DE29520980U1 - Device for testing a light-sensitive matrix sensor, in particular an electronic camera with a CCD matrix sensor or CMOS matrix sensor - Google Patents

Description

The innovation relates to a device for testing a light-sensitive matrix sensor, in particular an electronic camera with a CCD matrix sensor or CMOS matrix sensor according to the preamble of claim 1.

The protective devices used in safety technology are mainly light grids that work with an active light transmitter and an associated passive light receiver. A protective field can be created through the light path of one or more light beams, whereby the light sensor detects an interruption in the light beam and triggers protective measures, such as switching off a machine. Since light beams only spread in a straight line, the boundaries of the protective field can also only ever be straight.

so that a protective field built around a machine is usually a rectangle or polygon. For each line of this protective field, you need either a separate transmitter/receiver pair or at least deflecting mirrors that give the light beam a new desired direction. Another problem with the known light grids is that people and any objects cause the beam to be interrupted when entering or leaving the protective field. In many cases, however, it is desirable to have a selective protective field that allows certain objects, such as workpieces, robots or industrial trucks, to enter the protective field without triggering safety measures, but not people.

In an article by W. Grigulewitsch, "Modern Sensor Technology in Security Technology", special edition of "Die BG", issue 3/94, Erich Schmidt Verlag, Bielefeld, pp. 1-6, security technology systems based on infrared, laser light and ultrasound are examined, in particular with regard to the extent to which these sensors can distinguish between people and objects.

In order to check the sensor function of the infrared sensor, it is suggested, among other things, to monitor the thermal background noise of the sensor and to carry out a cyclic test with a heat pulse by briefly heating up a resistor mounted in front of the viewing window of the infrared sensor, thereby generating a heat pulse of defined size and evaluating the signal thus generated for amplitude and polarity using comparator thresholds.

On the other hand, this paper acknowledges that thermal sources of interference affect detection reliability, so that additional measures are proposed to eliminate possible interference.

When protecting property, such as buildings, land

·3
·

etc., electronic cameras have long been used, which often have a light-sensitive sensor in the form of a CCD matrix. However, monitoring is carried out using a monitor that must be observed by a person.

In an article by H.G. Graf et al. "Electronic vision", DEZ Elektronik 3/1995, pp. 52-57, a light-sensitive sensor based on a "CMOS matrix sensor" with high brightness dynamics was proposed, which is particularly suitable for surveillance tasks.

In general, it is desirable to use television cameras in protective devices, as they have several advantages compared to light curtains:

The television camera can be installed above a protective field and therefore takes up less space;
the protective field boundaries can be freely programmed so that any complex protective field configuration can be created;

Without any additional equipment, different protective field zones can be defined around a machine or, more generally, around an object, such as a warning zone and one or more protection zones, which trigger different measures when entered;
Finally, television cameras make it possible to distinguish between people and objects and also to distinguish between objects and thus trigger different measures when certain objects are detected, such as certain workpieces, etc. For example, a workpiece can be allowed to pass through to a machine without triggering a safety-relevant function, while the machine is stopped when a person enters the protective field defined around the machine.

However, a fundamental problem with the use of television cameras in security systems is checking the functionality or availability of the camera.

• ·4 · ♦

For example, due to insufficient lighting of the protective field monitored by the camera, the contrast between the people or objects to be detected and the surroundings can be so low that the person or objects can no longer be identified. The same can also occur due to dirt on the viewing window or the lens of the camera, despite sufficient lighting. Individual parts of the camera can also fail, such as some pixels of the CCD sensor matrix, without this error being detected, which then leads to at least a reduction in the safety function, which is unacceptable in any case, especially in the field of security technology.

The aim of the innovation is to improve the device of the type mentioned at the beginning so that a reliable self-test of the matrix sensor can be carried out. This task is solved by the features specified in claim 1. Advantageous embodiments and further developments of the innovation can be found in the subclaims.

The basic principle of the innovation is to arrange a predefined, multi-dimensional pattern in the detection area of the matrix sensor and to compare the signals generated by the matrix sensor when scanning this pattern with stored signals and, depending on this comparison, to generate a signal that indicates the availability of the matrix sensor.

The comparison signals are an image of the predefined multidimensional pattern, so that if the pattern detected by the matrix sensor and the stored comparison pattern match, one can assume that the sensor is working properly and that the lighting of the area monitored by the matrix sensor is sufficient, since otherwise the predefined multidimensional pattern introduced into the detection area would no longer be recognized. This pattern

can be constructed as complex as desired, so that the entire detection range of the matrix sensor can be checked. This pattern can also be changed over time, so that it is composed of any number of individual patterns that follow one another in time.

This test is preferably carried out cyclically. In a variant of the innovation described below, this test can also be carried out continuously. The innovation provides several variants for generating the predefined, multi-dimensional pattern, which is referred to below as the test pattern. In a first variant, the test pattern is generated by a liquid crystal display (so-called LCD) arranged in the beam path in front of the matrix sensor. Such LCDs are in turn constructed in the form of a matrix and can change their light permeability by applying electrical voltages, which can be used to generate any test pattern. These test patterns are superimposed on the image captured by the matrix sensor and displayed on the matrix sensor, where they are primarily displayed as differences in brightness. This test pattern is then captured by the sensor, converted into electrical signals and compared in an evaluation unit downstream of the matrix sensor with stored signals from a pattern, which is referred to below as the reference pattern. If the test pattern and reference pattern match, the sensor is ready for operation and fully available.

The degree of agreement between test and reference samples can be changed depending on safety requirements or environmental conditions. For example, a proper sensor function can be signaled if the degree of agreement is 95° or greater.

This variant has the advantage that the lighting of the monitored area is included in the test, since the matrix sensor can only recognize the test pattern if

if the brightness differences between the areas darkened by the liquid crystal display and the areas not darkened are still within the resolution range of the sensor. If this is the case, it can be assumed that objects that are to be detected, which naturally have to have a brightness difference, color difference or gray value difference compared to the surroundings, are also detected. This variant for checking the sensor can naturally only be carried out cyclically.

In a second variant of the innovation, a predefined test pattern is placed in the area monitored by the sensor and remains there permanently. For example, one or more strips with different gray values are placed at the boundaries of the protection field. The sensor compares the image captured by these strips with a corresponding stored reference pattern. As long as these gray value differences are still detected, it can be assumed that the lighting of the monitored area is sufficient and that the sensor is working properly at least in the area of the pixels of this test pattern. This variant of the innovation can work continuously, in the sense that the comparison between the test and reference pattern is carried out constantly, or intermittently in such a way that the test pattern is constantly captured, but the comparison with the reference pattern is only carried out in predetermined cycles.

In a third variant of the innovation, the test pattern is generated by projecting a pattern from a light source into the monitored area, for example using a slide projector, whereby this pattern can also be in the invisible wavelength range of light, for example in the infrared range, so that the test pattern is invisible to humans.

In the following, the innovation is explained in more detail using examples of implementation in connection with the drawing

explained. It shows:

Fig. 1 shows an embodiment of the innovation with a test pattern permanently arranged at the boundaries of the sensor’s detection range;

Fig. 2 shows an embodiment of the innovation in which the test pattern is generated by a liquid crystal display arranged in the beam path of the matrix sensor;

Fig. 3 a block diagram of the sensor, the liquid crystal display and an evaluation device; and

Fig. 4 shows another embodiment of the innovation, in which a test pattern is projected from a light source into the detection area of the matrix sensor.

In Fig. 1, a first protective field 2 and a second protective field 3 surrounding it, which can function as a warning zone, for example, are defined around an object 1 to be monitored, which can be a machine tool, for example. A camera 5 containing a matrix sensor is attached to a ceiling 4 above the object 1 and the protective fields 2 and 3. Furthermore, next to the camera 5 on the ceiling 4, a lighting system 6 is attached, the light beams 7 of which are directed at the object 1 and the protective fields 2 and 3.

A pattern is applied around the outer protective field 3, which here is formed from two strips 8 and 9 running parallel to each other, which have different brightness values or shades of gray in relation to each other and/or in relation to the environment.

The camera 5 is aligned so that it detects the object 1, the protection fields 2 and 3 and the strips 8 and 9, which are illuminated by the lamp 6. Since the strips 8 and 9 are fixed, for example on the floor

If they are glued on and have known brightness values, shades of gray and/or colors, they form a predefined two-dimensional pattern that lies within the detection range of the camera and its matrix sensor.

As will be explained in more detail in connection with Fig. 3, this test pattern formed by strips 8 and 9, possibly in interaction with their surroundings, is compared with a reference pattern.

Furthermore, in Fig. 1, a test body 10 is arranged in the protective field 3, the brightness, color and spatial configuration of which are also known. This test body is used to calibrate the camera and can also be part of the test pattern.

Fig. 2 shows another example of the innovation. The camera 5 has a housing 11, the light entry opening of which is closed by a liquid crystal display (so-called LCD) 12. External light, which is indicated by dashed lines, penetrates this liquid crystal display and reaches the matrix sensor 16 of the camera via a first converging lens 13, a pinhole 14 and a second converging lens 15. The liquid crystal display 12 can darken predefined areas by applying an electrical voltage, which then form the two-dimensional test pattern in conjunction with non-darkened areas. In the event that the area to be monitored is black or dark gray and the LCD test pattern would not provide sufficient contrast to the dark or gray background despite adequate lighting, the LCD display can be externally illuminated during the test phase, for example by infrared LEDs that are arranged outside the camera 5 and activated by a control system.

In Fig. 3, the camera 5 and the liquid crystal display 12, which serves as the viewing window of the camera,

The matrix sensor 16 of the camera transmits signals corresponding to the captured image to an image preprocessing and image recording control 17, which is indicated by the line 18. The image recording control is carried out via the line 19. The image preprocessing is carried out via a scanning processor of known design, which has a video output 17' to a monitor (not shown), a digital image signal output 22 to two processors 20 and 25 and a control input 21 which is connected to the first processor 20. The digitized image signal from the output 22 is fed to both processors 20 and 25, which each carry out the comparison independently of one another, i.e. in two channels. The second processor 25 controls the LCD 12 via a line 28 and an amplifier 29 and generates a defined reference pattern on it, which is stored in a memory module 25 ' of the processor or an external memory 25" and is fed to both processors 20 and 25. If necessary, this reference pattern can also be stored in a further memory 20', which is assigned to the first processor 20, so that both processors 20 and 25 receive the same reference pattern independently of one another. Both processors 20 and 25 then compare the image information of the LCD 12 present at the output 22 with the stored pattern.

As soon as a processor comes to a negative comparison result, the safety outputs 23 and 30 of both processors enter a locked state (machine STOP), regardless of whether a person is in the monitored area or not. Both processors 20 and 25 are connected to each other via lines 26 and 27 and also immediately exchange their comparison results with each other.

The processor 20 has a further control input 24, via which the protective field boundaries, certain patterns that may enter the protective field without triggering an alarm, etc., can be programmed.

• *

·

Since the evaluation is preferably carried out on two channels during normal operation, i.e. via the two independent processors 20 and 25, these programming signals on line 24 can also be fed to the second processor 25 via line 26.

Fig. 4 shows a third embodiment of the invention, which differs from that of Fig. 1 essentially in that the two-dimensional pattern is generated by a projector 32, which is also mounted on the ceiling 4 next to the camera 5. This projector 32 projects a test pattern 33 onto the object 1 to be monitored and an area surrounding it. This test pattern is shown here as a "checkerboard pattern", but can have any configuration. During the projection of this test pattern, the lighting 6 is preferably switched off. Of course, it is also possible to integrate the lighting 6 and the projection device 32 into a unit and to generate the test pattern by having an LCD arranged in the beam path of the illumination light generate a predefined two-dimensional pattern by applying electrical voltages. The evaluation, i.e. the comparison between the test and reference patterns, takes place in an analogous manner to that described in more detail in connection with Fig. 3.

Claims (4)

Protection claims 1. Device for testing a light-sensitive matrix sensor, in particular an electronic camera with a CCD matrix sensor or CMOS matrix sensor with an evaluation unit to which output signals from the matrix sensor can be fed, characterized in that a predefined multi-dimensional pattern (12; 8, 9; 33) is present in the detection area of the matrix sensor and that a comparison device (20, 25) is provided which compares the signals generated by the matrix sensor (16) with signals stored in a storage unit (20'; 25'; 25") and, depending on this comparison, generates a signal which indicates faulty or fault-free operation of the sensor. 2. Device according to claim 1, characterized in that a controllable pattern generator (12) is arranged in the beam path in front of the matrix sensor (16), which generates a pattern consisting of different shades of gray, different brightness values and/or different colors as a function of control signals, which pattern forms the predefined multi-dimensional test pattern. 3. Device according to claim 1, characterized in that a projector (32) is provided which projects a predefined pattern with different shades of gray, different brightness and/or different colors onto objects lying in the detection area of the matrix sensor (16). 4. Device according to claim 1, characterized in that the predefined multi-dimensional pattern is formed by one or more strips (8, 9) which • · &bgr; · or which have different shades of grey, different brightness and/or different colours compared to the surroundings or to each other and are arranged in the detection range of the matrix sensor (16). Device according to one or more of claims 1 to 4, characterized in that the evaluation unit (2) has evaluation units (20, 25) that operate independently of one another, to each of which the output signal of the matrix sensor (16) and signals corresponding to the reference pattern from one or two independent memories (25" or 20', 25') can be fed, and that both evaluation units (20, 25) exchange their evaluation results with one another and only generate a signal indicating trouble-free operation if the comparison results of both evaluation units (20, 25) are the same.