-215 opencv assertion_error ai_generated true

cv::error: (-215：断言失败) labels.size() == samples.rows 在函数 'cv::kmeans' 中

cv::error: (-215:Assertion failed) labels.size() == samples.rows in function 'cv::kmeans'

ID: opencv/kmeans-empty-labels

其他格式: JSON · Markdown 中文 · English
87%修复率
80%置信度
1证据数
2024-01-12首次发现

版本兼容性

版本状态引入弃用备注
4.5.5 active — — —
4.6.0 active — — —
4.7.0 active — — —
4.8.0 active — — —
4.9.0 active — — —

根因分析

传递给 kmeans 的 labels 数组大小与样本数量不一致，或者样本矩阵为空或行数为零。

English

The labels array provided to kmeans has a different size than the number of samples, or the samples matrix is empty or has zero rows.

generic

官方文档

https://docs.opencv.org/4.x/d5/d38/group__core__cluster.html#ga9a34e1b8c9c3f5a0b8b0c0a0c0a0c0a0

解决方案

  1. 让 OpenCV 自动分配 labels，传递 None：retval, labels, centers = cv2.kmeans(samples, K, None, criteria, attempts, flags)。避免手动尺寸问题。
  2. 如果需要手动提供 labels，确保形状：labels = np.zeros((samples.shape[0], 1), dtype=np.int32)。在调用 kmeans 前用 print(labels.shape) 验证。
  3. 将 samples 转换为 2D float32 矩阵：samples = np.float32(samples).reshape(-1, 2)（处理点集）。确保 samples.rows > 0。

无效尝试

常见但无效的做法:

  1. 30% 失败

    Pre-allocating labels with np.zeros((1, N)) instead of np.zeros((N, 1), dtype=np.int32) — shape mismatch

  2. 25% 失败

    Using np.random.randint to initialize labels without ensuring the dtype is np.int32

  3. 20% 失败

    Assuming the error is about K value being too large and reducing K, which doesn't fix the label size mismatch