Pycharm - OpenCV 04训练数据

4 训练数据

4.1 准备

准备工作

1. 下载训练数据 ‘orl_faces’
2. def前面该装的都装好。
3. project<new<directory<data<jm
4. project<new<directory<trainer

4.2 训练过程

import os
import cv2 as cv
import numpy as np
import sys
from PIL import Image

def getImageAndLabels(path):
    # 建立数组，然后往里面填数据
    facesSamples = []
    ids = []
    # os.listdir(path)获取文件，遍历名称
    imagePaths = [os.path.join(path, f) for f in os.listdir(path)]
    print(imagePaths)
    face_detector = cv.CascadeClassifier(
        'D:/LenovoQMDownload/opencv/sources/data/haarcascades/haarcascade_frontalface_default.xml')
    # 遍历列表中的图片
    for imagePath in imagePaths:
        # 打开图片，调用open，返回PIL_img
        PIL_img = Image.open(imagePath).convert('L')
        # 将图片转换成数组
        img_numpy = np.array(PIL_img, 'uint8')
        faces = face_detector.detectMultiScale(img_numpy)
        # 获取每张图片的id，切分图像的'路径'和'名称+后缀'
        print(os.path.split(imagePath))
        id = int(os.path.split(imagePath)[1].split('.')[0])
        for x, y, w, h in faces:
            # 提取人脸，进行切片img_numpy[y:y+h, x:x+w]，然后放到facesSamples中
            facesSamples.append(img_numpy[y:y+h, x:x+w])
            ids.append(id)
    return facesSamples, ids

if __name__ == '__main__':
    # 获取图片路径
    path ='./data/jm/'
    # 获取图像数组和id标签数组
    faces, ids = getImageAndLabels(path)
    # 获取循环对象
    recognizer = cv.face.LBPHFaceRecognizer_create()
    # 将ids列表转换成数组
    recognizer.train(faces, np.array(ids))
    # 保存文件
    recognizer.write('trainer/trainer.yml')

结果

['./data/jm/1.pgm', './data/jm/10.pgm', './data/jm/11.pgm', './data/jm/12.pgm', './data/jm/13.pgm', './data/jm/14.pgm', './data/jm/15.pgm', './data/jm/2.pgm', './data/jm/3.pgm', './data/jm/4.pgm', './data/jm/5.pgm', './data/jm/6.pgm', './data/jm/7.pgm', './data/jm/8.pgm', './data/jm/9.pgm']
('./data/jm', '1.pgm')
('./data/jm', '10.pgm')
('./data/jm', '11.pgm')
('./data/jm', '12.pgm')
('./data/jm', '13.pgm')
('./data/jm', '14.pgm')
('./data/jm', '15.pgm')
('./data/jm', '2.pgm')
('./data/jm', '3.pgm')
('./data/jm', '4.pgm')
('./data/jm', '5.pgm')
('./data/jm', '6.pgm')
('./data/jm', '7.pgm')
('./data/jm', '8.pgm')
('./data/jm', '9.pgm')

4.2.1 训练过程问题

列表输出顺序错乱

报错：module ‘cv2.face’ has no attribute 'LBPHFaceReconizer_creat

是的，没错，单词 ‘Reconizer’ 写错了！

报错：module ‘cv2.cv2’ has no attribute ‘face’

解决：
安装他俩解决
opencv-python
opencv-contrib-python

4.3 问题解决方案

4.3.1 列表排序函数

result = ["1.pgm", "11.pgm", "2.pgm" , "10.pgm"]
def return_sort_list(result):
    new_result = [int(value.split(".")[0]) for value in result]
    new_result.sort()
    my_result = [str(num) + ".pgm" for num in new_result]
    return my_result
new_list = return_sort_list(result)

print(new_list)

['1.pgm', '2.pgm', '10.pgm', '11.pgm']

4.3.1 sort拆分排序

1. 把列表中的每一个文件名的数字获取，放到一个列表里面

2. 对数字列表进行排序

3. 给数字列表的中每一个数字加上.png

result = ["1.pgm", "11.pgm", "2.pgm" , "10.pgm"]

new_result = [int(value.split(".")[0]) for value in result]
new_result.sort()
print(new_result)

my_result = [str(num) + ".pgm" for num in new_result]
print(my_result)

[1, 2, 10, 11]
['1.pgm', '2.pgm', '10.pgm', '11.pgm']

4.3.1 文件名获取过程

import os
imagePath = "/a/imgs/1.png"

# 获取路径中的文件名
file_name = os.path.basename(imagePath)
print(file_name)
# 分别获取整个文件的文件名和后缀
name, end_str = os.path.splitext(file_name)
print(name, end_str)
num = int(name)
print(num, type(num))

1.png
1 .png
1 <class 'int'>

num = int(os.path.splitext(os.path.basename(imagePath))[0])
print(num)