6D pose estimation method based on an instance segmentation network and iterative optimization

Abstract

The invention relates to a 6D pose estimation method based on an instance segmentation network and iterative optimization. The 6D pose estimation method comprises the following steps: 1) converting adepth image into an HHA feature map and a scene point cloud; 2) inputting the HHA feature map and the color image into an instance segmentation network added with a spatial transformation network, andperforming instance segmentation of any pose object to obtain an object type identification result and a mask segmentation result; 3) segmenting a target point cloud on the scenic spot cloud according to an instance segmentation result; And 4) according to an improved 4PCS algorithm and an ICP algorithm, performing matching and pose finishing on the segmented target point cloud and the model point cloud of the target CAD model, thereby obtaining an accurate pose estimation result. Compared with the prior art, the method has the advantages of accurate recognition, multi-type object recognition, high detection performance, high pose matching precision and the like.

Description

technical field [0001] The invention relates to the technical field of environment perception of robots, in particular to a 6D pose estimation method based on instance segmentation network and iterative optimization. Background technique [0002] The environmental perception technology of robots is an important scientific issue in the field of robotics research. In recent years, with the development of computer vision and deep learning technology, vision-based environmental perception has become a hot spot in academia and industry. Robots realize the perception of their environment and operating objects through the input of visual information, including recognition and state estimation, etc., and then realize interaction and complete tasks. Among them, the introduction of 3D vision technology enables robots to obtain more abundant information, which plays an important role in promoting the solution of robot environment perception problems. [0003] The present invention foc...

AI Technical Summary

Problems solved by technology

The industrial application prospects of such solutions are limited. In the case of cluttered backgrounds or stacked workpieces, it is difficult to identify workpieces in the field of view through traditional image processing solutions such as edge detection and template detection; at the same time, edge detection, etc. There are some parameters that need to be manually adjusted in the algorithm, and it is often difficult to adapt to different grasping objects, that is, different shapes, different sizes, or even mixed types of workpieces

[0006] From the perspective of pose estimation algorithms, existing patents are still limited to traditional methods of manually designing features and extracting features, and then matching the features extracted from the objects in the actual scene with the features of the model or template.

The artificial feature extraction and template matching scheme has low time efficiency on the one hand, and the search algorithm in the template matching process needs to consume a lot of computing resources. On the other hand, the traditional feature extraction strategy requires the target to be detected to have distinguishable texture features, while This is often difficult to meet in the application of the industrial field. Most of the workpieces in the industrial field are less textured. The traditional feature extraction method cannot meet this requirement. When faced with a variety of different types and shapes of industrial workpieces, Traditional solutions are also unable to adapt

[0007] To sum up, the traditional solutions can only obtain the two-dimensional plane attitude of the workpiece, or the accuracy and robustness of the algorithm are not strong, and cannot adapt to complex life and production application scenarios, and cannot guarantee the accuracy of various shapes, Various types of objects can accurately identify and calculate the pose, which cannot meet the requirements of complex life services and industrial production

Images