opencv摄像头识别手写数字的代码

以下是使用Opencv进行手写数字识别的代码示例：

import cv2
import numpy as np

def get_contour_precedence(contour, cols):
    
    tolerance_factor = 10
    origin = cv2.boundingRect(contour)
    return ((origin[1]//tolerance_factor)*tolerance_factor)*cols + origin[0]

def main():
    
    # Load the model
    net = cv2.dnn.readNet("model.pb")

    classes = []
    with open("model.txt", "r") as f:
        classes = [line.strip() for line in f.readlines()]

    # Capture from the default camera
    cap = cv2.VideoCapture(0)

    while True:
        _, frame = cap.read()

        # Convert the captured frame into grayscale
        gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

        # Apply adaptive thresholding to segment the foreground from the background
        thresh = cv2.adaptiveThreshold(gray, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY_INV, 9, 2)

        # Find contours in the binary image
        contours, _ = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)

        # Sort the contours in the order of y-coordinate
        contours.sort(key=lambda x:get_contour_precedence(x, frame.shape[1]))

        for contour in contours:
            # Get the bounding box for the contour
            x, y, w, h = cv2.boundingRect(contour)

            # Extract the digit from the binary image
            digit_roi = thresh[y:y+h, x:x+w]

            # Resize the digit image to (28, 28)
            digit_roi = cv2.resize(digit_roi, (28, 28))

            # Expand the image to (32, 32) for consistency with the training data
            digit_roi = cv2.copyMakeBorder(digit_roi, 2, 2, 2, 2, cv2.BORDER_CONSTANT, value=0)

            # Normalize the pixel values to the range (0, 1)
            digit_roi = digit_roi.astype("float32") / 255.0

            # Add an extra dimension to the image to make it compatible with the input shape of the model (1, 32, 32, 1)
            digit_roi = np.expand_dims(digit_roi, axis=-1)
            digit_roi = np.expand_dims(digit_roi, axis=0)

            # Feed the image to the model and get the predicted class label
            net.setInput(digit_roi)
            preds = net.forward()
            class_idx = np.argmax(preds)

            # Convert the class label to the corresponding digit
            digit = classes[class_idx]

            # Draw the digit and the bounding box on the original frame
            cv2.rectangle(frame, (x, y), (x+w, y+h), (0, 255, 0), 2)
            cv2.putText(frame, digit, (x+int(w/4), y+int(h/2)), cv2.FONT_HERSHEY_COMPLEX, 1, (0, 255, 0), 2)

        # Show the output
        cv2.imshow("Handwritten Digit Recognition", frame)

        # Press 'q' to quit
        if cv2.waitKey(1) == ord('q'):
            break

    cap.release()
    cv2.destroyAllWindows()

if __name__ == "__main__":
    main()

这段代码使用Opencv实现了从摄像头捕获图像，然后使用预训练好的模型对手写数字进行识别，并在屏幕上展示预测结果。在使用时，请确保正确设置模型的路径和模型的输入输出形状。