CN201935873U - Online image detection system for bottle cap - Google Patents

Abstract

The utility model belongs to the technical field of machine vision and image processing, and relates to an online image detection system for bottle caps. Above, the light source is placed obliquely above the conveyor belt, and the optical sensor is placed next to the conveyor belt directly below the camera to detect whether there is a bottle cap directly below the camera. The detection signal of the optical sensor triggers the camera to capture the bottle on the conveyor belt directly below it. The image of the bottle cap collected is sent to the computer through the image acquisition card. The light source of the utility model has a reasonable layout, the optical sensor triggers the camera to collect images online, and the computer realizes image feature extraction and high-speed automatic detection, which can improve the quality of the product packaging process.

Description

Bottle cap online image detection system

technical field

The utility model belongs to the technical field of machine vision and image processing, and relates to an online image detection system for bottle caps.

Background technique

Now many domestic manufacturers are more and more aware that quality and productivity are always the core factors. In the industrial field, machine vision technology is more and more used to monitor the production process and test product quality, so as to ensure the performance and appearance of the product. quality. At present, the product inspection of most domestic production lines is completed by manpower, which will not only increase the company's manpower investment, but also cause low efficiency. Therefore, there is a huge market demand for using machine vision technology to complete product defect detection, classification and grading. At the same time, it is also an area that the entire visual industry is currently keen on. At present, bottle caps such as plastic caps and aluminum caps can be produced quickly on industrial production lines. It is necessary to efficiently detect and eliminate possible defects in bottle caps to adapt to the production efficiency of fast production lines.

Utility model content

The purpose of this utility model is to provide an on-line inspection device for bottle caps that can realize on-line detection of bottle caps and avoid quality problems caused by product packaging processes. The scheme that the utility model solves its technical problem adopts is as follows:

An online image detection system for bottle caps, including a conveyor belt for conveying bottle caps, an optical sensor, a camera, a light source, an image acquisition card, and a computer, the camera is fixed directly above the conveyor belt, the light source is placed obliquely above the conveyor belt, and the optical sensor is placed Next to the conveyor belt directly below the camera, it is used to detect whether there is a bottle cap directly below the camera. The detection signal of the optical sensor triggers the camera to collect the image of the bottle cap on the conveyor belt directly below it, and the collected bottle cap image is passed through image acquisition. The card is fed into the computer.

Preferably, the optical sensor is an infrared sensing photoelectric sensor.

The bottle cap on-line detection system provided by the utility model has a reasonable light source layout, an optical sensor triggers a camera to collect images online, and a computer realizes image feature extraction and high-speed automatic detection, which can improve the quality of product packaging process.

Description of drawings

Figure 1: A schematic diagram of the appearance and composition of the hardware system of the utility model;

Fig. 2: the utility model software flowchart;

Fig. 3: The defect detection subroutine flowchart of the utility model;

Fig. 4: Schematic diagram of the circular ring straightening process of the present invention.

In Figure 1: 1 conveyor belt, 2 bottle caps, 3 optical sensors, 4 sensor signal lines, 5 cameras, 6 light sources, 7 transmission lines, 8 image acquisition cards, and 9 computers.

Detailed ways

Referring to Fig. 1, the bottle cap high-speed detection system of the present utility model is made up of transmission equipment, camera, special light source, transmission line, image acquisition card, computer. Specialized conveyor equipment includes conveyor belts and optical sensors. The camera of the camera can be a linear array CCD camera, which is installed directly above the conveyor belt, and is triggered by an optical sensor signal to collect the image of the bottle cap facing upwards in real time. The special light source obliquely illuminates the bottle cap at a certain angle, so that the edge of the bottle cap appears as a ring with a certain width in the collected image, so as to detect the gap of the bottle cap. The optical sensor is an infrared sensing photoelectric sensor, which is fixed next to the transmission line under the camera. The image acquisition card can be an external card with image front-end processing based on the PCI bus structure of the microcomputer, and forms a relatively complete real-time image acquisition and processing system with the existing resources of the microcomputer. Due to the large amount of image information and the long processing time, the computer needs to have a higher main frequency and stronger computing power. The camera collects real-time images, the size of which is 800×600, and sends them to the computer through the transmission line and image acquisition card for storage in BMP format and real-time detection.

The utility model is composed of a hardware part and a software part; the hardware part adopts a computer as a processing and control center; the software part consists of multiple subroutines such as image preprocessing, edge extraction, deformation detection, and defect detection.

Referring to Fig. 2, the software part includes: 1) image preprocessing subroutine, 2) edge extraction subroutine, 3) deformation detection subroutine, 4) defect detection subroutine.

1) The image preprocessing subroutine includes: image grayscale, noise removal, grayscale stretching, binarization and morphological filtering.

The image is grayscaled. The image captured by the camera in real time is 16-bit bitmap (RGB565 format) data. In order to facilitate the subsequent rapid image processing, the image data needs to be converted to make the color image into a 256-level grayscale image.

To remove noise, the image inevitably contains noise, and then the image is preprocessed by median filtering.

Gray-scale stretching, in order to enhance the contrast between the background area and the character area, the image is gray-scale stretched.

Binarization, binarize the grayscale image, adopt the method of the maximum inter-class variance and the minimum intra-class variance ratio, and adaptively calculate the gray threshold, the area smaller than this threshold is considered the target area, and the area greater than this threshold considered to be the background region.

Morphological filtering, which performs morphological filtering processing on binary images, and adopts a comprehensive operation combining dilation, erosion, opening operation and closing operation.

2) The edge extraction subroutine, the edge of the bottle cap in the binarized image appears as a ring with a certain width, and the edge ring image is extracted by dragging and dropping the area selection template composed of three concentric circles. The radius of the three circles can be Set independently, and name the three circles as outer circle, middle circle and inner circle respectively in order from outside to inside. The middle circle is used to estimate the cap radius, and the inner and outer circles are used to define the edge ring of the cap, where the inner circle matches the inner edge of the cap and the outer circle matches the outer edge of the cap.

3) The deformation detection subroutine calculates the circularity of the outer edge of the ring as shown in formula 1, and compares it with the preset circularity threshold. If the calculation result is lower than the circularity threshold, the bottle cap is judged as deformation, while generating a culling signal.

$\mathrm{<span\; class="notranslate">\; C</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; P</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; A</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Among them, P is the perimeter and A is the area.

4) Defect detection subroutine, detects whether there are defects such as burrs, burrs and gaps in the bottle cap, including: parameter calculation, ring straightening, preprocessing and width measurement, see Figure 3.

Parameter calculation: Calculate the position parameters required for the edge ring straightening projection transformation, that is, the sine and cosine values, and store them in the corresponding array. As shown in Figure 4, let the distance between a certain pixel point (xx, yy) to be stretched in the edge circle and the positive axis of the x-axis relative to the center of the circle (x0, y0) be a, where the sine sum The cosine value refers to the sine and cosine values of the included angle a, which is used to calculate the pixel value corresponding to the coordinates (x, y) of the expanded image in the original image, that is, the original edge circle in the projection transformation. If it is a hole point, further use Interpolate its pixel values.

The ring is straightened, and the outer circle in the area selection template is used as the reference, that is, the outer circle is used as the top edge of the rectangular area after projection transformation, and the difference between the radius of the inner and outer circles is used as the width of the rectangular area after projection transformation, and the edge ring is aligned along the right horizontal line Cut and project to the aforementioned rectangular area, where the pixel value of the hole point is calculated by bilinear interpolation algorithm.

Preprocessing, binarize the straightened ring image according to the threshold value of binarization in image preprocessing, and perform morphological filtering to eliminate isolated points.

Width measurement, the maximum and minimum width values of the ring are obtained through statistics and the average width of the ring is calculated. If the maximum and minimum width values exceed a certain percentage of the average width, it is considered that the bottle cap has defects such as burrs, burrs or gaps, and a rejection signal is generated at the same time .

The preferred embodiments of the utility model are introduced below, and this part is only an illustration of the utility model, rather than a limitation of the utility model and its application or use. Other implementation modes obtained according to the utility model also belong to the technical innovation scope of the utility model. The setting of relevant parameters in the scheme does not mean that only example values can be used.

The working process of the utility model: the system is powered on, the camera continuously collects the bottle cap image in real time, and transmits it to the microcomputer through the transmission line; the image preprocessing module performs enhancement, filtering and binary processing on the input image; according to the area formed by the concentric circle group The selection module extracts the ring image of the edge of the bottle cap; calculates the circularity of the outer edge of the ring, and compares it with the preset circularity threshold. If the circularity threshold is set to 0.8, if the calculation result is less than the preset threshold, then judge The bottle cap is deformed, and a rejection signal is generated, otherwise it enters the defect detection subroutine; the defect detection process first calculates the sine and cosine values required for the ring straightening projection transformation, and then straightens the gray ring image through the projection transformation, and then passes Binarization and filtering are used to process the straightened ring image, and finally the maximum and minimum width of the ring are counted and compared with the average width. If the maximum and minimum width exceed 20% of the average width, it is judged that the bottle cap is defective and will be rejected. Signal.

Claims (2)

1. An online image detection system for bottle caps, comprising a conveyor belt for transporting bottle caps, an optical sensor, a camera, a light source, an image acquisition card, and a computer, is characterized in that the camera is fixed directly above the conveyor belt, and the light source is placed on the conveyor belt Obliquely above, the optical sensor is placed next to the conveyor belt directly below the camera to detect whether there is a bottle cap directly below the camera. The detection signal of the optical sensor triggers the camera to capture the image of the bottle cap on the conveyor belt directly below it. The bottle cap image is sent to the computer through the image acquisition card. 2. The bottle cap online image detection system according to claim 1, wherein said optical sensor is an infrared sensing photoelectric sensor.

Images