Detection method and terminal for automobile glass surface and trademark printing defects

Abstract

The invention discloses an automobile glass surface and trademark printing defect detection method and a terminal, and belongs to the technical field of model detection, and the method comprises the following steps: 1, preparing in advance: preparing acquisition equipment which needs to detect automobile glass surface and trademark printing defects in advance, and carrying out calibration and other work; the safe use can be ensured; step 2, information acquisition: acquiring images of the automobile glass by using an acquisition device, acquiring the images for multiple times from multiple angles, and numbering and recording the acquired image information; and step 3, information processing: frame processing is carried out on the obtained multiple pieces of image information to obtain the edge of the image information, and then primary cutting is carried out to obtain accurate image information. According to the method and the terminal for detecting the automobile glass surface and trademark printing defects, acquisition, processing, comparison and analysis can be carried out, and the detection accuracy is ensured.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of model detection, and in particular to a detection method and terminal for surface and trademark printing defects of automobile glass. BACKGROUND

[0002] The automobile glass production industry has very high requirements for product quality, and the quality of trademark printing is one of the important factors. During the production of automobile curved glass, more than ten kinds of defects may occur on the surface, and various printing defects may also exist in the printed area of the trademark. Therefore, detection equipment needs to be used for detection, but there is no comparative analysis capability, which affects the accuracy of modeling.

[0003] After searching, a detection method and terminal for surface and trademark printing defects of automobile glass are found in Chinese patent document (authorized announcement number CN119180820B). A preset number of collection devices are numbered, and the collection devices collect images of the automobile glass to be tested according to a preset range, obtain a preset number of mother images and number them. Each mother image is cropped a preset number of times, and the cropped sub-images are detected by a pre-trained detection model. The mother images are spliced according to the number of each mother image and the position coordinates to obtain the original image of the automobile glass to be tested. The trademark image is extracted by a preset trademark extraction algorithm, and is matched with a preset accurate template and a defect template to obtain trademark defect information. The surface defect information and the trademark defect information are summarized and output and displayed on the original image of the automobile glass to be tested, realizing accurate and efficient detection of the automobile glass. The above-mentioned patent detection equipment can perform multiple and multi-angle detection, but has no comparative analysis capability, which affects the accuracy of modeling. SUMMARY

[0004] The purpose of the present application is to provide a detection method and terminal for surface and trademark printing defects of automobile glass to solve the problems raised in the background art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solution: a detection method for surface and trademark printing defects of automobile glass, comprising the following steps: Step 1, preliminary preparation: the collection device for detecting the surface and trademark printing defects of automobile glass is prepared in advance and calibrated to ensure safe use; Step 2, information acquisition: the collection device collects images of the automobile glass from multiple angles and multiple times, and records the collected image information with numbers; Step three, information processing: the obtained multiple image information is subjected to frame processing to obtain the edges of the image information, and then subjected to preliminary cropping to obtain accurate image information; Step four, information comparison: the obtained multiple accurate image information is subjected to superimposition comparison, and several image information with large differences are eliminated to obtain parameter information of the image information and complete database modeling; Step five, information analysis: the obtained image information to be detected is placed into the modeled database for information analysis, if the image information is intact, the parameter information is consistent, if the image information has deviation, the deviated parameter information is analyzed and marked, and then the image information is recorded into the database.

[0006] As a preferred embodiment, the collection device comprises an information collection module, an information cropping module, an information modeling module and an information analysis module.

[0007] As a preferred embodiment, the information collection module comprises: a collection unit for image collection; a numbering unit for sequentially numbering the collected image information and recording the collection date.

[0008] As a preferred embodiment, the information cropping module comprises: a scanning unit for scanning the frame of the image information to facilitate subsequent cropping; a cropping unit for reducing the size of the image information after scanning the frame to facilitate subsequent modeling and information analysis and improve analysis accuracy.

[0009] As a preferred embodiment, the information modeling module comprises: a comparison unit for comparing the image information two by two to obtain comparison results, obtaining image information of moderate position according to the comparison results, and marking image information of two side positions; a modeling unit for modeling the image information according to the comparison results to form a database.

[0010] As a preferred embodiment, the information analysis module comprises: an analysis unit for placing the image information to be detected into the information modeling module for analysis to obtain analysis results; a storage unit for storing the analysis results and the image information to train, verify and update the database.

[0011] A terminal for detecting defects of a surface and a trademark printed on an automobile glass, comprising a memory, a processor and a computer program stored on the memory and executable on the processor.

[0012] Compared with the prior art, the technical effects and advantages of the present application are: The automobile glass surface and trademark printing defect detection method and terminal can number image information, then crop the image information, reduce the burden of subsequent processing, and improve the accuracy of detection, thanks to the design of information acquisition and information processing. The automobile glass surface and trademark printing defect detection method and terminal can model information comparison, eliminate image information with large differences, reduce the influence degree, improve the accuracy of subsequent detection, and store information for comparison and analysis, so that the database of modeling is trained, and the accuracy of subsequent detection is improved, thanks to the design of information comparison and information analysis. The automobile glass surface and trademark printing defect detection method and terminal can acquire, process, compare, and analyze to ensure the accuracy of detection. BRIEF DESCRIPTION OF DRAWINGS

[0013] Fig. 1 The flowchart of the present application; Fig. 2 The framework diagram of the present application. DETAILED DESCRIPTION

[0014] In the following description, a large number of specific details are given in order to provide a more thorough understanding of the present application. However, it is obvious to those skilled in the art that the present application can be implemented without one or more of these details. In other examples, some technical features known in the art are not described in order not to obscure the present application.

[0015] Please refer to Figs. 1-2 An automobile glass surface and trademark printing defect detection method, the embodiment includes the following steps: Step one, preliminary preparation: the collection equipment for detecting the automobile glass surface and trademark printing defects is prepared in advance, and calibration and other work are performed. The core parameters of calibration include the focal length, exposure, white balance, angle deviation, light intensity, etc. of the collection equipment. The car glass surface is easy to reflect light, and the curvature of different models of glass is different. If calibration is not targeted at these characteristics, the image brightness, clarity, and distortion degree collected at different batches / angles will be inconsistent, which will directly affect the accuracy of subsequent comparison and analysis, and ensure safe use. Step two, information acquisition: the image of the automobile glass is collected by the collection device, and it is collected multiple times from multiple angles, and the angle range (such as 0°-90°, covering the glass edge, trademark dead angle) and the number of times threshold (such as 3 times, 5 times) are determined. For example, the edge defect of the curved glass may only be visible at a specific angle, so specific angle observation is performed; too many times will increase redundant calculation and reduce efficiency, so the number of times is selected according to actual needs, and the collected image information is numbered and recorded, and the key environmental parameters (such as light intensity, glass temperature, and device working state at the time of collection) are recorded. Step three, information processing: the obtained multiple image information is subjected to frame processing to obtain the edge of the image information, and then subjected to initial cutting to obtain accurate image information, and the edge extraction algorithm (such as Canny operator, sobel operator, etc.) is used for specific operation in complex scenes (such as glass surface with dust, trademark printing color close to background); the cutting reference (such as glass edge or trademark outer frame) is determined to avoid missing defect areas (such as small edge collapse of glass edge), image denoising (such as Gaussian noise, salt and pepper noise processing), reflection elimination (common problem on glass surface), contrast enhancement and other preprocessing steps. If the original image has reflection spots or noise, it will cause the edge extraction to be fuzzy and the contrast to have false deviation, affecting the subsequent analysis accuracy; Step four, information comparison: the obtained multiple accurate image information is compared, and several image information with large difference is removed, and the quantitative index of the difference (such as pixel deviation threshold, structural similarity SSIM threshold, feature point matching rate, etc.). Subjective judgment may cause effective images to be mistakenly deleted (such as images with special angles but containing real defects), or abnormal images (such as blurred images) to be retained, which may ultimately affect the accuracy of the database modeling reference, so the removal standard is selected according to actual needs to obtain the parameter information of the image information and complete the database modeling; Step five, information analysis: the image information to be detected is put into the modeled database for information analysis. If the image information is complete, the parameter information is consistent, if the image information has deviation, the parameter information of the deviation is analyzed and marked, and then the image information is recorded in the database.

[0016] The whole scheme is an automatic process. In actual detection, complex defects (such as edge mixed defects and slight printing color difference) may be misjudged by the algorithm, so artificial confirmation is performed regularly to ensure the reliability of the detection.

[0017] The collection device includes an information collection module, an information cropping module, an information modeling module, and an information analysis module, and the types of the collection device (such as an industrial camera, a linear array camera, an infrared camera, etc.) and the adaptation scene (such as the collection difference of curved glass and flat glass) are different. The resolution, frame rate, and anti-glare ability of different devices are different, and targeted selection is performed.

[0018] The information collection module includes: A collection unit for image collection; A numbering unit for sequentially numbering the collected image information and recording the collection date.

[0019] The information cropping module includes: A scanning unit for scanning the frame of the image information to facilitate subsequent cropping; A cropping unit for reducing the size of the image information after scanning the frame to facilitate subsequent modeling and information analysis and improve analysis accuracy.

[0020] The information modeling module includes: A comparison unit for comparing the image information pairwise to obtain a comparison result, obtaining the image information of the moderate position according to the comparison result, and marking the image information of the two side positions; A modeling unit for modeling the image information according to the comparison result to form a database.

[0021] As a preferred embodiment, the information analysis module includes: An analysis unit for placing the image information to be detected into the information modeling module for analysis to obtain an analysis result; A storage unit for storing the analysis result and the image information to train, verify, and update the database.

[0022] A terminal for detecting defects of automobile glass surface and trademark printing, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor.

[0023] It should be noted that in this document, relational terms such as one and two and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any such actual relationship or order between such entities or actions.

[0024] Although embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for detecting defects in the surface of automotive glass and in the printing of trademarks, characterized in that: Includes the following steps: Step 1: Preparation: Prepare the data collection equipment in advance for detecting defects on the surface of automotive glass and in the printing of trademarks, and perform calibration to ensure safe use. Step 2, Information Acquisition: The acquisition device acquires images of the car window from multiple angles and records the acquired image information by numbering. Step 3, Information Processing: Perform border processing on the obtained multiple image information to obtain the edges of the image information, and then perform initial cropping to obtain accurate image information; Step 4: Information Comparison: Compare the multiple accurate image information obtained, remove the few images with large discrepancies, obtain the parameter information of the image information, and complete the database modeling; Step 5: Information Analysis: Place the acquired image information to be detected into the pre-modeled database for information analysis. If the image information is intact, the parameter information matches. If the image information has deviations, analyze the deviation parameters and mark them. Then, record the image information into the database.

2. The method for detecting defects on the surface of automotive glass and in the printing of trademarks according to claim 1, characterized in that: The data acquisition device includes an information acquisition module, an information cropping module, an information modeling module, and an information analysis module.

3. The method for detecting defects on the surface of automotive glass and in the printing of trademarks according to claim 2, characterized in that: The information collection module includes: The acquisition unit is used for image acquisition. The numbering unit sequentially numbers the acquired image information and records the acquisition date.

4. The method for detecting defects on the surface of automotive glass and in the printing of trademarks according to claim 2, characterized in that: The information cropping module includes: The scanning unit is used to scan the borders of the image information to facilitate subsequent cropping; The cropping unit is used to reduce the size of the image information after the scanning border is completed, which facilitates subsequent modeling and information analysis and improves the accuracy of the analysis.

5. The method for detecting defects on the surface of automotive glass and in the printing of trademarks according to claim 2, characterized in that: The information modeling module includes: The comparison unit compares the image information in pairs to obtain the comparison results. Based on the comparison results, it obtains the image information of the median position and marks the image information of the two side positions. The modeling unit models the image information based on the comparison results, forming a database.

6. The method for detecting defects on the surface of automotive glass and in the printing of trademarks according to claim 1, characterized in that: The information analysis module includes: The analysis unit takes the image information to be detected and puts it into the information modeling module for analysis to obtain the analysis results. The storage unit stores the analysis results and image information, enabling the database to be trained, validated, and updated.

7. A detection terminal for defects in the surface of automotive glass and trademark printing, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps in the method for detecting defects on the surface of automotive glass and in the printing of trademarks as described in any one of claims 1-8.

Images