Crack image generation method based on cyclic consistency constraint and diffusion model

By generating realistic crack images through cyclic consistency constraints and diffusion models, the problems of data scarcity and poor model adaptability in surface crack identification are solved, and efficient crack identification in complex geological scenarios is achieved.

CN122244212APending Publication Date: 2026-06-19SOUTHWEST JIAOTONG UNIV
0 Cites 0 Cited by

Patent Information

Application Number
CN202610328976.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-18
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

In existing technologies, methods for expanding datasets for surface crack identification suffer from problems such as data scarcity, high annotation costs, and poor model robustness and adaptability. In particular, it is difficult to generate realistic and diverse images of natural bare ground cracks in complex geological scenarios, resulting in poor training performance of deep learning models.

Method used

A method based on cyclic consistency constraints and a diffusion model is adopted. The encoder maps the image to a low-dimensional latent space, constructs a bidirectional diffusion generation framework, and uses cyclic consistency constraints and adversarial loss function to train the generator to generate realistic crack images.

Benefits of technology

With small sample sizes and low costs, intelligent and controllable realistic crack datasets are generated, which improves the crack recognition performance of deep learning models in complex geological scenarios and solves the problems of data scarcity and poor model adaptability.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a crack image generation method based on cyclic consistency constraint and a diffusion model. The method comprises the steps of obtaining a training image set, a target image and a corresponding mask; mapping the image to a low-dimensional potential space through an encoder to obtain a training potential representation and a source potential representation, and performing down-sampling on the mask to obtain a corresponding training potential mask and a potential mask; a bidirectional diffusion generation framework comprising a first generation process and a second generation process is constructed, the first generation process is used for synthesizing the crack-free image into a crack image, and the second generation process is used for repairing the crack image into the crack-free image; training the bidirectional diffusion generation framework in the low-dimensional potential space by using cyclic consistency constraint; and performing iterative denoising through a target generation process determined in the trained bidirectional diffusion generation framework to generate a target potential representation, and performing reconstruction through a decoder to obtain a crack image. According to the method, the high-reality crack image is generated without pairing data.
Need to check novelty before this filing date? Find Prior Art