### 使用 OpenCV 和 CNN 实现手写文字识别 #### 准备工作 为了使用卷积神经网络(CNN)进行手写文字识别，需先安装必要的库。这通常涉及`tensorflow`或`keras`用于构建和训练模型，以及`opencv-python`来处理图像。 ```bash pip install tensorflow opencv-python numpy matplotlib scikit-learn ``` #### 数据准备 MNIST是一个常用的手写字体数据库，包含了大量0到9的手写数字图片及其标签，非常适合用来作为训练集[^2]。可以通过Keras内置函数轻松获取该数据集： ```python from keras.datasets import mnist import numpy as np (x_train, y_train), (x_test, y_test) = mnist.load_data() print('Training data shape:', x_train.shape) print('Test data shape', x_test.shape) # 归一化输入数据至0~1之间并调整维度以适应CNN输入要求 x_train = x_train.astype('float32') / 255. x_test = x_test.astype('float32') / 255. x_train = np.expand_dims(x_train, axis=-1) x_test = np.expand_dims(x_test, axis=-1) ``` #### 构建CNN模型 定义一个简单的CNN架构来进行分类任务: ```python from tensorflow.keras.models import Sequential from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout model = Sequential([ Conv2D(filters=32, kernel_size=(3, 3), activation='relu', input_shape=(28, 28, 1)), MaxPooling2D(pool_size=(2, 2)), Conv2D(filters=64, kernel_size=(3, 3), activation='relu'), MaxPooling2D(pool_size=(2, 2)), Flatten(), Dense(units=128, activation='relu'), Dropout(rate=0.5), Dense(units=10, activation='softmax') ]) model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy']) ``` #### 训练模型 设置好参数后即可开始训练过程，在此期间可以监控验证集上的表现以便及时发现过拟合等问题。 ```python history = model.fit( x=x_train, y=y_train, epochs=10, batch_size=128, validation_split=0.2 ) ``` #### 测试与评估 完成训练后，利用测试集检验模型性能，并计算准确度等指标。 ```python test_loss, test_acc = model.evaluate(x_test, y_test) print(f'Test accuracy: {test_acc:.4f}') ``` #### 应用OpenCV读取自定义图片 当想要预测新采集的手写数字图象时，可借助OpenCV读入文件并对之做适当预处理使之匹配先前设定的数据格式。 ```python import cv2 def preprocess_image(image_path): img = cv2.imread(image_path, cv2.IMREAD_GRAYSCALE) resized_img = cv2.resize(img, dsize=(28, 28)) normalized_img = resized_img.reshape((28, 28, 1)).astype('float32') / 255 return normalized_img image_to_predict = 'path/to/your/image.png' preprocessed_image = preprocess_image(image_to_predict) prediction = model.predict(np.array([preprocessed_image])) predicted_digit = prediction.argmax() print(f'Predicted digit is: {predicted_digit}') ```