Glass bead bubble detecting and counting system and method based on solution scheme

Abstract

The invention relates to the field of automatic target detection, and discloses a glass bead bubble detection counting system and method based on a solution scheme, and the system comprises a conveying device, a container tray, a visual detection system, a glass bead feeding device, a filtering separation device and a solution recovery device; a container tray is arranged on an original node on the conveying device and conveyed to a feeding node, and a glass bead feeding device feeds materials into the container tray; the loaded container tray is conveyed to a visual detection node by a conveying device, and enters a filtering and separating device node after being subjected to visual detection by a visual detection system; after the filtering and separating device node separates the solution and the glass beads in the container tray, the solution flows back to the solution recycling device, the container tray is conveyed back to the original node by the conveying device, and counting of bubbles of the glass beads is completed; according to the invention, automatic detection and counting of glass bead bubbles can be realized, and the detection efficiency and precision are improved.

Description

Technical Field

[0001] This invention relates to the field of automated target detection, and more particularly to a glass bead bubble detection and counting system and method based on a solution scheme. Background Technology

[0002] Traditional methods for counting bubbles in glass beads mainly include the following: 1. Manual Inspection Method: This is the most traditional method for inspecting the appearance of glass beads. Workers need to use tools such as magnifying glasses to manually inspect the surface of each glass bead one by one. This method is cumbersome, inefficient, and unreliable, and prone to human error.

[0003] 2. Semi-automatic testing equipment: Some glass bead manufacturers have adopted semi-automatic testing equipment based on image processing. This equipment automatically acquires high-resolution images of glass bead samples, which are then visually analyzed and statistically processed by staff through a software interface. This method improves testing efficiency, but still requires manual intervention, and the test results are susceptible to human factors. Summary of the Invention

[0004] The purpose of this invention is to propose a glass bead bubble detection and counting system and method based on a solution scheme, thereby solving the technical problems of low efficiency and inaccurate detection results of existing traditional glass bead counting methods.

[0005] Specifically, the present invention provides a glass bead bubble detection and counting system based on a solution solution, comprising: a conveying device, a container tray, a vision inspection system, a glass bead feeding device, a filtration and separation device, and a solution recovery device; The container pallet is placed at the original node on the conveying device and conveyed to the feeding node, where the glass bead feeding device feeds the container pallet. After being loaded, the container pallet is conveyed by a conveyor to the vision inspection node, where it undergoes visual inspection by the vision inspection system before entering the filtration and separation device node. After the filtration and separation device separates the solution and glass beads in the container tray, the solution flows back to the solution recovery device, and the container tray is conveyed back to the original node by the conveyor device to complete the counting of glass bead bubbles.

[0006] A solution-based method for detecting and counting glass bead bubbles, applied to a solution-based glass bead bubble detection and counting system, includes the following steps: S1. Material images are acquired by a vision inspection system; S2. Construct a glass bead target detection model. Input the material image into the glass bead target detection model to obtain the glass bead target image. S3. Use the bubble target detection model to detect bubbles in the glass bead target image, restore the bubble position to the original material image coordinate system, eliminate the number of duplicate detections based on the bubble position, and obtain the final count of glass bead bubbles.

[0007] A storage device that stores instructions and data for implementing a solution-based glass bead bubble detection and counting method.

[0008] A glass bead bubble detection and counting device based on a solution scheme includes: a processor and a storage device; the processor loads and executes instructions and data in the storage device to implement a glass bead bubble detection and counting method based on a solution scheme.

[0009] The beneficial effects of this invention are as follows: Compared with traditional methods and systems, this system achieves full automation, enabling fully automatic, rapid, and accurate detection and analysis of large batches of samples. The system employs a fully automated loading and unloading structure, automatically dispensing glass beads and solutions. A high-definition camera then captures images of the glass beads, and a professional image processing algorithm automatically identifies and counts air bubbles within the glass beads. Finally, the glass beads and solution are separated, with the solution recycled via a pump. The entire system significantly improves detection efficiency and accuracy. This is of great significance for quality control in glass manufacturing enterprises. Attached Figure Description

[0010] Figure 1 This is a schematic diagram of the glass bead bubble detection and counting system based on a solution-based approach according to the present invention. Figure 2 This is a schematic diagram of the detection of glass beads in solution; Figure 3 These are the actual detection images captured by this invention; Figure 4 This is a schematic diagram of the hardware device used in this application. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0012] Before formally describing the present invention, a general description of the solution of the present invention will be given first to facilitate understanding.

[0013] Please refer to Figure 1 The present invention provides a glass bead bubble detection and counting system based on a solution scheme, comprising: Conveying devices, container trays, vision inspection systems, glass bead feeding devices, filtration and separation devices, and solution recovery devices; The container pallet is placed at the original node on the conveying device and conveyed to the feeding node, where the glass bead feeding device feeds the container pallet. After being loaded, the container pallet is conveyed by a conveyor to the vision inspection node, where it undergoes visual inspection by the vision inspection system before entering the filtration and separation device node. After the filtration and separation device separates the solution and glass beads in the container tray, the solution flows back to the solution recovery device, and the container tray is conveyed back to the original node by the conveyor device to complete the counting of glass bead bubbles.

[0014] It should be noted that the conveying device can be a conveyor belt.

[0015] It should be noted that the feeding node includes a leveling unit, which is activated after the glass bead feeding device has finished feeding. This unit is used to mix the glass beads evenly with the solution and level the glass beads to prevent stacking. Specifically, the leveling unit can employ a piezoelectric ceramic vibrator combined with an eccentric wheel mechanism, with the amplitude (0.1–0.5 mm) and frequency (20–50 Hz) adjusted by a PID controller. Additionally, the leveling unit can be covered with a silicone anti-slip layer to ensure that the glass beads are evenly spread and do not shift during vibration. The piezoelectric ceramic vibrator, eccentric wheel mechanism, PID controller, etc., can be arbitrarily selected and combined according to actual conditions. This invention is only illustrative and not intended to limit the scope.

[0016] It should be noted that the visual inspection node includes a backlight unit. After the container tray is transferred to the visual inspection node, the backlight unit is activated to emit a light beam, facilitating visual inspection by the visual inspection system. Specifically, the backlight unit can employ a high-uniformity LED surface light source paired with a collimating lens array, with a wavelength of 630nm (red light has strong penetrability) and adjustable brightness (5000–10000 lux). The light source forms a 15° angle with the tray to eliminate specular reflection interference. A polarizing filter can also be added in this invention to eliminate surface reflection from the solution. The above-mentioned light source can be selected according to actual conditions; this invention is for illustrative purposes only and is not intended to limit the scope.

[0017] It should be noted that the filtration and separation device uses a mesh container. After the glass bead solution passes through the mesh container, part of the solution is recovered, and the glass beads are filtered onto the mesh.

[0018] It should be noted that the glass bead feeding device includes a plug. During feeding, a certain amount of solution is first injected into the container tray, and then the glass beads are injected. The plug is used to control the weight of the glass beads in the injected solution.

[0019] It should be noted that the visual inspection system includes a computer and an industrial camera. The industrial camera acquires images and transmits them to the computer, where they are processed by image analysis algorithms.

[0020] refer to Figure 2 , Figure 2 This is a schematic diagram of the detection of glass beads in a solution.

[0021] It should be noted that the solution used is benzyl benzoate solution, which has a similar refractive index to the glass bead sample. Placing the glass beads in the solution for visual inspection can avoid overexposed and shadowed areas caused by light refraction in the air.

[0022] It should be noted that after visual inspection, the container tray is rotated and unloaded before entering the filtration and separation device.

[0023] It should be noted that a large amount of glass bead sample is first manually poured into the hopper, and then a certain amount of solution is injected into the solution circulation system. The system is then started to operate automatically.

[0024] The system operates automatically according to the set procedures: Pour a certain amount of solution into an empty container tray.

[0025] Control the injection of 10g of glass bead sample into the container tray containing the solution.

[0026] Move the container pallet to the vision inspection station for vision inspection.

[0027] Tilt the container tray to pour the solution and glass beads onto the mesh-covered recycling station. The solution is recycled by a pump, and the glass beads on the mesh are poured into the waste bin.

[0028] A solution-based method for detecting and counting glass bead bubbles, applied to a solution-based glass bead bubble detection and counting system, includes the following steps: S1. Material images are acquired by a vision inspection system; S2. Construct a glass bead target detection model. Input the material image into the glass bead target detection model to obtain the glass bead target image. S3. Use the bubble target detection model to detect bubbles in the glass bead target image, restore the bubble position to the original material image coordinate system, eliminate the number of duplicate detections based on the bubble position, and obtain the final count of glass bead bubbles.

[0029] It should be noted that the glass bead target detection model in step S2 adopts the YOLO model; the bubble target detection model in step S3 adopts the rt-detr model.

[0030] Specifically, the glass bead target detection model can adopt the YOLOv8 model. Preferably, a "dual-path attention mechanism" can be introduced on the basis of the YOLOv8 model, which includes two branches: Branch 1: Extract global features using conventional convolution; Branch 2: Dilated convolution (dilation=3), enhances the capture of local features of small bubbles; Feature fusion: Dynamically weighted dual-path output through CBAM (Convolutional Block Attention Module) prioritizes focusing on areas with high bubble incidence (such as the edges of glass beads).

[0031] Specifically, rt-detr (Real-Time DEtection TRansformer) is a real-time optimized version of DETR, achieving high-speed inference (≥30fps) by reducing the number of encoder layers and using a dynamic matching strategy. Its advantage lies in end-to-end detection, eliminating the need for NMS post-processing.

[0032] Preferably, the present invention can also incorporate "scale-adaptive token selection" into the RT-DETR decoder: Step 1: Pre-generate bubble candidate regions (ROIs) using a lightweight CNN; Step 2: Only perform Transformer decoding on feature tokens within the ROI, reducing computation by 70%; It should be noted that in the Transformer architecture, feature tokens are the basic units that encode the spatial location of image features into serialized vectors.

[0033] Step 3: Design the bubble shape prior loss function: constrain the aspect ratio of the predicted box (the bubble is usually round, and the aspect ratio is ∈ [0.8, 1.2]).

[0034] The prior loss function for bubble shape is as follows: ,in: Indicates aspect ratio loss. , w i , h i They represent the first i The width and height of each prediction box, τ This represents the tolerance threshold (experimental value = 0.2, allowing aspect ratios ∈ [0.8, 1.2]), when... A penalty is imposed when the circle deviates from its designated shape.

[0035] Indicates a loss of density. ,in, A i Indicates the area of ​​the prediction box. Indicates the perimeter of the prediction box. This is used to measure similarity to a circle (circle = 1, rectangle ≈ 0.78). λ 1. λ 2 represents the weighting coefficient in this invention. λ 1 = 0.7 λ 2 = 0.3 (prioritize aspect ratio constraint).

[0036] Please refer to Figure 3 , Figure 3 These are the detection images actually captured by this invention.

[0037] The actual image of the glass bead has a resolution of 3480*2160, while the bubbles occupy about 25 to 35 pixels. The proportion of bubbles in the whole image is very small. No model can directly detect bubbles, and it is impossible to distinguish bubbles in the background (bubbles that already exist in the solution, rather than bubbles in the glass bead).

[0038] Therefore, this invention designs a glass bead bubble detection and counting method based on a solution-based approach. First, a YOLO object detection model is used to detect the glass beads, and the image is cropped to extract the glass bead image from the original image. Then, an RT-DETR object detection model is used to detect bubbles in the cropped image. Finally, the bubble positions are restored to the original image coordinate system, and duplicate bubble detections are eliminated based on their positions, thereby improving statistical accuracy.

[0039] Please see Figure 4 , Figure 4 This is a schematic diagram of the hardware device in operation according to an embodiment of the present invention. The hardware device specifically includes: a glass bead bubble detection and counting device 401 based on a solution scheme, a processor 402, and a storage device 403.

[0040] A solution-based glass bead bubble detection and counting device 401: The solution-based glass bead bubble detection and counting device 401 implements the solution-based glass bead bubble detection and counting method.

[0041] Processor 402: The processor 402 loads and executes the instructions and data in the storage device 403 to implement the glass bead bubble detection and counting system and method based on a solution scheme.

[0042] Storage device 403: The storage device 403 stores instructions and data; the storage device 403 is used to implement the glass bead bubble detection and counting method based on a solution scheme.

[0043] In summary, the beneficial effects of this invention are as follows: Compared with traditional methods and systems, this system achieves full automation, enabling fully automatic, rapid, and accurate detection and analysis of large batches of samples. The system employs a fully automated loading and unloading structure, automatically dispensing glass beads and solutions. A high-definition camera then captures images of the glass beads, and a professional image processing algorithm automatically identifies and counts air bubbles within the glass beads. Finally, the glass beads and solution are separated, with the solution recycled via a pump. The entire system significantly improves detection efficiency and accuracy. This is of great significance for quality control in glass manufacturing enterprises.

[0044] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A glass bead bubble detection and counting system based on a solution-based approach, characterized in that: include: Conveying devices, container trays, vision inspection systems, glass bead feeding devices, filtration and separation devices, and solution recovery devices; The container pallet is placed at the original node on the conveying device and conveyed to the feeding node, where the glass bead feeding device feeds the container pallet. After being loaded, the container pallet is conveyed by a conveyor to the vision inspection node, where it undergoes visual inspection by the vision inspection system before entering the filtration and separation device node. After the filtration and separation device separates the solution and glass beads in the container tray, the solution flows back to the solution recovery device, and the container tray is conveyed back to the original node by the conveyor device to complete the counting of glass bead bubbles.

2. The glass bead bubble detection and counting system based on a solution scheme as described in claim 1, characterized in that: The feeding node includes a leveling unit, which is activated after the glass bead feeding device has finished feeding. The leveling unit is used to mix the glass beads with the solution evenly and to level the glass beads to prevent them from stacking.

3. The glass bead bubble detection and counting system based on a solution scheme as described in claim 1, characterized in that: The visual inspection node includes a backlight unit. After the container tray is transferred to the visual inspection node, the backlight unit is activated to emit a light beam, which facilitates visual inspection by the visual inspection system.

4. The glass bead bubble detection and counting system based on a solution scheme as described in claim 1, characterized in that: The filtration and separation device uses a mesh container. After the glass bead solution passes through the mesh container, part of the solution is recovered, and the glass beads are filtered onto the mesh.

5. A solution-based method for detecting and counting glass beads bubbles, applied to the solution-based glass bead bubble detection and counting system as described in any one of claims 1 to 4, characterized in that: Includes the following steps: S1. Material images are acquired by a vision inspection system; S2. Construct a glass bead target detection model. Input the material image into the glass bead target detection model to obtain the glass bead target image. S3. Use the bubble target detection model to detect bubbles in the glass bead target image, restore the bubble position to the original material image coordinate system, eliminate the number of duplicate detections based on the bubble position, and obtain the final count of glass bead bubbles.

6. The glass bead bubble detection and counting method based on a solution scheme as described in claim 5, characterized in that: In step S2, the glass bead target detection model uses the YOLO model; in step S3, the bubble target detection model uses the rt-detr model, where the prior loss function for bubble shape is designed as follows: in: Indicates aspect ratio loss. , w i , h i They represent the first i The width and height of each prediction box, τ Indicates the tolerance threshold, when Punishment should be imposed at the appropriate time; Indicates a loss of density. ,in, A i Indicates the area of ​​the prediction box. Indicates the perimeter of the prediction box. , used to measure similarity to a circle; λ 1. λ 2 represents the weighting coefficient.

7. A storage device, characterized in that: The storage device stores instructions and data for implementing the glass bead bubble detection and counting method based on a solution scheme as described in any one of claims 5 to 6.

8. A glass bead bubble detection and counting device based on a solution-based method, characterized in that: include: A processor and a storage device; the processor loads and executes instructions and data in the storage device to implement the glass bead bubble detection and counting method based on a solution scheme as described in any one of claims 5 to 6.

Images