KD High-Precision Assembly and Quality Traceability Management System Based on Visual Recognition and AI
By constructing a dual-mode feature separation mechanism for the surface normal vector distribution field, the problems of inaccurate assembly path planning and easy damage to external labels are solved, realizing high-precision assembly and label-free traceability, and improving the reliability of assembly accuracy and quality traceability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- XINJIANG FENGHAO ZHIXING AUTOMOBILE TECHNOLOGY CO LTD
- Filing Date
- 2026-02-11
- Publication Date
- 2026-05-26
AI Technical Summary
Existing technologies cannot effectively distinguish between workpiece position deviations and body flexibility deformation, resulting in inaccurate assembly path planning. Furthermore, external labels are easily damaged or fail in complex industrial environments, making it impossible to achieve high-reliability body quality traceability.
By constructing a dual-mode feature separation mechanism based on the distribution field of surface normal vectors, low-pass smoothing filtering is used to extract manifold surfaces and construct a flexible deformation compensation model. Combined with high-pass residual screening, random high-frequency normal perturbation data is extracted to generate microscopic quality fingerprints, thereby realizing flexible assembly and traceability without external tags.
It achieves simultaneous decoupling of macroscopic geometric deformation and microscopic surface texture in a single optical scan, improving assembly accuracy and generating unique microscopic fingerprint features, ensuring high reliability of assembly accuracy and quality traceability.
Smart Images

Figure CN122089153A_ABST