Method for detecting defects of glass bottle body by using camera
By using the camera and specific detection methods in the glass bottle detection, including the bottle mouth, bottle width and black dot model detection, the existing technology is difficult to detect small defects, realized high accuracy defect detection, improved the product Qualification rate.
Patent Information
- Application Number
- CN202411825446.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-12
- Publication Date
- 2025-01-28
AI Technical Summary
It is difficult to detect small defects in existing glass bottle defect detection methods, such as the fine projection of the bottle mouth ring, shear printing, widths of different heights of bottle diameters, etc., resulting in a decrease in the pass rate of product delivery.
Methods to detect glass bottle body defects, including bottle mouth detection (protruding detection and scissors), bottle width detection, and black dot model detection. The defects on the model.
The detection accuracy of glass bottle defects has been improved, the detection effect of 100%is achieved, the flawed glass bottle flows into the market, and the grade rate of the glass bottle is increased.
Smart Images
Figure CN119375249A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of glass bottle detection methods, and in particular relates to a method for detecting defects of a glass bottle body by using a camera. Background Art
[0002] The production and manufacturing of glass bottles is to inject molten glass into the row machine, blow it through the corresponding glass production mold, and then go through various subsequent processing processes such as cooling and tempering to obtain the finished glass bottles. Due to the particularity of the production process, various problems may occur during the blowing process, resulting in uneven blowing defects, inconsistent thickness and width of the bottle body, irregular shape of the bottle body, various small flaws and other defects, which affect normal use and cannot be delivered to customers. Therefore, in the process of glass production and manufacturing, it is necessary to inspect the finished glass products and eliminate glass bottles with various defects to prevent defective glass bottles from entering the packaging stage.
[0003] The existing method of detecting defects in glass bottles generally involves taking photos of the glass bottles from multiple angles, analyzing the photos using a computer detection program, and comparing the image differences between the glass bottle and the detection frame to analyze and determine whether the glass bottle meets the standards. This method can effectively detect most defects in glass bottles. However, due to the shape of the glass bottles, especially for special shapes such as glass bottles with varying bottle diameters and ring-shaped bottle mouths, the existing detection method is difficult to detect some minor defects, such as the slight protrusion of the glass bottle mouth ring, the shear marks of the bottle mouth, and the width of the bottle diameter at different heights. It is impossible to effectively detect and eliminate them, resulting in a decrease in the qualified rate of product delivery, which causes great damage to the reputation and economic benefits of the manufacturing company. Therefore, it is necessary to conduct research based on the existing detection methods to improve the accuracy of detecting glass bottle defects. Summary of the invention
[0004] The invention provides a method for detecting defects of a glass bottle body by using a camera. The method further comprises bottle mouth detection, bottle width detection and black spot sample plate detection on the basis of an existing detection module.
[0005] Bottle mouth inspection includes two parts: protrusion inspection and scissor mark inspection. Protrusion inspection is to inspect the upper end surface of the mouth ring of the bottle mouth to determine whether the upper end of the mouth ring has a protrusion defect caused by damage to the mouth mold during the blowing process. During the inspection, a mouth ring outer frame is set for the bottle mouth inspection part in the inspection program by taking a photo from the side of the glass bottle, so that the mouth ring outer frame frames the entire mouth ring part of the bottle mouth, and keeps the upper side of the mouth ring outer frame flush with the upper edge of the inspection bottle mouth. When the inspection program finds that the upper side of the mouth ring in the photo exceeds the upper edge of the mouth ring outer frame, it is judged that the place is an abnormal point on the mouth ring of the bottle mouth and the inspection fails.
[0006] Scissors mark detection is to detect the upper end surface of the bottle mouth to determine whether there is a scissors mark defect on the upper end surface of the bottle mouth. The upper end surface of the bottle mouth is photographed from top to bottom for detection. A cross-bottle mouth line consisting of two concentric detection frames is set in the detection program to make the outer detection frame flush with the outer edge of the bottle mouth, and the inner detection frame flush with the inner edge of the bottle mouth. Then the two detection frames are detected through the detection program. When it is found that the brightness between the two detection frames is inconsistent with other areas, it is determined that there is an image abnormality between the two detection frames, and a scissors mark defect appears on the upper end of the bottle mouth, and the detection fails. In order to make the detection accuracy able to effectively eliminate defective bottles, the parameters of the cross-bottle mouth line are adjusted as follows: the sensitivity value is set to 5, the width value is 40mm, and the angle difference is 55.
[0007] Bottle width detection is to detect the overall width of the bottle body to determine whether there is any dent in the bottle body caused by insufficient blowing during the blowing process. The glass bottle is photographed from the side for detection, and 3-5 detection frames are set in the detection program. The detection frames are set at different heights to detect the width of the bottleneck, shoulder and body of the bottle respectively. When the bottle diameter changes a lot, the number of detection frames is increased to 5. Multi-segment detection is used to determine whether the glass bottle has a defect of insufficient width. When the width data detected by the detection frame does not meet the predetermined value, the glass bottle is judged to have a defect and the detection fails.
[0008] The black spot sample test is to test the detection accuracy of the detection line by sticking a black spot sample patch on the sample bottle and then putting the sample bottle into the detection line for detection. When performing black spot sample detection, it is necessary to ensure that the black spot sample can be correctly identified in the detection program without affecting the detection accuracy of the glass bottle by the detection program. Therefore, the number of black spot sample patches is set to 12, which are respectively attached to the upper, middle and lower sections of the glass bottle, and 4 are attached to each section of the circumference of the glass bottle body to ensure that the black spot sample patch can be accurately identified during the detection process. At the same time, the black spot sample patch is set as a transparent film disc style, and an opaque black spot is set in the center of the transparent film disc. In this way, the part other than the black spot can be prevented from affecting the detection accuracy, and the black spot sample patch can be tightly attached to the glass bottle.
[0009] The method for detecting defects on the body of a glass bottle using a camera provided by the present invention can accurately detect the protrusion of the mouth of the glass bottle, shear defects, body width and sample detection while ensuring the existing detection items, thereby improving the detection accuracy of the defects on the glass bottle and achieving a 100% detection effect, effectively preventing defective glass bottles from entering the market and improving the delivery yield rate of glass bottles. Through the application of the present invention, the existing detection procedures can be used to detect minor defects on glass bottles, while improving the defect detection rate without incurring additional equipment costs, which can bring better economic benefits to enterprises and can be widely promoted and applied in the industry. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] The present invention is further described in detail below with reference to the accompanying drawings.
[0011] Figure 1 This is a schematic diagram of the protrusion detection program panel; Figure 2 This is a schematic diagram of the scissors print detection program panel; Figure 3 This is a schematic diagram of the abnormal state detected by the scissors print detection program panel; Figure 4 This is a schematic diagram of the bottle width detection program panel; Figure 5 This is a schematic diagram of the black dot sample patch being pasted on the sample bottle. DETAILED DESCRIPTION
[0012] To further illustrate the concept of the present invention, the specific implementation of the present invention will be further described below in conjunction with the accompanying drawings: A method for detecting defects in glass bottles using a camera, using glass visual inspection equipment for inspection, using a high-definition camera on the equipment to collect images of glass bottles at various angles, using the built-in inspection program of the equipment to process, analyze, compare and identify the collected images, and determine whether the glass bottle has defects. The inspection method provided by this method is mainly aimed at bottle mouth inspection, bottle width inspection and black spot sample inspection, combined with other conventional inspection modules, to form an overall inspection process for glass bottles.
[0013] The bottle mouth inspection is divided into two steps. The first step is the protrusion inspection of the bottle mouth, which mainly detects whether there are tiny protrusions on the mouth ring of the bottle mouth. The shooting direction is to take a photo of the bottle mouth from the side horizontally to ensure that the upper end of the bottle mouth is a plane in the captured image. The mouth ring frame is set in the inspection program. The height of the mouth ring frame is consistent with the height of the mouth ring part, and the width is greater than the diameter of the mouth ring part, so that the upper and lower lines of the mouth ring frame coincide with the upper and lower ends of the mouth ring part respectively. Figure 1 As shown, the detection program compares the photographed mouth ring with the mouth ring outer frame. When the upper end of the mouth ring after comparison exceeds the upper line of the mouth ring frame, it can be determined that a protruding defect occurs in the mouth ring part and the detection result is fed back.
[0014] The second step is the scissors mark detection, which detects the scissors mark defects on the bottle mouth. The scissors mark defects are caused by uneven cooling due to the temperature difference caused by the cutting of the scissors blades when shearing the glass liquid. Due to the size of the bottle mouth, the scissors mark on the bottle mouth is extremely small and not obvious, and cannot be detected by general detection methods. This step takes a picture of the mouth of the glass bottle from a top-down angle so that the glass bottle mouth appears as a complete circle in the photo. In the detection program, a cross-bottle mouth line is set. The cross-bottle mouth line consists of two concentric circular detection frames. The outer detection frame is flush with the outer edge of the bottle mouth, and the inner detection frame is flush with the inner edge of the bottle mouth, so that the cross-bottle mouth line completely frames the upper end face of the bottle mouth. The detection program compares the image brightness of the part of the bottle mouth framed in the two detection frames to determine whether there is a scissors mark defect. Set the detection parameter sensitivity value to 5, the width value to 40mm, and the angle difference to 55. Figure 2 When an abnormal point is detected, as shown in the figure, Figure 3 As shown, it is determined that there is a scissors mark defect on the bottle mouth and the detection result is fed back.
[0015] Bottle width detection is to take a photo of the glass bottle from the side to detect the overall width of the glass bottle. Since glass bottles have different diameters, 3-5 detection frames are set according to the different diameters of the glass bottles in the detection program, such as Figure 4 As shown, three detection frames are set for the bottleneck, shoulder and body of the bottle respectively, and the widths of the three parts are detected. When the detection program finds that the bottle width data detected by the detection frames of the three parts reach the set error during the detection process, the glass bottle is completely qualified. If the width data of any part does not meet the predetermined value, it is judged that the glass bottle has a width defect and the detection result is fed back.
[0016] During daily production line inspections, in order to monitor the accuracy of the inspection equipment, some sample bottles are put on the inspection line. The sample bottles are glass bottles with minor defects. After being put on the production line, we check whether the inspection equipment can accurately detect the sample bottles, so as to monitor whether the inspection equipment is operating normally. Since the sample bottles are defective bottles, it is necessary to ensure that they can be quickly removed in the event of missed inspections, and at the same time, it is also necessary to avoid affecting the normal inspection work during the inspection process. After the black dot sample patch is affixed, the black dot sample inspection is carried out. 12 black dot sample patches are affixed to the sample bottle, and the affixing positions are respectively in the upper, middle and lower sections, and 4 are affixed along the circumference of each section, such as Figure 5 As shown, the main body of the black dot sample patch is a transparent film disc, and an opaque black dot is set in the center of the transparent film disc. Figure 5 As shown in the red box, the black spot can be observed while the surrounding area of the black spot will not affect the detection result.
[0017] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various changes and variations may occur. Any modifications, equivalent substitutions, improvements, etc. made using the present invention should be included in the scope of protection of the present invention.
Claims
1. A method for detecting defects on a glass bottle body using a camera, characterized in that: Including bottle mouth inspection, bottle width inspection and black spot sample inspection; The bottle mouth detection includes protrusion detection and scissor mark detection; the protrusion detection is to take a photo of the side of the bottle mouth for detection, and a mouth ring outer frame is set in the bottle mouth detection part, the mouth ring outer frame frames all the mouth rings of the bottle mouth part, and the upper side of the mouth ring outer frame is flush with the upper edge of the detected bottle mouth. When a part exceeding the upper edge of the mouth ring outer frame is detected, it is judged that a protruding abnormal point appears on the mouth ring of the bottle mouth, and the detection fails; The scissors mark detection is to take photos of the upper end of the bottle mouth from top to bottom. A cross-bottle mouth line consisting of two concentric detection frames is set. The outer detection frame is flush with the outer edge of the bottle mouth, and the inner detection frame is flush with the inner edge of the bottle mouth. The image abnormal points between the two detection frames are detected through the detection program. When it is detected that the brightness between the two detection frames is inconsistent with other areas, it is judged that a scissors mark defect appears on the upper end of the bottle mouth and the detection fails.
2. The method for detecting defects of glass bottles using a camera according to claim 1, characterized in that: The parameters of the cross-bottle mouth line are: sensitivity value is set to 5, width value is 40mm, and angle difference is 55.
3. The method for detecting defects of glass bottles using a camera according to claim 2, characterized in that: Bottle width detection is to take a photo of the glass bottle from the side, set up 3-5 detection frames, and detect the width of the bottleneck, shoulder and bottle body by setting detection frames of different heights. Through multi-segment detection, it is determined whether the glass bottle has width defects. When the width data detected by the detection frame does not meet the predetermined value, it is determined that the glass bottle has defects and the inspection fails.
4. The method for detecting defects of glass bottles using a camera according to claim 3, characterized in that: The black spot sample detection is to stick a black spot sample patch on the sample bottle and then put the sample bottle into the detection line for detection, so as to test the detection accuracy of the detection line.
5. The method for detecting defects of glass bottles using a camera according to claim 4, characterized in that: A total of 12 black dot sample patches are affixed to the sample bottle, 4 of which are affixed along the circumference of the upper section, the middle section and the lower section respectively. The black dot sample patches are transparent film discs with an opaque black dot in the center.