import cv2 import os import numpy as np from sklearn import svm import joblib def read_images(folder): images = [] labels = [] for filename in os.listdir(folder): label = filename.split('.')[0] img = cv2.imread(os.path.join(folder,filename)) if img is not None: images.append(img) labels.append(label) return images, labels # 提取特征向量 def extract_features(images): features = [] for img in images: gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) hist = cv2.calcHist([gray],[0],None,[256],[0,256]) features.append(hist.flatten()) return features # 读取图像和标签 images, labels = read_images('C:/Users/Administrator/Desktop/111') # 提取特征向量 features = extract_features(images) # 训练模型 clf = svm.SVC() clf.fit(features, labels) # 保存模型到文件 joblib.dump(clf, 'model.pkl') clf = joblib.load('E:/xiangmu/measure/model.pkl') # 预测新图像 img = cv2.imread('C:/Users/Administrator/Downloads/2fa446fe46477a2850d029fedce20ae6.jpeg') gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) hist = cv2.calcHist([gray],[0],None,[256],[0,256]) features = np.array(hist.flatten()).reshape(1, -1) label = clf.predict(features)[0] print('识别结果：', label) 我想知道怎么获取返回结果的概率值

在这个例子中，你可以使用`predict_proba`方法来获取预测结果的概率值。具体来说，可以将新图像的特征向量作为输入，使用`predict_proba`方法来获取属于各个类别的概率值。代码如下：

import cv2
import os
import numpy as np
from sklearn import svm
import joblib

def read_images(folder):
    images = []
    labels = []
    for filename in os.listdir(folder):
        label = filename.split('.')[0]
        img = cv2.imread(os.path.join(folder,filename))
        if img is not None:
            images.append(img)
            labels.append(label)
    return images, labels

def extract_features(images):
    features = []
    for img in images:
        gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
        hist = cv2.calcHist([gray],[0],None,[256],[0,256])
        features.append(hist.flatten())
    return features

# 读取图像和标签
images, labels = read_images('C:/Users/Administrator/Desktop/111')

# 提取特征向量
features = extract_features(images)

# 训练模型
clf = svm.SVC(probability=True) # 启用概率估计功能
clf.fit(features, labels)

# 保存模型到文件
joblib.dump(clf, 'model.pkl')

# 加载模型
clf = joblib.load('model.pkl')

# 预测新图像
img = cv2.imread('C:/Users/Administrator/Downloads/2fa446fe46477a2850d029fedce20ae6.jpeg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
hist = cv2.calcHist([gray],[0],None,[256],[0,256])
features = np.array(hist.flatten()).reshape(1, -1)

# 预测结果的概率值
proba = clf.predict_proba(features)[0]
print('概率值：', proba)

# 预测结果
label = clf.predict(features)[0]
print('识别结果：', label)

这里需要将SVM分类器的`probability`参数设置为True来启用概率估计功能，然后使用`predict_proba`方法来获取预测结果的概率值。在这个例子中，`proba`变量是一个一维数组，包含了新图像属于各个类别的概率值。