Alright, let's break this down step by step to figure out how to calculate the right rotation angle and get your camera facing that left-side QR code perfectly. First, let's align on the basics with your given coordinate system (y-axis = up/down, x-axis = left/right):

核心旋转逻辑 #

When your camera is positioned to the left of the QR code, the QR code sits in the right portion of the camera's field of view. To face it directly, you need to rotate the camera around the y-axis (this is called the yaw angle in camera terms).

From the perspective of looking down the y-axis (top-down view), this rotation should be clockwise — it swivels the camera's gaze to the right, directly toward the QR code.

如何计算旋转角度 #

There are two reliable ways to get the exact angle, depending on whether you have camera calibration data:

方法1：基于QR码位置与相机内参（高精度） #

If you have your camera's intrinsic parameters (like focal length f), you can use trigonometry to calculate the precise angle:

1. Detect the QR code in your image and get its center coordinates (qr_center_x, qr_center_y).
2. Find your image's center coordinates (img_center_x, img_center_y).
3. Use the horizontal offset between the two centers to compute the angle:

   angle_rad = arctan( (qr_center_x - img_center_x) / f )
   angle_deg = rad2deg(angle_rad)
   

   • A positive angle_deg means you need to rotate clockwise (camera swivels right to face the QR code).
   • A negative value would mean the QR code is on the left (opposite of your scenario), so you'd rotate counterclockwise.

方法2：基于QR码的倾斜检测（快速校准） #

If you don't have camera calibration data, you can use the QR code's inherent tilt in the image to estimate the angle:

1. Detect the QR code's bounding rotated rectangle (most QR code detection libraries return this).
2. The angle of this rotated rectangle is exactly the tilt of the QR code in your frame. Reverse this angle to get the camera rotation needed to make the QR code appear "straight" (and thus the camera is facing it directly).

代码示例（Python + OpenCV） #

Here's a practical implementation to calculate the angle and send a control command (adjust the camera command part to match your hardware's protocol):

import cv2
import numpy as np

def calculate_yaw_angle_for_qr(img_path, camera_focal_length=None):
    # Load and preprocess the image
    img = cv2.imread(img_path)
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    img_h, img_w = gray.shape

    # Initialize QR detector
    qr_detector = cv2.QRCodeDetector()
    data, bbox, _ = qr_detector.detectAndDecode(gray)

    if bbox is not None:
        # Get QR code center
        bbox_points = bbox[0]
        qr_center_x = np.mean(bbox_points[:, 0])
        img_center_x = img_w / 2

        if camera_focal_length:
            # High-precision calculation using focal length
            angle_rad = np.arctan( (qr_center_x - img_center_x) / camera_focal_length )
            angle_deg = np.rad2deg(angle_rad)
        else:
            # Fallback: use QR code's rotation angle
            rect = cv2.minAreaRect(bbox_points)
            angle = rect[2]
            # Adjust angle to standard rotation direction
            angle_deg = -angle if angle > 45 else 90 - angle

        return round(angle_deg, 2)
    else:
        print("No QR code detected in the image.")
        return None

# Example usage
target_angle = calculate_yaw_angle_for_qr("wall_qr_image.jpg", camera_focal_length=800)
if target_angle:
    print(f"Send this command to your camera: `ROTATE_Y {target_angle}`")
    # Replace the line below with your actual camera control API call
    # camera_controller.rotate_y(target_angle)

关键注意事项 #

• Rotation direction alignment: Double-check your camera's control protocol — some systems define clockwise yaw as a negative angle, so you might need to flip the sign of the calculated angle.
• Focal length accuracy: For method 1, use a calibrated focal length (you can calibrate your camera using OpenCV's calibration tools) to avoid angle errors.
• QR code skew: If the QR code itself is tilted on the wall (not just because of camera position), combine the position offset angle with the QR code's own tilt angle for perfect alignment.

内容的提问来源于stack exchange，提问作者Rachel