相关文章

  • 光学标记阅读器的开源库[关闭]

    Closed 这个问题不符合堆栈溢出指南 help closed questions 目前不接受答案 我想要一个用于光学标记读取 OMR 的开源库 有这样的库吗 我想在我的 java 应用程序中使用这个库 zxing 可能对你有用 http

  • 为什么我不能使用“exclude”从 python 轮子中排除“tests”目录？

    考虑以下包结构 与以下setup py内容 from setuptools import setup find packages setup name dfl client packages find packages exclude te

  • Python Nose 导入错误

    我似乎无法理解鼻子测试框架 https nose readthedocs org en latest 识别文件结构中测试脚本下方的模块 我已经设置了演示该问题的最简单的示例 下面我会解释一下 这是包文件结构 init py foo py t

  • 底图上的子图

    我有一张英国地图和 121 个地点 每个地点有 3 个值 我想绘制 121 个位置中每个位置的三个值的小条形图 目前 这些值绘制为markersize属性 看起来像这样 密集恐惧症情节 https i stack imgur com 5fv

  • 子进程改变目录

    我想在子目录 超级目录中执行脚本 我需要首先进入该子目录 超级目录 我无法得到subprocess进入我的子目录 tducin localhost Projekty tests ve python Python 2 7 4 default

  • Python的reduce()短路了吗？

    If I do result reduce operator and False 1000 得到第一个结果后它会停止吗 自从False anything False 相似地 result reduce operator or True 10

  • 如何从谷歌云存储桶读取音频文件并在datalab笔记本中使用ipd播放

    我想在数据实验室笔记本中播放我从谷歌云存储桶中读取的声音文件 这个怎么做 import numpy as np import IPython display as ipd import librosa import soundfile as

  • 在Python中创建一个新表

    我正在尝试从数控机床中提取数据 事件每毫秒发生一次 我需要过滤掉一些用管道 分隔的变量分隔符 PuTTy exe 程序生成的日志文件 我尝试阅读熊猫 但列不在同一位置 df pd read table data log sep 日志文件的一

  • 在 Flask (WSGI) 中使用全局单例，我是否需要担心竞争条件？ [复制]

    这个问题在这里已经有答案了 Flask 的 hello world 演示是 from flask import Flask app Flask name app route def hello return Hello World if n

  • 从服务器下载图像（cUrl，但接受建议）C++

    我试图通过从服务器 网站 下载图像来设置旋转背景图像 并尝试使用curl 来执行此操作 但是在执行此操作方面取得了0 成功 我的代码的 缩短的 版本如下 我没有收到错误 但是 如何 临时 保存该图像以将其显示为背景 是否有图像 类型变量 或

  • 如何使用 Pandas 将巨大的 CSV 转换为 SQLite？

    我有一个巨大的表 大约 60 GB 采用存档的 CSV 文件形式 我想将其转换为 SQLite 文件 我现在所做的事情如下 import pandas import sqlite3 cnx sqlite3 connect db sqlite

  • 网页抓取 - 前往第 2 页

    如何访问数据集的第二页 无论我做什么 它都只返回第 1 页 import bs4 from urllib request import urlopen as uReq from bs4 import BeautifulSoup as sou

  • 如何修复错误“AttributeError：‘模块’对象在 python3 中没有属性‘客户端’？

    以下是我的代码 import http h1 http client HTTPConnection www bing com 我认为没问题 但是 python 给了我以下错误 AttributeError 模块 对象没有属性 客户端 我想知

  • 在Python中删除带有重音符号的字符串中的所有非字母字符

    我正在尝试使用 Python 3 7 从包含重音符号的字符串中删除所有非字母字符 空格除外 我尝试了以下方法 import re text 29 1981 4 2008 clean text re sub W d text print cl

  • Spark中的count和collect函数抛出IllegalArgumentException

    当我使用时抛出此异常时 我尝试在本地 Spark 上加载一个小数据集count 在 PySpark 中 take 似乎有效 我试图搜索这个问题 但没有找到原因 看来RDD的分区有问题 有任何想法吗 先感谢您 sc stop sc Spark

  • 如何向 SCons 构建添加预处理和后处理操作？

    我正在尝试在使用 SCons 构建项目时添加预处理和后处理操作 SConstruct 和 SConscript 文件位于项目的顶部 预处理动作 生成代码 通过调用不同的工具 gt 不知道在此预处理之后将生成的确切文件 可以创建用于决定生成哪

  • 如何设置 matplotlib 表中列的背景颜色

    我在一个目录中有多个 txt 文件 例如 d memdump 0 txt 1 txt 10 txt 示例文本文件如下 Applications Memory Usage kB Uptime 7857410 Realtime 7857410

  • 从另一个 python 脚本获取返回信息

    我在 Linux 上 我有一个 python 脚本 我想从另一个 python 脚本调用它 我不想将其作为模块导入 为了一层安全性 现在为了学术练习 因为我想弄清楚这一点 我实际上想让一个脚本使用 os system 或另一个类似的函数 并

  • 最小硬币找零问题——回溯

    我正在尝试用最少数量的硬币解决硬币找零问题 采用回溯法 我实际上已经完成了它 但我想添加一些选项 按其单位打印硬币数量 而不仅仅是总数 这是我下面的Python代码 def minimum coins coin list change mi

  • Python 枚举子集迭代

    我想迭代以下枚举的子集 class Items enum Enum item1 0 item2 1 item3 2 item4 3 item5 4 item6 5 item7 6 item8 7 说我想 for item in Items

随机推荐

热门标签