运行报错

[yolic66]

Traceback (most recent call last): File "c:\Users\74773\Desktop\Sfm-python-master\Sfm-python-master\revise_v2.py", line 329, in main() File "c:\Users\74773\Desktop\Sfm-python-master\Sfm-python-master\revise_v2.py", line 286, in main key_points_for_all, descriptor_for_all, colors_for_all = extract_features(img_names) ^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "c:\Users\74773\Desktop\Sfm-python-master\Sfm-python-master\revise_v2.py", line 46, in extract_features return np.array(key_points_for_all), np.array(descriptor_for_all), np.array(colors_for_all) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (14,) + inhomogeneous part.

[Arrogant-Q]

因为SIFT检测出来的两个图像特征数量不一样，导致最后转为numpy数组的时候，每个list中的两个元素长度不一样

[ddjiayou]

请问怎样解决呢

[zhaosiyuan1098]

我遇到了一模一样的问题，请问目前有什么解决方法吗？

[pyre001]

将numpy库降级成1.23.0版本可解决该问题