Three-dimensional object detection method based on generation and refinement of occlusion representation

By generating and refining occlusion representations, the problem of object recognition in 3D target detection involving sparse point clouds and incomplete shapes is solved, improving the robustness and generalization ability of the detector and achieving high-precision object recognition.

CN117218607BActive Publication Date: 2026-06-12NANJING UNIV OF SCI & TECH
0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
NANJING UNIV OF SCI & TECH
Filing Date
2023-08-27
Publication Date
2026-06-12

AI Technical Summary

⚠Technical Problem

Existing 3D object detection methods cannot effectively identify object positions when processing sparse and incomplete point cloud data, and do not consider the alignment and domain differences between objects when generating target point clouds, resulting in poor detector generalization.

⚗Method used

We employ a method of generating and refining occlusion representations. Initial occlusion representations are generated through a candidate box representation encoding voting strategy and a centrosymmetric method in spherical space. Weights are then refined in cylindrical space by combining density and distance weight allocation strategies. This constructs a shape learning network to improve the robustness of the detection network.

🎯Benefits of technology

This improved the detection performance of the 3D object detector for occluded objects, enabled end-to-end training, and enhanced the generalization ability and detection accuracy of the detection network.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
Patent Text Reader

Abstract

The application discloses a three-dimensional target detection method based on generation and refining of occlusion representation, which comprises the following steps: generating initial occlusion representation by adopting a candidate box lower representation coding voting strategy and a center-symmetry method based on object centroid in a spherical space; constructing a shape learning network based on representation movement to generate occlusion representation similar to a prototype target; assigning weight to each representation according to prior knowledge of density and distance based on the completed occlusion representation, refining the occlusion representation with higher weight in a cylindrical space, and weighting the weight to a feature channel generated by the occlusion representation in the detection process; and generating high-quality object features through optimization iteration of the detection network and the shape learning network. The application is suitable for traditional three-dimensional target detection networks, and the method based on generation and refining of occlusion representation can improve the detection performance of the traditional three-dimensional target detector for severely occluded objects.
Need to check novelty before this filing date? Find Prior Art