The X-ray system for quality control uses a Pentium PC to manage the detection process and all the functions of the system. The picture below shows the general layout of the Inspection Unit highlighting the main interactions between the basic elements of the system. In particular, the PC and the VIRTUAL PLC carry out all the main processing and control phases as well as interfacing with the other elements of the unit.
The user interface allows to modify the setup parameters of a product elaboration and change the status of the system. Except for extraordinary repairs, user intervention on the machine is restricted to the user interface.
The Virtual PLC informs the user interface of any alarms or faults on the inspection line. The User interface and Virtual PLC exchange information regarding the characteristics of the type of product to be inspected in order to adapt line status to the product. Concerning timing, motor speed, reject action, etc. (I/O Management) and the status of the interface (operating, in maintenance mode, in automatic mode) other signals are provided for calibration of the detector. Another important signal exchanged between the PC and Virtual PLC allows management of faulty product reject.
The process of X-ray formation starts when a signal is sent from the PC to the X-ray generator containing information necessary to set power level and frequency output of the X-ray generator.
When the signal arrives to the generator, the X-ray generator output will be again amplified, and it will be sent to the X-ray Tube via a high voltage cable.
When the high voltage signal strikes the target (Anode) in the tube, X-rays are emitted.
X-ray energy is received and processed using a detector.
The sensitive element which receives X-ray energy, within the detector, is called CCD (Charged Couplet Device), which transforms X-ray energy into an analog output.
The analog output is sent to an Analog to Digital converter (ADC) and a number (equivalent in the grayscale to the amount of x-ray energy received) is generated.
The amount of X-ray energy a contaminant absorbe is relative to its density and size. Looking at these grayscale readings it is easy to understand that there must be a contaminant in the product, because the grayscale value in the body area remarkably decreases.
The general processing parameters of the X-ray detection are used to find zone of the object with have different dense value from the average value. In this zone usually we can find foreign body.
The very light zones of the image (moving from the product to zones external to the product) are generally not very homogeneous from a photometry point of view (dark/light anomaly variations), and these irregularities could lead to erroneous pinpointing of contaminants.
Processing methods (algorithms) based on the analysis of the photometry value for single image pixels generally allow medium or large sized contaminants or contaminants with high contrast (metal) to be pinpointed.
It is important to note that, once the threshold level has been defined, all pixels with a photometry value less than the threshold value are considered anomalous pixels, but only those that meet very precise photometry and morphological conditions are actually considered pixels corresponding to contaminants.
Processing methods (algorithms) based on the analysis of the photometry contrast for single image pixels generally allow medium or small sized contaminants or contaminants with quite high contrast to be pinpointed.
All pixels with a photometry contrast value less than the threshold value are considered anomalous pixels.
The processing method (algorithms) based on the analysis of the photometry gradient of the pixels in the image generally allows contaminants with quite a high contrast with respect to the image to be pinpointed, regardless of the size of the contaminants.
In addition, the gradient method also allows contaminants with little contrast to be pinpointed, provided they are characterized by angular surfaces or have fracture plains that are perpendicular to each other (see figure underlying).
The figure shows the main steps involved in the gradient method: starting with the image of a certain product, the photometry gradient value for each pixel is determined, and all pixels with a gradient greater than the parameter value are considered anomalous pixels. The anomalous pixels pinpointed in this way are used to construct groups of anomalous pixels, and all the anomalous groups thus defined with specific photometry and morphological characteristics are considered contaminants.
Analysis of the contrast value for dark groups is a very powerful algorithm, capable of pinpointing a vast range of small and medium sized contaminants, even with little contrast.
Starting from the individual dark pixels pinpointed, the system automatically constructs dark groups of which the dark pixels are a part, based on a certain number of photometry and geometric rules, as shown in figure. For each dark group the photometry contrast value with respect to the zones adjacent to the image is then determined, and all groups with a contrast value greater than the parameter value are considered contaminants.
Detectable products features
Contaminant detection performances depend on the following product characteristics : thickness, homogeneity, density.
When homogeneity and density are the same, a thick product is more difficult to inspect compared to a thin one because the former creates a darker image than the latter; this happens because the x-ray energy which results on the detector (see previous paragraphs) after crossing the thick product is less than the energy which results on the detector after crossing the thin one.
- Thick cheese - Thin cheese
When thickness and density are the same, an homogeneous product is easier to inspect than a non homogeneous one, because the former creates an image with less "noise" than the latter. In fact a change of homogeneity can deceive the inspection system and inducing to detect a contaminant.
- Hamburger - Bacon
When thickness and homogeneity are the same, a product with higher density is more difficult to inspect compared to a less dense one, because the former creates a darker image than the latter; this happens because the x-ray energy which results on the detector after crossing the higher density product is less than the energy which results on the detector after crossing the lower density one.
- • High density cheese - Low density cheese
X-ray machines are equipments are completely compatible and can be easily be connected to the customer's production line.
Installing an X-ray machine does not mean revolutionizing the production line, but simply adding a final quality control.
New technologies allow the monitoring of information coming from the inspection system via a single remote workstation (even for more than one equipment on the production lines of the same factory):
- monitoring of the system status
- possible alarm signals
- production statistics
It is not possible to modify the system status.
Furthermore, thanks to the remote control kit, a skilled technician can take over the control of the inspection system, thus avoiding him to physically visit the customer premises.