MediaPipe手部识别示例代码（Python/C++）

上回说到，mediapipe如何安装，这回我们来看看mediapipe是如何 识别 手的位置和返回坐标的。

首先我们调用mediapipe库

import mediapipe as mpimport cv2import numpy as np

之后我们使用此代码进行识别

mp_drawing = mp.solutions.drawing_utilsmp_drawing_styles = mp.solutions.drawing_utils.DrawingSpecmp_hands = mp.solutions.handsIMAGE_FILES = ["p2.jpg"]with mp_hands.Hands(    static_image_mode=True,    max_num_hands=2,    min_detection_confidence=0.5) as hands:  for idx, file in enumerate(IMAGE_FILES):        image = cv2.flip(cv2.imread(file), 1)        results = hands.process(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))         print('Handedness:', results.multi_handedness)    if not results.multi_hand_landmarks:      continue    image_height, image_width, _ = image.shape    annotated_image = image.copy()    for hand_landmarks in results.multi_hand_landmarks:            print(          f'#5: (',          f'{hand_landmarks.landmark[mp_hands.HandLandmark.INDEX_FINGER_MCP].x * image_width}, '          f'{hand_landmarks.landmark[mp_hands.HandLandmark.INDEX_FINGER_MCP].y * image_height})\n'          f'#6: (',          f'{hand_landmarks.landmark[mp_hands.HandLandmark.INDEX_FINGER_PIP].x * image_width}, '          f'{hand_landmarks.landmark[mp_hands.HandLandmark.INDEX_FINGER_PIP].y * image_height})\n'          f'#7: (',          f'{hand_landmarks.landmark[mp_hands.HandLandmark.INDEX_FINGER_DIP].x * image_width}, '          f'{hand_landmarks.landmark[mp_hands.HandLandmark.INDEX_FINGER_DIP].y * image_height})\n'          f'#8: (',          f'{hand_landmarks.landmark[mp_hands.HandLandmark.INDEX_FINGER_TIP].x * image_width}, '          f'{hand_landmarks.landmark[mp_hands.HandLandmark.INDEX_FINGER_TIP].y * image_height})\n'          f'#12: (',          f'{hand_landmarks.landmark[mp_hands.HandLandmark.MIDDLE_FINGER_TIP].x * image_width}, '          f'{hand_landmarks.landmark[mp_hands.HandLandmark.MIDDLE_FINGER_TIP].y * image_height})\n'          f'#16: (',          f'{hand_landmarks.landmark[mp_hands.HandLandmark.RING_FINGER_TIP].x * image_width}, '          f'{hand_landmarks.landmark[mp_hands.HandLandmark.RING_FINGER_TIP].y * image_height})\n'          f'#20: (',          f'{hand_landmarks.landmark[mp_hands.HandLandmark.PINKY_TIP].x * image_width}, '          f'{hand_landmarks.landmark[mp_hands.HandLandmark.PINKY_TIP].y * image_height})\n'                )      mp_drawing.draw_landmarks(          annotated_image,          hand_landmarks,          mp_hands.HAND_CONNECTIONS,          mp_drawing_styles(),          mp_drawing_styles())    cv2.imwrite(        'annotated_image' + str(idx) + '.png', cv2.flip(annotated_image, 1))    print(cv2.imread("hand.jpg").shape)   

可能有朋友要问，怎么能知道返回的是哪个点的位置呢，我们看一看它手指关节点 对照 表，然后就可以返回它们的x，y，z的值了。但是要注意，mediapipe的坐标值都是经过归一化的，如果需要绝对 坐标 ，需要分别对应地乘上图片的宽和高。

此图是示例代码返回的数值，因为示例图片为某非公开项目的图片，因此不予公布效果图。

免责声明：本文系网络转载或改编，未找到原创作者，版权归原作者所有。如涉及版权，请联系删