Python+OpenCV虹膜检测与直径测量：解决MediaPipe误检问题

问题概述 #

• 眼睛闭合时仍会检测并绘制虹膜圆圈
• 面部移动时出现双重检测圈
• 需要实现虹膜直径的测量

解决方案细节 #

1. 解决眼睛闭合时的误检测 #

通过计算上下眼睑的垂直距离判断眼睛状态：

• 提取眼睛上下边界的关键点（左眼：386/374，右眼：159/145）
• 计算两点间的欧氏距离，当距离小于设定阈值时判定为闭眼，跳过虹膜绘制

2. 解决移动时的双重检测圈 #

• 验证虹膜关键点的置信度（visibility属性），仅当置信度高于0.5时使用该点数据
• 确保每次循环基于新读取的帧绘制，避免帧残留干扰

3. 虹膜直径测量 #

• 直接从最小外接圆的半径计算像素直径（直径 = 2 * 半径）
• 将直径数据实时绘制在画面上，若需物理尺寸，可结合摄像头参数与拍摄距离换算

完整修改代码 #

import cv2 as cv
import numpy as np
import mediapipe as mp

mp_face_mesh = mp.solutions.face_mesh

LEFT_EYE = [362, 382, 381, 380, 374, 373, 390, 249, 263, 466, 388, 387, 386, 385, 384, 398]
RIGHT_EYE = [33, 7, 163, 144, 145, 153, 154, 155, 133, 173, 157, 158, 159, 160, 161, 246]

LEFT_IRIS = [474, 475, 476, 477]
RIGHT_IRIS = [469, 470, 471, 472]
# 眼睛开合判断关键点：上眼睑/下眼睑
LEFT_EYE_OPENNESS = [386, 374]
RIGHT_EYE_OPENNESS = [159, 145]
# 闭眼阈值（可根据实际情况调整）
EYE_CLOSE_THRESHOLD = 12

cap = cv.VideoCapture(0)

with mp_face_mesh.FaceMesh(
        max_num_faces=1,
        refine_landmarks=True,
        min_detection_confidence=0.5,
        min_tracking_confidence=0.5
) as face_mesh:
    while True:
        ret, frame = cap.read()
        if not ret:
            break
        frame = cv.flip(frame, 1)
        rgb_frame = cv.cvtColor(frame, cv.COLOR_BGR2RGB)
        img_h, img_w = frame.shape[:2]
        results = face_mesh.process(rgb_frame)
        mask = np.zeros((img_h, img_w), dtype=np.uint8)

        if results.multi_face_landmarks:
            face_landmarks = results.multi_face_landmarks[0]
            mesh_points = np.array([np.multiply([p.x, p.y], [img_w, img_h]).astype(int)
                                    for p in face_landmarks.landmark])
            
            # 计算眼睛开合度
            left_eye_dist = np.linalg.norm(mesh_points[LEFT_EYE_OPENNESS[0]] - mesh_points[LEFT_EYE_OPENNESS[1]])
            right_eye_dist = np.linalg.norm(mesh_points[RIGHT_EYE_OPENNESS[0]] - mesh_points[RIGHT_EYE_OPENNESS[1]])

            # 处理左眼虹膜
            if left_eye_dist > EYE_CLOSE_THRESHOLD:
                # 检查虹膜关键点置信度
                left_iris_visible = all(face_landmarks.landmark[idx].visibility > 0.5 for idx in LEFT_IRIS)
                if left_iris_visible:
                    (l_cx, l_cy), l_radius = cv.minEnclosingCircle(mesh_points[LEFT_IRIS])
                    center_left = np.array([l_cx, l_cy], dtype=np.int32)
                    cv.circle(frame, center_left, int(l_radius), (0, 255, 0), 2, cv.LINE_AA)
                    cv.circle(mask, center_left, int(l_radius), (255, 255, 255), -1, cv.LINE_AA)
                    # 计算并显示左眼虹膜直径
                    left_diameter = 2 * l_radius
                    cv.putText(frame, f"Left Iris Diameter: {left_diameter:.1f}px", (20, 40), 
                               cv.FONT_HERSHEY_SIMPLEX, 0.7, (0, 255, 0), 2)
            
            # 处理右眼虹膜
            if right_eye_dist > EYE_CLOSE_THRESHOLD:
                # 检查虹膜关键点置信度
                right_iris_visible = all(face_landmarks.landmark[idx].visibility > 0.5 for idx in RIGHT_IRIS)
                if right_iris_visible:
                    (r_cx, r_cy), r_radius = cv.minEnclosingCircle(mesh_points[RIGHT_IRIS])
                    center_right = np.array([r_cx, r_cy], dtype=np.int32)
                    cv.circle(frame, center_right, int(r_radius), (0, 255, 0), 2, cv.LINE_AA)
                    cv.circle(mask, center_right, int(r_radius), (255, 255, 255), -1, cv.LINE_AA)
                    # 计算并显示右眼虹膜直径
                    right_diameter = 2 * r_radius
                    cv.putText(frame, f"Right Iris Diameter: {right_diameter:.1f}px", (20, 80), 
                               cv.FONT_HERSHEY_SIMPLEX, 0.7, (0, 255, 0), 2)

        cv.imshow('Mask', mask)
        cv.imshow('Iris Detection', frame)
        key = cv.waitKey(1)
        if key == ord('q'):
            break
cap.release()
cv.destroyAllWindows()

代码说明 #

• 眼睛开合判断：通过上下眼睑关键点的距离判断闭眼状态，阈值可根据摄像头分辨率调整
• 置信度验证：过滤低置信度的虹膜关键点，避免移动时的错误检测
• 直径显示：实时计算并绘制虹膜的像素直径，方便直观查看

内容的提问来源于stack exchange，提问作者Roman Bhuvanesh