CN-121981954-A - Bipolar plate flatness detection method based on image moire patterns

Abstract

The invention relates to a bipolar plate flatness detection method based on image mole patterns, which is characterized by comprising the following steps of S1, setting image acquisition conditions, S2, mole pattern image acquisition and preprocessing, S3, mole pattern feature enhancement and extraction, namely separating mole patterns from the background by adopting a multi-segment threshold segmentation method to form a binary feature image, S4, warp analysis and result output, namely carrying out qualitative and/or quantitative determination of surface flatness based on the extracted mole pattern morphological features. Compared with the prior art, the invention has the advantages of high sensitivity, low cost, high efficiency, workpiece damage prevention, reliable result, comprehensive analysis dimension and the like.

Inventors

• ZHAO LIYAN
• XU CHENLIANG
• LIN SHAOYU
• LIU WANYING
• CHEN YANXI
• Christopher Ayman
• ZHANG WEIMIN

Assignees

• 同济大学

Dates

Publication Date

20260505

Application Date

20251225

Claims (10)

1. 1. The bipolar plate flatness detection method based on the image moire patterns is characterized by comprising the following steps of: The method comprises the following steps of S1, setting image acquisition conditions, namely regulating and controlling the relative position between an imaging system and a bipolar plate to be detected, controlling the included angle between the optical axis of a camera and the normal line of the bipolar plate to be 30-60 degrees, controlling the shooting distance to be 250-600 mm, and providing illumination so as to excite the periodic flow path structure of the bipolar plate and the pixel grid interference of an image sensor to generate Moire patterns sensitive to surface warpage; s2, acquiring and preprocessing a moire image, namely acquiring an original digital image of a bipolar plate flow channel region under the imaging condition of S1, and preprocessing the image by resolution scaling and contrast enhancement to strengthen the moire characteristic; S3, mole pattern feature enhancement and extraction, namely separating mole patterns from the background by adopting a multi-segment threshold segmentation method on the preprocessed image to form a binarized feature image; and S4, warp analysis and result output, namely, carrying out qualitative and/or quantitative judgment on the surface flatness based on the extracted mole pattern morphological characteristics, and outputting a detection result.
2. 2. The method for detecting flatness of bipolar plates based on image moire patterns according to claim 1, wherein in S1, an included angle between an optical axis of the camera and a normal line of the bipolar plate is 45 °.
3. 3. The method for detecting flatness of bipolar plates based on moire patterns according to claim 1, wherein in S2, the resolution scaling is to scale the original image to 0.10 to 0.90 times the original resolution using the CUBIC interpolation algorithm to find the scale with the most significant moire patterns.
4. 4. The method for detecting flatness of bipolar plates based on moire patterns according to claim 1, wherein in S2, contrast enhancement is to enhance contrast between moire pattern area and background by image processing algorithm.
5. 5. The method for detecting flatness of bipolar plates based on image moire patterns according to claim 1, wherein in S4, the qualitative determination is to match the moire pattern features with a pre-stored library of warp patterns to determine the type and position of warp.
6. 6. The method for detecting flatness of bipolar plates based on image moire as claimed in claim 5, wherein the type of warpage includes protrusions and depressions.
7. 7. The method for detecting flatness of bipolar plates based on image moire according to claim 1, wherein in S4, the quantitative evaluation is performed by measuring geometric parameters of moire features and calculating the height or depth of warpage according to a pre-calibrated conversion relationship.
8. 8. The method for detecting flatness of bipolar plates based on image moire patterns according to claim 7, wherein the geometric parameters include the long-short axial ratio of moire elliptical rings, and the curvature radius of the stripes.
9. 9. The method for detecting flatness of bipolar plates based on image moire patterns according to claim 1, further comprising a partition analysis step of dividing a bipolar plate flow channel area into a plurality of analysis subareas and calibrating moire pattern features in combination with three-dimensional scanning data before the step S4.
10. 10. The method for detecting the flatness of a bipolar plate based on image moire patterns according to claim 1, wherein the method is integrated in a visual detection system, can automatically extract moire pattern features through an image processing algorithm and output flatness interpretation results so as to support on-line detection and quality control of a production line.

Description

Bipolar plate flatness detection method based on image moire patterns Technical Field The invention relates to the technical field of bipolar plate detection, in particular to a bipolar plate flatness detection method based on image moire patterns. Background The fuel cell is increasingly used in the fields of transportation, stationary power supply, etc. The fuel cell stack is formed by alternately stacking a Membrane Electrode Assembly (MEA) and a Bipolar Plate (BPP), wherein the Bipolar Plate performs the core functions of gas distribution, electrical conduction, heat dissipation, support and the like, and the dimensional stability and surface flatness of the Bipolar Plate directly affect the sealing performance and the reaction efficiency of the stack. With the development of the fuel cell to the high power and large area, the large-breadth bipolar plate is more prone to appearance defects such as warping and sinking in the processing, carrying and storing processes, so that the assembly is difficult, the flow channels are not aligned or the local sealing is invalid. Therefore, it is important to quickly and accurately detect the surface flatness of the bipolar plates before the galvanic pile assembly. The currently commonly used bipolar plate flatness detection mode mainly comprises the following steps: (1) Contact detection methods (e.g., dial gauge, feeler gauge). The method is easy to generate contact rebound or indentation, has large influence of human factors, is difficult to realize high-precision repeated detection, and is not suitable for flexible or thin bipolar plates. (2) A laser profile scanner. The three-dimensional morphology with higher precision can be obtained, but the scanning speed is low, the equipment cost is high, and the batch detection of the production line is not facilitated. (3) Common visual inspection methods. Based on feature point or edge extraction, bipolar plate has dense flow paths, strong light reflection and high structure periodicity, and is difficult to directly extract warp information from a common image. However, bipolar plate flow channel regions have significant periodic geometric features, and frequency differences from the pixel grid of digital image sensors can naturally produce moire patterns in the image. Moire is very sensitive to micro deformation, and surface warpage can be amplified and presented in the form of fringe bending. The detection method based on the moire has the advantages of non-contact, low cost, high speed, simplicity in imaging and the like, but a special detection method for the flow channel structure and the warping characteristic of the bipolar plate is lacking at present. Therefore, it is needed to provide a rapid detection method for flatness of a bipolar plate based on moire patterns, which realizes visualization and quantitative analysis of the warpage pattern and degree of the bipolar plate by controlling shooting parameters and combining an image processing algorithm, so as to meet the requirement of the fuel cell manufacturing industry on high-efficiency quality detection. Disclosure of Invention The invention aims to overcome the defects of the prior art, and provides the non-contact detection method which has the advantages of high sensitivity, low cost, high efficiency, strong pertinence, strong detection robustness, reliable result and comprehensive analysis dimension, and can avoid workpiece damage. The aim of the invention can be achieved by the following technical scheme: the invention provides a bipolar plate flatness detection method based on image mole patterns, which comprises the following steps: the method comprises the following steps of S1, setting image acquisition conditions, namely regulating and controlling the relative position between an imaging system and a bipolar plate to be detected, controlling the included angle between the optical axis of a camera and the normal line of the bipolar plate to be 30-60 degrees, controlling the shooting distance to be 250-600 mm, and providing uniform and stable illumination so as to excite the periodic flow channel structure of the bipolar plate to interfere with the pixel grid of an image sensor to generate Moire patterns sensitive to surface warpage; s2, acquiring and preprocessing a moire image, namely acquiring an original digital image of a bipolar plate flow channel region under the imaging condition of S1, and preprocessing the image by resolution scaling and contrast enhancement to strengthen the moire characteristic; S3, mole pattern feature enhancement and extraction, namely separating mole patterns from the background by adopting a multi-segment threshold segmentation method on the preprocessed image to form a binarized feature image; and S4, warp analysis and result output, namely, carrying out qualitative and/or quantitative judgment on the surface flatness based on the extracted mole pattern morphological characteristics, and outputting a detection resul