PCB Solder Joint 3D Reconstruction With Neural Rendering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current PCB production processes rely on costly high-precision laser sensors for 3D imaging of solder joints or rely on 2D image-based detection methods that are cost-effective but dependent on inspector experience, lacking a high-precision 3D reconstruction method using existing 2D AOI equipment.
Innovation Solution
A method utilizing neural rendering to reconstruct high-resolution 3D models of solder joints from 2D images captured by conventional AOI equipment, integrating Signed Distance Field (SDF) and Bidirectional Reflectance Distribution Function (BRDF) within a neural rendering framework to model light interaction with solder joints, without requiring additional hardware.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-precision laser sensors are used for 3D imaging of solder joints, then manufacturing precision is improved, but device cost increases
Solution Approach 1:
The patent creates a virtual 3D copy of the solder joint using neural rendering algorithms based on 2D images from conventional cameras. Instead of directly measuring 3D geometry with expensive laser sensors, the system synthesizes a realistic 3D representation by learning the mapping from 2D images to 3D structures, achieving high precision without costly hardware
Solution Approach 2:
The patent replaces the mechanical laser sensing system with a computational approach using neural networks. The physical laser measurement system is substituted by an algorithmic system that processes 2D images through trained neural rendering models to infer 3D geometry, transforming a hardware-intensive problem into a software-based solution
2Device complexity
If 2D image-based detection methods are used, then device cost is reduced, but measurement precision deteriorates
Solution Approach 1:
The patent transforms 2D image data into 3D spatial information by introducing a third dimension through neural rendering. The system takes conventional 2D images as input and generates corresponding 3D representations, effectively adding dimensional information computationally to enhance detection precision while using simple 2D imaging hardware
Solution Approach 2:
The patent changes the representation parameters from direct 2D pixel values to implicit 3D geometric parameters through neural network transformations. By learning the parameter mapping from 2D image space to 3D structure space, the system extracts precise geometric measurements that would normally require complex 3D sensing equipment
3Device complexity
If conventional 2D AOI equipment is used for 3D reconstruction, then device cost is reduced, but manufacturing precision deteriorates
Solution Approach 1:
The patent introduces neural rendering algorithms as an intermediary between conventional 2D AOI equipment and 3D reconstruction output. The neural network acts as a computational mediator that translates simple 2D images into accurate 3D representations, bridging the gap between inexpensive 2D hardware and high-precision 3D reconstruction requirements
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables high-precision 3D reconstruction of solder joints at a low cost using existing 2D AOI equipment, enhancing rendering efficiency and accuracy while utilizing conventional inspection tools.
Implementation Method 1
combines the BRDF with the calibrated incident angle to represent the light intensity of solder joint images
Data Source
AI summary
A 3D reconstruction method for PCB solder joints based on neural rendering, including using the signed distance function (SDF) to represent the solder joint surface, combining the bidirectional reflectance distribution function (BRDF) with the calibrated incident angle to represent the light intensity of solder joint images, and then unifying them into the neural rendering framework to optimize. The method uses existing solder joint inspection equipment to collect solder joint image, and reconstructs high resolution 3D reconstruction of solder joint from a single image with low cost.