OpenCV完结篇——计算机视觉（人脸识别 || 车牌识别）

源代码•宸2023-12-01 21:56

文章目录


Haar人脸识别方法



scaleFactor调整哈尔级联器的人脸选框使其能框住人脸

官方教程指路

每个特征都是通过从黑色矩形下的像素总和减去白色矩形下的像素总和获得的单个值

级联器模型文件位置

python 复制代码 
# -*- coding: utf-8 -*-
import cv2
import numpy as np

cv2.namedWindow('img', cv2.WINDOW_NORMAL)
#第一步，创建Haar级联器
facer = cv2.CascadeClassifier('./haarcascade_frontalface_default.xml')

#第二步，导入人脸识别的图片并将其灰度化
img = cv2.imread('E:/pic/Pic/11.jpg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

#第三步，进行人脸识别
#[[x,y,w,h]]
faces = facer.detectMultiScale(gray, 1.1, 5)

for (x, y, w, h) in faces:
    cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 0), 3)

cv2.imshow('img', img)
if cv2.waitKey(0) & 0xff == 27:
    cv2.destroyAllWindows()

Haar识别眼鼻口

python 复制代码 
# -*- coding: utf-8 -*-
import cv2
import numpy as np

cv2.namedWindow('img', cv2.WINDOW_NORMAL)
#第一步，创建Haar级联器
facer = cv2.CascadeClassifier('./haarcascade_frontalface_default.xml')
eye = cv2.CascadeClassifier('./haarcascade_eye.xml')
mouse = cv2.CascadeClassifier('./haarcascade_mcs_mouth.xml')

#第二步，导入人脸识别的图片并将其灰度化
img = cv2.imread('E:/pic/Pic/11.jpg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

#第三步，进行人脸识别
#[[x,y,w,h]]
faces = facer.detectMultiScale(gray, 1.1, 5)
eyes = eye.detectMultiScale(gray, 1.1, 5)
mouses = mouse.detectMultiScale(gray, 1.1, 5)

for (x, y, w, h) in faces:
    cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 0), 3)

for (x, y, w, h) in eyes:
    cv2.rectangle(img, (x, y), (x + w, y + h), (255, 255, 0), 3)

# for (x, y, w, h) in mouses:
#     cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 255), 3)

cv2.imshow('img', img)
if cv2.waitKey(0) & 0xff == 27:
    cv2.destroyAllWindows()

识别嘴就会不精确了

python 复制代码 
# -*- coding: utf-8 -*-
import cv2
import numpy as np

cv2.namedWindow('img', cv2.WINDOW_NORMAL)
#第一步，创建Haar级联器
facer = cv2.CascadeClassifier('./haarcascade_frontalface_default.xml')
eye = cv2.CascadeClassifier('./haarcascade_eye.xml')
mouse = cv2.CascadeClassifier('./haarcascade_mcs_mouth.xml')

#第二步，导入人脸识别的图片并将其灰度化
img = cv2.imread('E:/pic/Pic/11.jpg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

#第三步，进行人脸识别
#[[x,y,w,h]]
faces = facer.detectMultiScale(gray, 1.1, 5)
eyes = eye.detectMultiScale(gray, 1.1, 5)
mouses = mouse.detectMultiScale(gray, 1.1, 5)

for (x, y, w, h) in faces:
    cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 0), 3)

for (x, y, w, h) in eyes:
    cv2.rectangle(img, (x, y), (x + w, y + h), (255, 255, 0), 3)

for (x, y, w, h) in mouses:
    cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 255), 3)

cv2.imshow('img', img)
if cv2.waitKey(0) & 0xff == 27:
    cv2.destroyAllWindows()

识别鼻子

python 复制代码 
# -*- coding: utf-8 -*-
import cv2
import numpy as np

cv2.namedWindow('img', cv2.WINDOW_NORMAL)
#第一步，创建Haar级联器
facer = cv2.CascadeClassifier('./haarcascade_frontalface_default.xml')
eye = cv2.CascadeClassifier('./haarcascade_eye.xml')
mouse = cv2.CascadeClassifier('./haarcascade_mcs_mouth.xml')
nose = cv2.CascadeClassifier('./haarcascade_mcs_nose.xml')

#第二步，导入人脸识别的图片并将其灰度化
img = cv2.imread('E:/pic/Pic/11.jpg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

#第三步，进行人脸识别
#[[x,y,w,h]]
faces = facer.detectMultiScale(gray, 1.1, 5)
eyes = eye.detectMultiScale(gray, 1.1, 5)
mouses = mouse.detectMultiScale(gray, 1.1, 5)
noses = nose.detectMultiScale(gray, 1.1, 5)

for (x, y, w, h) in faces:
    cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 0), 3)

for (x, y, w, h) in eyes:
    cv2.rectangle(img, (x, y), (x + w, y + h), (255, 255, 0), 3)

for (x, y, w, h) in mouses:
    cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 255), 3)

for (x, y, w, h) in noses:
    cv2.rectangle(img, (x, y), (x + w, y + h), (0, 0, 255), 3)

cv2.imshow('img', img)
if cv2.waitKey(0) & 0xff == 27:
    cv2.destroyAllWindows()

只要不测口，还是比较准确的

python 复制代码 
# -*- coding: utf-8 -*-
import cv2
import numpy as np

cv2.namedWindow('img', cv2.WINDOW_NORMAL)
#第一步，创建Haar级联器
facer = cv2.CascadeClassifier('./haarcascade_frontalface_default.xml')
eye = cv2.CascadeClassifier('./haarcascade_eye.xml')
mouse = cv2.CascadeClassifier('./haarcascade_mcs_mouth.xml')
nose = cv2.CascadeClassifier('./haarcascade_mcs_nose.xml')

#第二步，导入人脸识别的图片并将其灰度化
img = cv2.imread('E:/pic/Pic/11.jpg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

#第三步，进行人脸识别
#[[x,y,w,h]]
faces = facer.detectMultiScale(gray, 1.1, 5)
# eyes = eye.detectMultiScale(gray, 1.1, 5)
# mouses = mouse.detectMultiScale(gray, 1.1, 5)
# noses = nose.detectMultiScale(gray, 1.1, 5)

i = 0

for (x, y, w, h) in faces:
    cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 0), 3)
    roi_img = img[y: y+h, x:x+w]
    eyes = eye.detectMultiScale(roi_img, 1.1, 5)
    for (x, y, w, h) in eyes:
        cv2.rectangle(roi_img, (x, y), (x + w, y + h), (255, 255, 0), 3)
    noses = nose.detectMultiScale(roi_img, 1.1, 5)
    for (x, y, w, h) in noses:
        cv2.rectangle(roi_img, (x, y), (x + w, y + h), (0, 0, 255), 3)
    # mouses = mouse.detectMultiScale(roi_img, 1.1, 5)
    # for (x, y, w, h) in mouses:
    #     cv2.rectangle(roi_img, (x, y), (x + w, y + h), (0, 255, 255), 3)
    
    # i += 1
    # winname = 'face' + str(i)
    # cv2.imshow(winname, roi_img)

# for (x, y, w, h) in mouses:
#     cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 255), 3)

# for (x, y, w, h) in noses:
#     cv2.rectangle(img, (x, y), (x + w, y + h), (0, 0, 255), 3)

cv2.imshow('img', img)
if cv2.waitKey(0) & 0xff == 27:
    cv2.destroyAllWindows()

测口准确度太低！！！

python 复制代码 
# -*- coding: utf-8 -*-
import cv2
import numpy as np

cv2.namedWindow('img', cv2.WINDOW_NORMAL)
#第一步，创建Haar级联器
facer = cv2.CascadeClassifier('./haarcascade_frontalface_default.xml')
eye = cv2.CascadeClassifier('./haarcascade_eye.xml')
mouse = cv2.CascadeClassifier('./haarcascade_mcs_mouth.xml')
nose = cv2.CascadeClassifier('./haarcascade_mcs_nose.xml')

#第二步，导入人脸识别的图片并将其灰度化
img = cv2.imread('E:/pic/Pic/11.jpg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

#第三步，进行人脸识别
#[[x,y,w,h]]
faces = facer.detectMultiScale(gray, 1.1, 5)
# eyes = eye.detectMultiScale(gray, 1.1, 5)
# mouses = mouse.detectMultiScale(gray, 1.1, 5)
# noses = nose.detectMultiScale(gray, 1.1, 5)

i = 0

for (x, y, w, h) in faces:
    cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 0), 3)
    roi_img = img[y: y+h, x:x+w]
    eyes = eye.detectMultiScale(roi_img, 1.1, 5)
    for (x, y, w, h) in eyes:
        cv2.rectangle(roi_img, (x, y), (x + w, y + h), (255, 255, 0), 3)
    noses = nose.detectMultiScale(roi_img, 1.1, 5)
    for (x, y, w, h) in noses:
        cv2.rectangle(roi_img, (x, y), (x + w, y + h), (0, 0, 255), 3)
    mouses = mouse.detectMultiScale(roi_img, 1.1, 5)
    for (x, y, w, h) in mouses:
        cv2.rectangle(roi_img, (x, y), (x + w, y + h), (0, 255, 255), 3)
    
    # i += 1
    # winname = 'face' + str(i)
    # cv2.imshow(winname, roi_img)

# for (x, y, w, h) in mouses:
#     cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 255), 3)

# for (x, y, w, h) in noses:
#     cv2.rectangle(img, (x, y), (x + w, y + h), (0, 0, 255), 3)

cv2.imshow('img', img)
if cv2.waitKey(0) & 0xff == 27:
    cv2.destroyAllWindows()

Haar+Tesseract进行车牌识别




安装很简单，这里贴一个安装教程

配置出现问题的，可以看看这篇博客

测试一下，识别文字还是很准的！！！

python 复制代码 
# -*- coding: utf-8 -*-
import cv2
import numpy as np

# cv2.namedWindow('img', cv2.WINDOW_NORMAL)
#第一步，创建Haar级联器
carplate = cv2.CascadeClassifier('./haarcascade_russian_plate_number.xml')

#第二步，导入带车牌的图片并将其灰度化
img = cv2.imread('./chinacar.jpeg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

#第三步，进行车牌定位
#[[x,y,w,h]]
carplates = carplate.detectMultiScale(gray, 1.1, 5)

for (x, y, w, h) in carplates:
    cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 0), 3)

cv2.imshow('img', img)
if cv2.waitKey(0) & 0xff == 27:
    cv2.destroyAllWindows()
python 复制代码 
# -*- coding: utf-8 -*-
import cv2
import numpy as np

# cv2.namedWindow('img', cv2.WINDOW_NORMAL)
#第一步，创建Haar级联器
carplate = cv2.CascadeClassifier('./haarcascade_russian_plate_number.xml')

#第二步，导入带车牌的图片并将其灰度化
img = cv2.imread('./chinacar.jpeg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

#第三步，进行车牌定位
#[[x,y,w,h]]
carplates = carplate.detectMultiScale(gray, 1.1, 5)

for (x, y, w, h) in carplates:
    cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 0), 3)

# 对获取到的车牌进行预处理
# 1.提取ROI
roi = gray[y: y+h, x:x+w]
# 2.进行二值化
ret, roi_bin = cv2.threshold(roi, 10, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)

cv2.imshow('img', img)
cv2.imshow('roi_bin', roi_bin)
if cv2.waitKey(0) & 0xff == 27:
    cv2.destroyAllWindows()



python 复制代码 
# -*- coding: utf-8 -*-
import cv2
import numpy as np

# 引入tesseract库
import pytesseract

# cv2.namedWindow('img', cv2.WINDOW_NORMAL)
#第一步，创建Haar级联器
carplate = cv2.CascadeClassifier('./haarcascade_russian_plate_number.xml')

#第二步，导入带车牌的图片并将其灰度化
img = cv2.imread('./chinacar.jpeg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

#第三步，进行车牌定位
#[[x,y,w,h]]
carplates = carplate.detectMultiScale(gray, 1.1, 5)

for (x, y, w, h) in carplates:
    cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 0), 3)

# 对获取到的车牌进行预处理
# 1.提取ROI
roi = gray[y: y+h, x:x+w]
# 2.进行二值化
ret, roi_bin = cv2.threshold(roi, 10, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)

pytesseract.pytesseract.tesseract_cmd = r'D:\Program Files\Tesseract_OCR\tesseract.exe'
print(pytesseract.image_to_string(roi, lang='chi_sim+eng', config='--psm 8 --oem 3'))

cv2.imshow('img', img)
cv2.imshow('roi_bin', roi_bin)
if cv2.waitKey(0) & 0xff == 27:
    cv2.destroyAllWindows()

具体实现还需要进一步优化！！！

深度学习基础知识

深度学习是计算机视觉最为重要的方法














dnn实现图像分类





python 复制代码 
# -*- coding: utf-8 -*-
import cv2
from cv2 import dnn
import numpy as np

# 1.导入模型，创建神经网络
# 2.读取图片，转成张量
# 3.将张量输入到网络中，并进行预测
# 4.得到结果，显示

# 导入模型，创建神经网络
config = "./bvlc_googlenet.prototxt"
model = "./bvlc_googlenet.caffemodel"
net = dnn.readNetFromCaffe(config, model)

# 读取图片，转成张量
img = cv2.imread('./smallcat.jpeg')
blob = dnn.blobFromImage(img, 1.0, (224, 224), (104, 117, 123))

# 将张量输入到网络中，并进行预测
net.setInput(blob)
r = net.forward()

# 读取类目
classes = []
path = './synset_words.txt'
with open(path, 'rt') as f:
    classes = [x [x.find(" ") + 1:] for x in f]

order = sorted(r[0], reverse=True)
z = list(range(3))

for i in list(range(0, 3)):
    z[i] = np.where(r[0] == order[i])[0][0]
    print('No.', i + 1, ' matches:', classes[z[i]], end='')
    print('category row is at:', z[i] + 1, ' ', 'posibility:', order[i])


之后我会持续更新，如果喜欢我的文章，请记得一键三连哦，点赞关注收藏，你的每一个赞每一份关注每一次收藏都将是我前进路上的无限动力 ！！！↖(▔▽▔)↗感谢支持！

相关推荐
风象南5 小时前
普通人用AI加持赚到的第一个100块
人工智能·后端
牛奶5 小时前
2026年大模型怎么选？前端人实用对比
前端·人工智能·ai编程
牛奶5 小时前
前端人为什么要学AI？
前端·人工智能·ai编程
地平线开发者7 小时前
SparseDrive 模型导出与性能优化实战
算法·自动驾驶
董董灿是个攻城狮7 小时前
大模型连载2：初步认识 tokenizer 的过程
算法
地平线开发者8 小时前
地平线 VP 接口工程实践（一）：hbVPRoiResize 接口功能、使用约束与典型问题总结
算法·自动驾驶
罗西的思考8 小时前
AI Agent框架探秘：拆解 OpenHands（10）--- Runtime
人工智能·算法·机器学习
冬奇Lab9 小时前
OpenClaw 源码精读（2）：Channel & Routing——一条消息如何找到它的 Agent？
人工智能·开源·源码阅读
冬奇Lab9 小时前
一天一个开源项目（第38篇）：Claude Code Telegram - 用 Telegram 远程用 Claude Code，随时随地聊项目
人工智能·开源·资讯
格砸10 小时前
从入门到辞职｜从ChatGPT到OpenClaw，跟上智能时代的进化
前端·人工智能·后端
热门推荐
01GitHub 镜像站点02OpenClaw 使用和管理 MCP 完全指南03OpenClaw + 飞书（Feishu）环境搭建指南04【OpenClaw 本地实战 Ep.3】突破瓶颈：强制修改 openclaw.json 解锁 32k 上下文记忆05Claude Code + GLM4.7 避坑指南：解决 Unable to connect to Anthropic services06OpenClaw优化飞书API 额度已耗尽问题07小黑课堂计算机二级WPSoffice题库软件下载安装教程（2026年3月最新版）08Clawdbot部署教程：解决‘gateway token missing’授权问题的完整步骤09Window 10部署openclaw报错node.exe : npm error code 12810OpenClaw大龙虾机器人完整安装教程