Coral Soil Particle Tracking With 3D Breakage Observation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods like triaxial testing and computer vision techniques fail to accurately track and observe particle breakage in coral soil, leading to incomplete understanding of its mechanical properties due to limitations in materials and particle-scale behavior.
Innovation Solution
A particle tracking method based on a multimodal model using a transparent rubber triaxial membrane, binocular vision, bilinear interpolation, and optimized BoT-SORT algorithm for real-time tracking and segmentation of coral soil particles, incorporating self-attention and feature enhancement for improved accuracy and zero-shot recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional triaxial testing with conventional materials is used, then mechanical properties can be measured, but particle-scale behavior and breakage processes cannot be observed
Solution Approach 1:
The patent applies transparency changes to the triaxial membrane material, transitioning from opaque conventional materials to transparent rubber membrane. This allows light to pass through the membrane, enabling optical observation and imaging of particle-scale behavior and breakage processes during triaxial testing, thereby resolving the contradiction between measurement precision and device complexity
Solution Approach 2:
The patent introduces transparent membrane material as an intermediary between the traditional testing system and the observation system. This intermediary allows simultaneous mechanical testing and optical observation without interfering with the mechanical properties, enabling particle-scale measurement while maintaining the integrity of the triaxial testing function
2Measurement precision
If PIV technique is used, then overall displacement can be calculated, but individual particle tracking is not possible
Solution Approach 1:
The patent segments the continuous displacement field into discrete particle-level measurements by using image processing techniques to identify and track individual particles. This segmentation allows each particle to be tracked separately, providing individual particle tracking capability while maintaining processing efficiency through automated image analysis algorithms
Solution Approach 2:
The patent transitions from region-based displacement measurement to particle-based tracking by adding the dimension of particle identification. Through binocular vision and image processing, the system extracts particle positions, shapes, and trajectories, enabling individual particle tracking while maintaining productivity through efficient computational methods
3Measurement precision
If PTV technique is used, then individual particle tracking is achieved, but particle fragmentation detection is not possible
Solution Approach 1:
The patent applies preliminary action by pre-training deep learning models with large datasets of particle images before actual tracking. This pre-training enables the system to recognize particle boundaries, shapes, and fragmentation patterns automatically during tracking, achieving breakage detection without increasing real-time system complexity
Solution Approach 2:
The patent uses parameter changes in particle morphology (shape factors, area, perimeter) detected through image processing to identify fragmentation events. By monitoring changes in these parameters over time, the system can detect particle breakage automatically, enhancing measurement precision while managing device complexity through software-based detection
4Measurement precision
If deep learning models are trained with large datasets, then recognition accuracy improves, but training time and computational resources increase
Solution Approach 1:
The patent applies preliminary action by pre-training deep learning models offline with large datasets before deployment. This allows the model to learn complex particle patterns and achieve high recognition accuracy in advance, so that during actual particle tracking, only inference is needed, significantly reducing real-time computational requirements and training time
Solution Approach 2:
The patent uses synthetic particle images and simulated data to create training datasets that replicate real-world conditions. By copying and augmenting limited real data with synthetic data, the model achieves high accuracy without requiring extensive collection of real particle images, reducing both data collection time and training time while maintaining recognition precision
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
Enhances the speed and accuracy of coral soil particle segmentation, reduces tracking errors, and provides comprehensive observation of particle breakage mechanisms without requiring specialized model training.
Implementation Method 1
Obtaining a three-dimensional cloud image of surface particles by three-dimensional reconstruction of the images taken in S2 through a binocular vision algorithm
Data Source
AI summary
The present disclosure belongs to the field of particle tracking technology of coral particles in soil, and specifically discloses a particle tracking method for coral particles in soil based on a multimodal model. The method comprises a particle tracking method that enables counting the coral soil particles in an image by using an improved small sample counting model. Coral soil particles in the figures are segmented, and contour coordinates and morphological characteristics of coral soil particles are obtained. Multiple images are put into an optimized BoT-SORT algorithm, and a three-dimensional motion trajectory of the surface coral soil particles is obtained during the test. This method allows for comprehensive observation throughout a triaxial test, and it not only increases the speed and accuracy of coral soil particle segmentation and reduces the tracking error rate, but also improves overall precision.