Source code and scheme for target recognition, detection and 6D attitude estimation (the most advanced method and data set)

3D Accurate positioning and pose estimation of objects are very important for many higher-level tasks , For example, robot operation （ Such as
Amazon Picking Challenge）、 Scene interpretation and augmented reality . The recent introduction of consumer grade depth sensors has allowed for the traditional 2D The method is substantially improved , Because it can capture more detailed 3D Geometric features . However , There are still some challenges to be solved , Include ： Foreground occlusion , The background is messy , Large scale and postural changes , Multiple instance objects ！

【 So far this column 16 Chapter , subscribe 30+, Originality is not easy. , Subscribe to the source code solution 】

The scope of this article is to list the most advanced methods and data sets that can be used to further help drive research , Don't be wordy. Let's look at the effect below .

Advanced algorithms ：（ The connection on the right side is described in detail ）

title