Search

CN-121998932-A - Device and method for monitoring state of laser shock enhanced absorption layer

CN121998932ACN 121998932 ACN121998932 ACN 121998932ACN-121998932-A

Abstract

The invention discloses a device and a method for monitoring the state of a laser shock enhanced absorption layer, which relate to the technical field of surface enhancement and comprise a visual image sensor array, an image preprocessing module, an image analysis module and a state evaluation module; the system comprises a visual image sensor array, an image preprocessing module, an image analysis module and a state evaluation module, wherein the visual image sensor array acquires visible spectrum images and near infrared band images of a laser impact strengthening processing area at multiple angles, the image preprocessing module preprocesses the acquired images to obtain preprocessed images at different angles, the image analysis module is provided with an image recognition model based on ResNet-UNet mixed architecture and attention mechanism, the preprocessed images are recognized to obtain image analysis results, and the state evaluation module evaluates the images according to the image analysis results by adopting a multi-threshold judgment mechanism to obtain evaluation results. The invention can monitor the change of the absorption layer state in the laser shock peening process in real time, discover the damage condition in time and effectively improve the stability and consistency of the laser shock peening processing.

Inventors

  • GUO WEI
  • SHI JIAXIN
  • ZHANG HONGQIANG

Assignees

  • 北京航空航天大学

Dates

Publication Date
20260508
Application Date
20260121

Claims (10)

  1. 1. The device for monitoring the state of the laser shock enhanced absorption layer is characterized by comprising a visual image sensor array, an image preprocessing module, an image analysis module and a state evaluation module; the visual image sensor array is used for collecting visible spectrum images and near infrared band images of the laser shock peening processing area at multiple angles; The image preprocessing module is used for preprocessing the acquired images to obtain preprocessed images with different angles; The image analysis module is deployed with an image recognition model based on ResNet-UNet mixed architecture and an attention mechanism, and recognizes the preprocessed image to obtain an image analysis result, wherein the image recognition model comprises an encoder, a self-adaptive feature fusion module, an attention module and a decoder, the encoder extracts multi-level features, the self-adaptive feature fusion module generates multi-view multi-band fusion features according to multi-level feature fusion, the attention module weights and fuses the multi-view multi-band fusion features to obtain a comprehensive fusion feature, and the decoder performs pixel level classification recognition on the multi-class typical damage features of an absorption layer on the comprehensive fusion feature to obtain an image analysis result; And the state evaluation module is used for evaluating by adopting a multi-threshold judgment mechanism according to the image analysis result to obtain an evaluation result.
  2. 2. The apparatus for monitoring the state of a laser shock peening absorption layer according to claim 1, wherein the visual image sensor array comprises a plurality of high-speed industrial cameras which are respectively installed in front of a laser shock peening processing area and distributed in a conical shape, and collect visible spectrum images of different angles and near infrared band images of different bands.
  3. 3. The apparatus for monitoring the state of a laser shock peening absorption layer according to claim 1, wherein the encoder in the image recognition model extracts multi-level features from low-level texture to high-level semantics from the preprocessed image using residual connection; The self-adaptive feature fusion module is used for respectively calculating weights of the multi-level features extracted from the preprocessed image corresponding to the visible spectrum image and the multi-level features extracted from the preprocessed image corresponding to the near infrared band image, and carrying out weighted fusion on the two groups of multi-level features according to the weights to obtain multi-view multi-band fusion features; The attention module comprises a space attention layer, a channel attention layer and a multi-scale fusion layer, wherein the space attention layer learns the importance weight of the multi-view multi-band fusion feature at the space position; The decoder restores the spatial resolution through up-sampling and jump connection, performs pixel level segmentation on the comprehensive fusion features, and identifies multiple types of typical damage features.
  4. 4. The apparatus for monitoring the state of a laser shock peening absorbing layer according to claim 1, wherein the image preprocessing module uses a hardware acceleration architecture based on FPGA and DSP chips, and the preprocessing operation includes improved bilateral filtering denoising, adaptive histogram equalization enhancement, geometric correction of multi-point perspective transformation, and normalization of images acquired at different angles into a unified format.
  5. 5. The apparatus of claim 1, wherein the plurality of types of typical damage features include linear features, localized ridges, irregular area deletions, and wherein the image analysis results include no apparent damage features or one or more of the plurality of types of typical damage features, and wherein the damage features correspond to parameters associated with the damage areas.
  6. 6. The apparatus for monitoring the state of a laser shock peening absorption layer according to claim 1, wherein the evaluation index is calculated based on the image analysis result, the damaged state of the absorption layer is evaluated based on the evaluation index and a multi-threshold judgment mechanism, and the obtained evaluation result includes a normal state, a slightly damaged state, and a severely damaged state.
  7. 7. An apparatus for monitoring the condition of a laser shock peening absorption layer according to claim 1, wherein the image recognition model is trained using a cross entropy loss function and an optimizer.
  8. 8. The apparatus for monitoring the state of a laser shock peening absorption layer according to claim 1, further comprising an adaptive learning module for algorithmically optimizing an image recognition model by continuously collecting processing data during laser shock peening processing using an online incremental learning strategy and a knowledge distillation technique; The system is characterized by further comprising a material identification model, wherein the material identification model comprises a material feature encoder and a classifier, the material feature encoder is used for preprocessing and feature extraction of acquired processing data and encoding spectrum and texture characteristics of different materials into feature vectors, the classifier is used for identifying the feature vectors and determining the material type of an absorption layer, the material identification model and the image identification model adopt a combined training strategy and share part of feature extraction layers, the acquired data are processed by utilizing material features and damaged features through cooperative optimization of a multi-task learning framework, and an image analysis result is optimized.
  9. 9. The apparatus for monitoring a state of a laser shock peening absorption layer according to claim 6, further comprising a feedback control module for triggering a multi-stage early warning mechanism according to the evaluation result, sending an early warning signal in a slightly damaged state, sending an early warning signal in a severely damaged state, and automatically suspending a processing flow.
  10. 10. A method for monitoring the condition of a laser shock peening absorber layer, characterized in that it is applied to a device for monitoring the condition of a laser shock peening absorber layer according to any one of claims 1 to 9, comprising the steps of: step 1, collecting a surface image of an unprocessed absorption layer as a reference image, and establishing a reference image library; step 2, continuously collecting multi-angle images of the surface of the absorption layer in the laser impact process; step 3, preprocessing the collected multi-angle images; training an image recognition model by using a reference image library, and recognizing multiple types of typical damage features in the preprocessed image by using the image recognition model to obtain an image analysis result; and 5, determining a multi-threshold judging mechanism according to the characteristics of the reference images in the reference image library, and judging the damage severity of the absorption layer by adopting the multi-threshold judging mechanism according to the image analysis result to obtain an evaluation result.

Description

Device and method for monitoring state of laser shock enhanced absorption layer Technical Field The invention relates to the technical field of surface strengthening, in particular to a device and a method for monitoring the state of a laser shock enhanced absorption layer. Background In Laser Shock Peening (LSP) processing, the type and state of an absorber layer, which is a key medium for converting laser energy into a plasma shock wave, have an important influence on the strengthening effect. In the laser shock peening process, the integrity of the absorber layer plays a decisive role in the process. When the laser energy density is too high, the pulse width is not proper or the repetition rate is not properly selected, the absorption layer may be damaged and fall off due to local overheating or too strong impact. Once the absorption layer is damaged, the laser directly acts on the surface of the material, and the surface of the material can generate severe thermal effects due to the lack of energy conversion and protection of the absorption layer, so that local melting, gasification and even ablation phenomena are caused. The method can not achieve the expected strengthening effect, but can cause the degradation of the surface quality of the material, generate microscopic defects and reduce the fatigue performance. Therefore, in actual processing, the absorption layer must be ensured to be kept in a complete state all the time by optimizing the combination of process parameters, so as to realize an ideal strengthening effect. The monitoring of the absorption layer state in the current laser shock peening process has obvious defects. The traditional manual monitoring mode is low in working efficiency, is easily influenced by subjective factors, and cannot timely discover and respond to the damage condition of the absorption layer. This hysteresis may lead to a deterioration of the surface quality of the material, resulting in an unstable processing quality and an increase in production costs. At present, the technology for detecting the damage of the absorption layer in the laser shock peening process is still in a blank state. The prior studies have focused mainly on optimization of the performance of the absorbent layer, Therefore, how to monitor the change of the absorption layer state in the laser shock peening process in real time and discover the damage condition in time is a problem that needs to be solved by those skilled in the art. Disclosure of Invention In view of the above problems, the present invention provides a device and a method for monitoring the state of a laser shock-reinforced absorption layer, which overcome or at least partially solve the above problems, and based on a multi-scale feature fusion neural network, the change of the absorption layer state in the laser shock-reinforced process can be monitored in real time, so as to discover the damage condition in time and trigger early warning, thereby effectively improving the stability and consistency of the laser shock-reinforced processing. In order to achieve the above purpose, the present invention adopts the following technical scheme: in a first aspect, an embodiment of the present invention provides an apparatus for monitoring a state of a laser shock peening absorption layer, including a visual image sensor array, an image preprocessing module, an image analysis module, and a state evaluation module; the visual image sensor array is used for collecting visible spectrum images and near infrared band images of the laser shock peening processing area at multiple angles; The image preprocessing module is used for preprocessing the acquired images to obtain preprocessed images with different angles; The image analysis module is deployed with an image recognition model based on ResNet-UNet mixed architecture and an attention mechanism, and recognizes the preprocessed image to obtain an image analysis result, wherein the image recognition model comprises an encoder, a self-adaptive feature fusion module, an attention module and a decoder, the encoder extracts multi-level features, the self-adaptive feature fusion module generates multi-view multi-band fusion features according to multi-level feature fusion, the attention module weights and fuses the multi-view multi-band fusion features to obtain a comprehensive fusion feature, and the decoder performs pixel level classification recognition on the multi-class typical damage features of an absorption layer on the comprehensive fusion feature to obtain an image analysis result; And the state evaluation module is used for evaluating by adopting a multi-threshold judgment mechanism according to the image analysis result to obtain an evaluation result. Preferably, the visual image sensor array comprises a plurality of high-speed industrial cameras which are respectively arranged in front of the laser shock peening processing area and distributed in a conical manner to form multi-angle