-215 opencv assertion_error ai_generated true

cv::error: (-215:断言失败) objectPoints.size() == imagePoints.size() && objectPoints.size() >= 4 在函数 'cv::solvePnP' 中

cv::error: (-215:Assertion failed) objectPoints.size() == imagePoints.size() && objectPoints.size() >= 4 in function 'cv::solvePnP'

ID: opencv/calib3d-solvepnp-empty-points

其他格式: JSON · Markdown 中文 · English
90%修复率
88%置信度
1证据数
2024-07-12首次发现

版本兼容性

版本状态引入弃用备注
4.8.0 active — — —
4.9.0 active — — —
4.10.0 active — — —

根因分析

3D 物体点的数量与 2D 图像点的数量不匹配，或者点对应关系少于 PnP 求解所需的 4 个。

English

The number of 3D object points does not match the number of 2D image points, or there are fewer than 4 point correspondences required for PnP solving.

generic

官方文档

https://docs.opencv.org/4.x/d9/d0c/group__calib3d.html#ga549c2075d5aa0ed5f5a6e6d7f3e4b5b9

解决方案

  1. Ensure both vectors have equal size and contain at least 4 points. Use a checkerboard detection to generate at least 4 corners: cv::findChessboardCorners returns a vector of corners that can be paired with known 3D object points.
  2. If using ArUco markers, ensure at least 4 markers are detected and their corners are properly paired with 3D model points using cv::aruco::estimatePoseSingleMarkers for each marker.

无效尝试

常见但无效的做法:

  1. Adding random extra points to objectPoints to match imagePoints size without verifying correspondence 95% 失败

    Incorrect correspondences lead to wildly wrong pose estimates or solver failure; PnP requires accurate point pairs.

  2. Using only 3 points because the user thinks 3 is enough for pose estimation 100% 失败

    solvePnP requires at least 4 non-coplanar points for a unique solution; 3 points can only provide up to 4 possible solutions.