Machine vision-based glucose solution quality inspection system design

1 Introduction

As industrial technology development and improvement of the quality inspection standard, high-speed production line of traditional medicine and water quality testing methods are slow, detect quality affected by anthropogenic factors and the impact of the work environment difficult to guarantee the conditions of hygiene, to meet modern quality control and process control requirements, the need to study more advanced detection technology to complete quality testing of pharmaceutical products.

Machine vision applications in health care is very extensive, modern many difficult surgery in machine vision with the complete, such as skin testing, the Visual brain CT Imaging, surgery, etc.

Medical machine vision also is applied in some pharmaceutical products, medical and some scientific research, such as glucose liquid detection, detection of chromosomal testing, RBC, etc. [1].

In order to effectively address the modern medicine production line of products for real-time detection issues that this article for glucose solution quality testing has designed a test system.

The design goal is: the use of machine vision technology, use more than one CCD camera, special lighting, work performance of industrial computer to replace the artificial, achieving glucose quality of automatic detection. This experimental system designed to develop high-precision, high-speed automatic detection of smart glucose unit provides effective basis [2].

2 glucose testing process

Medicinal glucose injection of production process as shown in Figure 1, the first through the high concentration of glucose and proper proportion of distilled water are glucose solution, and glucose solution filling into a high temperature and UV disinfection container inside a package.

Package of product after the quality inspection to detect objects primarily:

(1) liquid inside of suspended solids, such as no impurities, fiberglass, etc.;

(2) defects in the bottle;

(3) bottle not cleaned of dirt;

Figure 1 glucose production process

Traditional methods of detection, detection is artificial detect shop located in a separate darkroom, with detection of glucose using artificial manner on the Assembly line, assembly line workers by detection of the speed control, when drugs to detect the presence of workers, workers from the line out, in a dedicated light box judgement whether qualified product quality.

Artificial detect the existence of significant disadvantages:

The modern production line 1) are getting higher and higher speed, manual testing is very difficult to meet the detection rate of line;

2) modern manufacturing stressed real-time, online, non-contact detection to ensure full control of the manufacturing process, increase production efficiency and product qualified rate, this is the traditional testing methods are not available;

3) modern product manufacturing accuracy greatly increases, demand the corresponding high-precision detection method.

4) conventional artificial cannot continue to maintain a consistent test results; different detection differences between; unable to achieve real-time process control;

5) traditional detection cannot adapt to the modern quality control and statistical process control (SPC).

3 test system structures

For glucose liquid inside tiny glass scrap, rubber scraps or other insoluble particles, in a static situation generally precipitation in the bottle bottom or top of the bottle, artificial detect need glucose liquid upside down, so that it occurs in liquid movement, while also driving solid impurities, impurities with drug body motion to a central area, and then trigger the camera to capture images can take a clear image of defective parts or by way of judging the moving object to determine the existence of impurities.

Based on this principle, this experimental system design of a high-speed rotating platform, rotating platform can bring high-speed spin the bottle and then play out, so driving the movement to detect impurities.

3.1 platform hardware chart

As shown in Figure 2, the question detecting glucose bottle jams in go bed, bed time driven rotating spin the bottle for machine vision installed in the dedicated led light source as shown in the source box, CCD camera will image transmitted to IPC.

Following on the hardware platform for details.

Figure 2 glucose impurity detection device chart

(1) handle: according to the figure shown in the direction of rotation knob can lead pallets movement up and down, in order to place the bottle;

(2) motor: normal AC motor, drive the transformation of the rotational spindle speed of bed;

(3) top pallets: glucose pallets in the cradle, and when rotation handhold pallets for convenient placement bottle up and down movement;

(4) light box: light box to place a ring UV source, through the bottle bottom-up radiation medicine;

(5) CCD camera: responsible for shooting sports medicine;

(6) of the IPC: responsible for CCD camera image sequence;

3.2 work process

(1) turn the handle to lift the top pallets, place the question detecting glucose bottle to base;

(2) lay down the handle, close it, open the go card, go bed of spindle rotating speed driven pallets, due to the friction of pallets driving the bottle also do high speed of rotation;

(3) stop go bed, bottle card stop glucose bottle stop spinning, but liquid in inertial effect to continue after a slow stop rotation;

(4) in this process, continuous shooting of 7 images in the sequence of the moving image, moving target for recognition;

(5) when the recognition to exceed the allowed indicator, judging this bottle of liquor is not qualified

3.3 visual imaging

System configuration

Image acquisition part will complete the line of motion image acquisition, image acquisition part acquisition image quality is good or bad will directly influence the detection efficiency [3].

Image acquisition part mainly by CCD camera. CCD camera exposure image signal, from image acquisition card image signal acquisition. The objective of this experiment apparatus selects two orientation of the camera to detect objects to be detected, a down-beat-a side shot, on some question to detect objects can be multi-faceted detection. The camera is Pulnix company TM6703 [4], acquisition card use a Matrox company Cornora2 [5].

Matrox Graphics is a Matrox Corona Ⅱ, production Inc. image controller, you can capture every scan/progressive scan component RGB signals and single/dual black-and-white analog video signal; 3-channel A/D converter 10bit; 24-bit RS-422/LVDS digital interface; simulation case acquisition rate 30MHz, RS-422 digital mode up to 25 MHz, LVDS digital mode up to 40 MHz; connect two RGB or six analog black-and-white video signal; 32-bit/33MHz PCI bus master mode; add-on board for real-time acquisition and storage; support for both analog VGA and independent digital VGA or TV output [4].

Pulnix company TM6703 is 1/2 inch, 648 * 484, shutter speed 1/60 ~ 1/32000 seconds while with asynchronous reset function.

When the pulse effect VINIT, reset the camera's CCD scanning clear. When in asynchronous mode and external VINIT high level signal, asynchronous functionality will automatically be selected, the readout is prohibited until a trigger. The following is the asynchronous reset camera modes:

1) control pulse width external VINIT: shutter speed control with pulse width;

2) fast internal trigger mode: when VINIT's falling edge and the external HD is the same time, signal delay, otherwise the crawl does not, there will be a delay of 0-1HD;

3) slow internal trigger mode: you can select the shutter speed 1/250-1/2000 seconds, if VINIT and external HD-drop along the same, and you start the integral charging, discharging VINIT drop along the camera.

Output delay depends on the selected shutter speed.

4 testing process

Through the previous analysis, detection of suspended matter in glucose experiment need a high speed of rotation of the platform.

The entire test system operation process as shown in Figure 3:

First the system, all device initialization, according to glucose suspended solids test requirements for configuring camera and capture card parameters.

Then, start rotating platform, rotating platform driven rotating 30 seconds for glucose liquid, high speed of rotation stop Rotary platform, delay of 10 seconds after the liquid slowly stopped, at which point the camera image sequences.

Figure 3 glucose impurity detection process

5 innovative point of this article

(1) this article will machine vision applications in a quality inspection for glucose solution, enables non-contact detection, in drug safety and precision are the traditional manual detection has improved significantly.

(2) of this article according to the characteristics of glucose liquid impurities design Rotary measuring device.

(3) according to glucose liquid impurities detection features designed exclusively for Visual Imaging programme.

6 references

[1] Rafael C. Gonzalez , Richard E. Woods .Digital Image Processing (second Edition).Pubulishing House of Electronics Industry , Beijing.

2002.

[2] Zheng Shi-tea, wool seldomly, hubo. machine vision recognition cotton disease patterns feature extraction [J]. a micro-computer information, 2007, 04-1: P290-292.

[3] Wang Wei, Li Tao Wang Nan, etc. Machine vision-based development of prior examination experiment platform. China instrument, 2006, (4) .62-64.

[4] Matrox Corona2 installation and hardware Reference[z].

Matrox Electronic Systems Ltd. April 23, 2001.

[5] PULNIX TMC6703 Manual[z].PULNIX America Inc. 2004.