Depth-Image Material Inspection for Fast Irregular Volume Grading

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional methods for grading irregularly shaped materials, such as farm produce, are inefficient and inaccurate, and existing 3D model-based volume calculation techniques are time-consuming and resource-intensive, especially for complex geometric structures.

Innovation Solution

A material inspection system that directly extracts depth information and uses a deep learning model to calculate volume, eliminating the need for 3D model construction, and includes components like a conveying portion, image-taking portion, and a volume calculation module with a deep learning model to process depth images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional 3D model construction methods are used to calculate volume, then volume calculation can be achieved, but computation time and resource usage are excessive

Engineering Contradiction:

Improvevolume calculation accuracyVSAvoidcomputation time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts only the essential depth information from depth images and feeds it directly into a deep learning model for volume calculation, eliminating the time-consuming 3D model construction process while maintaining measurement accuracy

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent replaces the traditional mechanical 3D model construction approach with a deep learning-based direct calculation system that processes depth images to obtain volume results without intermediate model building steps

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If conventional 3D model construction methods are used to calculate volume, then volume calculation can be achieved, but computation resources are excessively consumed

Engineering Contradiction:

Improvevolume calculation accuracyVSAvoidcomputation resource usage

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent extracts only the necessary depth information from images and feeds it directly into a deep learning model, eliminating resource-intensive 3D model construction while preserving calculation accuracy

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent uses depth images as direct input copies for volume calculation through deep learning, replacing the need to create and process full 3D model representations, thereby reducing computational resource requirements

Inventive Principle:

Principle #26Copying

3Productivity

If fixed-sized holes are used for grading materials, then grading can be performed, but accuracy is poor for irregularly shaped materials

Engineering Contradiction:

Improvegrading efficiencyVSAvoidgrading accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent replaces mechanical fixed-hole grading methods with an image processing and deep learning system that analyzes depth information to accurately determine material volume and grading category, achieving high accuracy for irregularly shaped materials

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent changes the grading parameter from simple dimensional measurement through fixed holes to complex volume calculation based on depth image analysis and deep learning, enabling accurate grading of irregularly shaped materials

Inventive Principle:

Principle #35Parameter changes

4Adaptability or versatility

If variable-gap grading machines are used to improve grading flexibility, then adjustable grading is possible, but operational complexity and downtime increase

Engineering Contradiction:

Improvegrading adjustment flexibilityVSAvoidmachine operation complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent replaces mechanical variable-gap adjustment mechanisms with a digital image processing and deep learning system that automatically adapts to different material types and grading requirements through software algorithms, eliminating mechanical complexity

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent implements dynamic adaptability through deep learning models that can be trained and adjusted for different material types and grading criteria, providing flexibility without mechanical moving parts or manual adjustments

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS20260004445A1Material inspection system

Publication Date: 2026.01.01 FENG CHIA UNIVERSITY

AI summary

A material inspection system includes a supporting structure, a conveying portion, an image-taking portion, a driving portion, and a volume calculation module. The conveying portion is provided on the supporting structure and can convey a material from a first end to a second end of the conveying portion. The image-taking portion has an imaging lens whose field of view encompasses at least a portion of the moving path of the material on the conveying portion. The driving portion is located on the conveying portion and can drive the material into rotation so that the image-taking portion can take a plurality of depth images of different portions of the material while the material is in rotation. The volume calculation module extracts the depth information in each depth image and feeds the depth information into a deep learning model to calculate the volume of the material.