CN101007308A - Pearl real time detection and classifying system based on mechanical vision - Google Patents

Abstract

The invention discloses a real-time pearl detection and grading system based on machine vision, which comprises a computer vision identification component arranged under a feeding conduit of a pearl feeding device, and a pearl sorting device arranged under the computer vision identification component. Use machine vision technology to collect images of the pearl surfaces that are falling freely in sequence, and transmit the collected images to hardware devices such as DSP and computer for processing, and make instant judgments on the processing results according to the pearl classification standards, and finally pass the judgment The information is sent to the sorting device, so that the pearls with different characteristics can be automatically classified by the equipment. The image of the entire surface of the pearl can be obtained, and the detection of all appearance quality indicators such as the size, shape, smoothness, blemish and color of the pearl can be completed at the same time, and the rapid real-time detection and grading of the pearl can be realized. The system can adapt to different grading speeds, and has a simple structure and strong adaptability.

Description

Pearl real-time detection and grading system based on machine vision

technical field

The invention relates to a machine vision-based detection and grading system, in particular to a machine vision-based pearl real-time detection and grading system.

Background technique

my country is one of the first countries in the world to discover and use pearls. In 2005, the annual output of pearls in my country reached more than 1,500 tons, accounting for more than 95% of the world's total output, and the output of freshwater pearls accounted for 99% of the world's output. There are more than 400,000 people engaged in the pearl industry and related industries, and there are nearly a thousand enterprises engaged in the processing and sales of freshwater pearls, with an output value of more than 6 billion yuan. At present, freshwater pearl industry clusters represented by Zhuji, Zhejiang, and seawater pearl breeding trade bases in Beihai and Zhanjiang have been formed nationwide. Different pearls have different values according to their characteristics such as size, color, shape, luster and blemishes, and in various practical applications, pearls of different grades are also required, which requires the detection and grading of pearls . At present, the sorting method used by enterprises in our country is basically the traditional manual method. This method requires a lot of labor, is slow, and consumes a lot of manpower. At the same time, this subjective evaluation is subject to personal vision, color discrimination, emotion, light And other aspects of the impact, and only a single index can be screened when grading, not only low efficiency but also poor accuracy.

In order to unify the standards of pearl sorting and processing and improve the quality and production capacity of pearl sorting and processing, it is self-evident that it is superior to use computer vision technology to replace manual pearl quality inspection and grading. First of all, visual measurement can eliminate the influence of human subjective factors, avoiding the detection results that vary from person to person. In addition, it can perform unified sorting of multiple indicators, and can realize one-time grading of each indicator of pearls. At present, this technology has been applied in the field of pearl grading, but it can only be classified according to a single requirement such as shape classification, and it cannot be uniformly graded for smoothness and blemishes; at the same time, it cannot achieve smoothness and blemishes. According to the requirements of analyzing each surface of the pearl, it can only obtain the image information of part of the surface of the pearl to be tested, and the missed detection rate of pearl surface defects is relatively high. In addition, due to its slow speed, the fast online detection of pearls cannot be realized.

Contents of the invention

The object of the present invention is to provide an on-line detection and grading system for real-time detection and grading of pearl quality. Use machine vision technology to collect images of the pearl surfaces that are falling freely in sequence, and transmit the collected images to hardware devices such as DSP and computer for processing, and make instant judgments on the processing results according to the pearl classification standards, and finally pass the judgment The information is transmitted to the sorting device, so that the pearls with different characteristics can be automatically classified by the equipment.

The technical scheme that the present invention adopts is as follows:

A computer vision recognition component is set under the feeding conduit of the pearl feeding device, and a pearl sorting device is set under the computer vision recognition component; the computer vision recognition component includes a light source and a camera, a protective cover, a sensor and a computer; The upper and lower ends of the cover are respectively connected with the image acquisition card and the control module in the computer through lines, and provide the position information of each pearl; the light source and the camera are evenly spaced in the three-dimensional space around the protective cover, consisting of four cameras are connected to the image acquisition card and control module in the computer via lines; the pearl position information provided by the sensor triggers the dynamic collection of pearl images, and processes them through the program in the computer, and the processing results are judged in real time according to the pearl classification standard .

The pearl feeding device: includes a feeding port, a feeding barrel, a claw, a dial, a DC motor and a feeding conduit; the upper end of the feeding barrel is provided with a feeding port, and the inner bottom of the feeding barrel is provided with a dial and The claw is coaxially installed on the dial, the lower end of the feeding tube is connected to the feeding conduit, and the DC motor installed outside the lower end of the feeding drum drives the dial and the claw to rotate. A group of small through-holes are evenly opened on the same circumference, and the small through-holes are aligned with the orifice of the feeding conduit.

The pearl sorting device: the sorting device is an intermittently operating turntable system, including a feeding tray, a feeding tray switch, an electromagnetic action switch, an electromagnetic switch control part, a feeding conduit and a DC motor; A number of feeding holes are distributed. After the pearl passes the shooting area, it automatically enters a certain feeding hole of the feeding tray. At this time, the computer system records the number of the feeding hole where the pearl is located at the same time. Intermittent operation, and each movement turns a partition box position. When the partition box where the pearl is located is transferred to the grading receiver corresponding to the pearl, the computer immediately triggers the corresponding electromagnetic control switch. switch, so that the pearl enters the corresponding receiving container through the feeding conduit.

The pearl feeding device can make the pearls be conveyed in the way of free fall at a certain frequency, and the feeding speed can be adjusted very simply by controlling the speed of the turntable; when the pearls enter the position of the computer vision recognition module, the pearls can be guaranteed All surfaces can be detected at the same time, so that the image information of the inspected pearl can be obtained accurately, effectively and comprehensively, and the quality characteristic information can be extracted from it. Through the analysis of computer software, the detection of all appearance quality indicators such as pearl size, shape, color, smoothness and blemishes can be completed at the same time, and the grade of each pearl can be comprehensively judged according to the judgment standards drawn up in advance. And the result can be sent to the sorting device for grading and screening of pearls.

The beneficial effect of the present invention is that all surface images of pearls at 360 degrees can be obtained at the same time, corrected and analyzed, and the quality indicators of pearls such as size, shape, color, blemish, and cleanliness can be detected quickly and effectively. And the grading is realized through the automatic grading part, which improves the automation level and accuracy of pearl quality detection and grading. After being applied to actual production, it can improve the production efficiency of pearls, improve the quality level and competitiveness of products.

Description of drawings

Fig. 1 is an overall frame diagram of the present invention;

Fig. 2 is a structural representation of the present invention;

Fig. 3 is a schematic diagram of a feeding device of the present invention;

Fig. 4 is a schematic diagram of a sorting device of the present invention;

Fig. 5 is the computer vision recognition device of the present invention;

Fig. 6 is a spatial arrangement diagram of the camera in the computer vision recognition device of the present invention;

Fig. 7 is a simplified network diagram of the pearl sorting 1394 control bus;

Fig. 8 is a schematic diagram of pearl sorting control and signal transmission.

In the figure: 1 is the pearl feeding device, 2 is the computer vision recognition part, 3 is the pearl mechanical grading device, 4 is the feeding port, 5 is the barrel, 6 is the claw, 7 is the dial, 8 is the DC motor , 9 is a feeding conduit, 10 is a blanking tray, 11 is a blanking tray switch, 12 is an electromagnetic action switch, 13 is an electromagnetic switch control part, 14 is a blanking conduit, 15 is a DC motor, 16 is a camera and a light source, 17 is a protective cover, 18 is a pearl, 19 is a sensor, and 20 is a computer.

Detailed ways

The present invention will be further described below in conjunction with accompanying drawing and embodiment

As shown in FIGS. 1 and 2 , the present invention consists of a pearl feeding device 1 , a computer vision recognition component 2 , and a pearl mechanical grading device 3 . A computer vision recognition unit 2 is arranged below the feeding conduit 9 of the pearl feeding device 1 , and a pearl sorting device 3 is arranged below the computer vision recognition unit 2 . The designed feeding device 1 allows the pearls to enter the imaging area in a free-fall manner, which can ensure that the subsequent computer vision recognition component 2 obtains an image of the entire surface of the pearls. Its working process is: when the pearl falls from the feeding conduit of the feeding device 1, the position sensor 19 is triggered, and the trigger signal is transmitted to the image acquisition unit, and the grading processing software in the computer processes the acquired image to obtain the pearl The grading results of pearls are obtained, and the results are used to control the action of the feeding device to achieve the grading of pearls.

As shown in FIG. 3( a ), the pearl feeding device includes: a feeding port 4 , a feeding cylinder 5 , a finger 6 , a dial 7 , a DC motor 8 and a feeding conduit 9 . Figure 3 (b) shows the front view of the claw 6 and the dial 7. There are 16 holes evenly on the dial 7 (the number of holes can be adjusted according to the output requirements), and an aperture is arranged at the lower end of the feeding cylinder 5 . When working, the motor 8 drives the puller 7 to rotate in the barrel 5, and the pearls will automatically enter the small hole of the puller 7, and the pearls in the hole rotate with the turner 7. When above the hole, the pearl automatically falls into the conduit 9, thereby completing the feeding work of a pearl. Pearls of different sizes can be processed by adjusting the diameters of the dial 6 and the small holes of the feeding disc 7, and the feeding speed can be adjusted by two methods: (1) adjusting the number of small holes on the dial 7; (2) Adjust the rotating speed of DC motor 8.

Fig. 4(a) is a schematic diagram of the sorting device of the present invention. The mechanical sorting unit is an intermittently operating turntable system. It consists of a blanking tray 10, a blanking tray switch 11, an electromagnetic action switch 12, an electromagnetic switch control part 13, a blanking conduit 14, and a DC motor 15. As shown in Figure 4 (b), a plurality of blanking holes (also known as partition boxes, whose number is adjusted according to the number of pearl classifications) are distributed above the blanking tray 10 . After the pearl passes through the shooting area, it automatically enters a certain blanking hole of the blanking tray 10, and the computer system records the number of the blanking hole where the pearl is located at the same time. The entire feeding tray 10 runs intermittently according to the feeding frequency of the pearls, and each time it moves through a partition box position, when the partition box where the pearls are located is transferred to the grading receiver corresponding to the pearls, the computer immediately triggers the bottom of the corresponding feeding hole. The electromagnetic control switch 12, and by pulling the switch 11 below it, make the pearl enter the corresponding receiving container through the feeding conduit 14. Here, the number of switches 11 and electromagnetic control switches 12 is the same as the number of blanking holes on the blanking tray 10 .

As shown in Figure 5, it is a computer vision recognition component. Computer vision recognition includes light source and camera 16, protective cover 17, sensor 19, computer 20, etc. The sensors 19 are respectively located at the upper and lower ends of the protective cover 17 , are respectively connected with the image acquisition card and the control module in the computer 20 through lines, and provide the position information of each pearl 18 . The light source and the camera 16 are installed at intervals around the protective cover 17 according to the arrangement shown in FIG. The position information of the pearl 18 provided by the sensor 19 triggers the dynamic collection of the pearl image, and processes it through the program in the computer 20, and makes an instant judgment on the processing result according to the pearl classification standard.

Fig. 6 is a spatial arrangement diagram of the cameras in the computer vision recognition device of the present invention. Since the images captured by each camera cover nearly half of the surface of the pearl, only a single or two commonly used correspondingly arranged camera devices may have blind spots on some joint points, so in order to obtain information on the entire surface of the pearl, the present invention Four cameras evenly spaced in the three-dimensional space are designed, so that the detection of the entire pearl surface can be completed.

Fig. 7 is a simplified network diagram of the pearl sorting 1394 control bus. The control scheme adopts a high-speed 1394b control bus. In addition to the main control computer, the entire control system includes 6 peripheral devices: feeding controller, camera, camera exposure time controller, DSP image processor, sorting controller, lighting controller .

Figure 8 Schematic diagram of pearl sorting control and signal transmission. In the present invention, all the peripheral devices are connected to the 1394b bus by using the 1394b bus interface, so as to realize the data communication and the transmission of control signals among the devices. The specific control process is as follows:

(1) Before the sorting starts, the operator selects the sorting task, and according to the different sorting tasks, the computer sends out initialization instructions for each peripheral device through the 1394 bus to initialize each peripheral device. Such as feeding interval setting of the feeder, camera exposure time setting (the camera adopts mode 1 exposure mode), sorting channel encoding setting in the DSP image processor, decoding setting in the sorting controller, lighting brightness setting, etc.

(2) During the pearl sorting process, when pearls fall from the feeder, the photoelectric switch senses and sends out an initial pulse signal. The pulse signal passes through the pulse width controller, and sends out a certain length of exposure pulse and synchronization pulse according to the initialization requirements. The sync pulse makes the camera and DSP image processor work synchronously.

(3) After all the cameras receive the exposure pulse, after a certain time delay (delay generated in the distance between the pearl passing through the photoelectric sensor and the camera), the pearls are simultaneously photographed in parallel, and the graphics are transmitted through the 1394 bus to the respective DSP image processors.

(4) After the DSP image processor receives the image signal, it processes the image in parallel respectively, generates the sorting channel coding signal according to the sorting requirements set during the sorting initialization, and then immediately sends the sorting channel code number to 1394 on the bus, and notify the sorting controller to receive channel coded signals from the 1394 bus.

(5) After the sorting controller receives the channel coded signal from the 1394 bus, it identifies whether it is the coded signal in the channel according to the initialization setting result in its own decoder. When the set separation box of the mechanical sorting device is transferred to the pearl grading receiver corresponding to the coded signal, the sorting control switch is turned on, and the aerodynamic separator is started, so that the pearl enters the corresponding receiving container.

In the above control scheme, in addition to initial setting of the entire sorting device, the main control computer also undertakes the following tasks:

(1) Monitor the status of each peripheral device. During the sorting process, all peripheral devices need to report their abnormal status to the host through the 1394 bus in time. Once the host receives the abnormal status report, it must deal with it in time.

(2) Count the sorting results of each sorting channel. When the DSP image processor sends the sorting channel code number to the 1394 bus, the main control computer also receives the coded signal, and then decodes it to obtain the channel number, and counts the sorting results.

(3) Arbitrate the sorting results. After the DSP image processor processes the image, when it is difficult to judge the image processing result, it can send an arbitration request to the main control computer through the 1394 bus, and the host computer makes the final decision. After the host makes a decision, it sends a channel coding signal to the 1394 bus to control the sorting of channels. For example, in the sorting of pearl smoothness, the images taken by each camera may contain different surface defects. After processing the respective image signals, each DSP image processor may determine which category the pearl should be classified into. If it is difficult to make a judgment, the host can be requested to arbitrate.

All the components used above can be purchased from the market.

Claims (3)

1. The pearl real-time detection and grading system based on machine vision is characterized in that: a computer vision identification component (2) is set below the feeding conduit (9) of the pearl feeding device (1), and a computer vision identification component (2) is placed below the computer vision identification component (2). Pearl sorting device (3) is set; Described computer vision identification part (2) comprises light source and video camera (16), protective cover (17), sensor (19) and computer (20); Sensor (19) is respectively positioned at protective The upper and lower ends of the cover (17) are respectively connected with the image acquisition card and the control module in the computer (20) by lines, and provide the positional information of each pearl (18); ) is evenly distributed in the three-dimensional space around, and is composed of four cameras, which are respectively connected to the image acquisition card and the control module in the computer (20) via lines; the pearl (18) position information provided by the sensor (19) triggers the pearl The dynamic collection of images is processed by the program in the computer (20), and the processing results are judged in real time according to the pearl classification standard. 2. the pearl real-time detection and grading system based on machine vision according to claim 1, is characterized in that: described pearl feeding device (1): comprises feeding port (4), feeding cylinder (5), The claw (6), the dial (7), the DC motor (8) and the feeding conduit (9); The dial (7) and the claw (6) coaxially installed on the dial (7), the lower end of the feeding tube (5) is connected to the feeding conduit, and the lower end of the feeding tube (5) is installed outside the lower end of the feeding tube (5). The DC motor (8) drives the dial (7) and the claw (6) to rotate, and a group of small through holes are evenly opened on the same circumference of the dial 7, and the small through holes are aligned with the feeding conduit (9) orifice. 3. The pearl real-time detection and grading system based on machine vision according to claim 1, characterized in that: the pearl sorting device (3): the sorting device is an intermittently operated turntable system, comprising a feeding tray (10), blanking tray switch (11), electromagnetic action switch (12), electromagnetic switch control part (13), blanking conduit (14) and DC motor (15); above the blanking tray (10) are distributed A plurality of feeding holes, after the pearl passes the shooting area, it automatically enters a certain feeding hole of the feeding tray (10), at this time, the computer system simultaneously records the number of the feeding hole where the pearl is located, and the entire feeding tray (10) presses the pearl The feeding frequency runs intermittently, and every time it moves through a partition box position, when the partition box where the pearl is located is transferred to the corresponding grading receiver of the pearl, the computer immediately triggers the corresponding electromagnetic control switch (12), and by pulling the bottom The feed tray switch (11) under the feed tray (10) makes the pearls enter the corresponding receiving container by the feed conduit (14).

Images