Possible problem in new VoxelGrid implementation from PCL 1.5.0

最新推荐文章于 2024-07-17 14:51:08 发布

songjinghao

最新推荐文章于 2024-07-17 14:51:08 发布

阅读量1k

点赞数

文章标签： VoxelGrid PCLPoint Cloud Libra

本文链接：https://blog.youkuaiyun.com/songjinghao/article/details/8681539

版权

在使用 PCL 1.5.0 的 VoxelGrid 过滤器时，如果输入输出变量相同，可能会导致过滤过程崩溃。通过使用输出临时变量和正确调整大小，可以解决此问题。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

今天在学习PCL的时候，发现了一个问题，上网查找的结果跟大家分享一下，希望对大家以后的学习有所帮助。

在此，感谢发现并提出解决方案的Mike先生。

In the VoxelGrid filter, with the new implementation of VoxelGrid (PCL 1.5.0), if you are using the same variable for input and output as explained in the "Aligning object templates to a point cloud" tutorial, the filtering process will crash (it was working correctly in 1.4.0). It will work correctly if you use "SetInputCloud" from the VoxelGrid class.

In PCL 1.5.0, at line 280 in voxel_grid.hpp (output.points.resize (total);), it will resize the variable "input_" from the base class "pcl::PCLBase" too (because variable "input_" and "output" are the same reference). Because of this, in voxel_grid.hpp lines 327, 328, 329 or 343, it will crash if you are trying to access the vector input_->points[] outside his new size. It can be easily fixed by using a variable called "output_temp" instead of using "output" in the function "pcl::VoxelGrid<PointT>::applyFilter (PointCloud &output)" and, at the last line of the function, resize "output" and copy "output_temp" in "output".