Evaporator defect detection method and system based on deep learning

By jointly modeling multi-view image data and using a deep learning target detection model, multi-scale features of the evaporator are extracted and consistency comparisons are performed. This solves the problem of unstable defect identification under changes in evaporator structural parameters in existing technologies and achieves high-precision error-proof assembly inspection.

CN122265173APending Publication Date: 2026-06-23HEFEI JUQUE ELECTRONICS CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202610293768.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-11
Publication Date
2026-06-23

AI Technical Summary

Technical Problem

Existing technologies cannot reliably identify the inherent relationships between various structural parameters during evaporator assembly, leading to misjudgments and poor detection stability, making it difficult to adapt to the flexible production needs of multiple specifications and batches.

Method used

Multi-view image data joint modeling is adopted. Multi-scale features such as the number of septa, arrangement direction and spatial location of gaps are extracted by deep learning target detection model. The features are recombined using a hybrid attention mechanism. Finally, the consistency is compared with the preset structural parameters to determine the consistency of the evaporator structure.

Benefits of technology

It achieves stable identification of evaporator defects, avoids misjudgment, improves the stability and versatility of detection, and is suitable for automated production detection of products with multiple specifications.

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

Abstract

The invention provides an evaporator defect detection method and system based on deep learning, and relates to the field of anomaly detection. The method comprises the following steps: acquiring multi-view image data of an evaporator, and extracting multi-scale features including spacer number features, arrangement direction features and gap space position information through a backbone network layer of a deep learning target detection model; performing nonlinear enhancement on the multi-scale features through a feature enhancement network layer of the model to obtain enhanced multi-scale features, and extracting recombination features through a mixed attention mechanism layer; and based on the recombination features, through the detection layer of the model, obtaining detection results of the evaporator spacer and the gap, and comparing the detection results with preset structure parameters to obtain defect detection results. The method and device are used in the evaporator defect detection process, and the technical problem that in the prior art, under the condition that the environment and evaporator structure parameters change, the evaporator defects cannot be stably recognized is solved.
Need to check novelty before this filing date? Find Prior Art