读手用opencv需要大家安装opencv-python，安装后执行如下代码，启动摄像头cap=cv2.VideoCapture(0)0代表系统默认摄像头。用opencv在视频流中创建一个正方形cv2.rectangle(overlay,(x,y),(x+self.rect_width,y+self.rect_width),(255,0,0),-1)获取食指坐标usingmedialpipeRecognizehumanhandsimportmediapipeasmpself.mp_drawing=mp.solutions.drawing_utils#Recognizepalmresults=hands.process(self.image)forhand_landmarksinresults.multi_hand_landmarks:landmark_list=[]#存储手掌关键点坐标forlandmark_id,enumerate(hand_landmarks.landmark)中的finger_axis：landmark_list.append([landmark_id,finger_axis.x,finger_axis.y,finger_axis.z])获取食指的x、y、z坐标。＃食指食指index_finger_tip=landmark_list[8]index_finger_tip_x=ratio_x_to_to_pixel（index_finger_tip[1]））):x=self.rect_left_x#盒子左上角的x坐标y=self.rect_left_y#盒子左上角的y坐标if(x
