DE102008031861A1 - Tracking sensor for detecting optical marks on path to guide vehicle, has color filter and comparator that examines scenes using sensor with and without filters and compares scenes with one another - Google Patents

Abstract

The sensor (6) has a color filter (7) that is selected for color spectrum of marks (2) or path, where the marks have color spectrum like the path and physical filter is utilized as filter. A comparator (9) i.e. microcontroller, examines scenes using the sensor with and without filters and compares the scenes with one another, where the marks on the path are differentiated by the filter effect and are fixed in a position. Sensor elements with different spectral sensitivity are utilized for filter. A camera (8) is connected to the color filter.

Description

The The invention relates to a track sensor for detecting optical Trademarks on a railway according to the preamble of Claim 1.

For the tracking of vehicles are linear optical marks or other designations strokes, stripes or contrast stripes, as from DE 100 13 767 A1 known, used. A single mark on the web, e.g. B. a short dash or a point, however, can also be used as a reference or way mark for distance or position determination. It is important that a contrast of the brand to the rest of the web or to the road surface arises. Such a contrast comes about through two adjoining surfaces which differ in their intensity and / or in the wavelength of the reflected light. Preferably, one surface is configured in a light color, including white, and the other surface in a dark color, which includes black. But it can also be similar or equal bright or dark surfaces that differ only in their spectrum, in other words the wavelength, so the color.

through an optical sensor and an evaluation, the contrast can be determined in his position and thus the vehicle this Contrast strips are tracked.

at contrast stripes, contrast patterns, strokes, lines or marks, hereinafter referred to as optical marks or just trademarks, usually have bright or dark, so white / black contrasts enforced. One reason lies u. a. in that black and white tracking sensors, e.g. Black and white cameras, are more available and cheaper than color cameras. In the Scene, ie the track area considered with the track sensor, in the brand is sought, black and white contrasts can be represent with great brightness difference. Strong contrasts can be easily recognized.

Of the Trace sensor consists of sensor elements that emit light in the visible or possibly also adjacent infrared or ultraviolet region convert into a corresponding electrical signal. in principle The technology also works in the IR and UV range, though is subsequently spoken by light or optical sensors.

While one for pursuing a linear mark for tracking two or better several, like over 100 sensor elements, eg. As single-line cameras or 2D cameras Can be used for a transverse brand, only as a waymark is used, a single sensor element suffice. On crossing the brand z. B. detected by changing the brightness.

adversely in the known methods is that when aging, bleaching, Pollution and in particular in case of unfavorable shadow effect or Moisture as well as combination of these disorders, the Marks on the road surface are no longer clearly recognized. disturbing affects, if by a strong shade a wrong railroad / brand contrast is produced. The same is disturbing when vehicles with their wheels from a wet area to a dry one Drive concrete carriageway and then the wet and therefore dark track z. B. is confused with a contrast strip. Also the rubber abrasion from the wheels can due to the then formed light / dark contrast be confused with a brand.

task Therefore, it is the object of the present invention to provide a track sensor for detection of creating optical marks on a web, in particular in case of disturbances in the light intensity, z. B. by Shadow effect, moisture, rubber abrasion and other, especially dark soiling - below Generally called disturbances - also a reliable one Detection of optical marks possible.

These Task is solved by a track sensor according to claim 1. The other claims contain further inventive Versions.

It is advantageous according to claim 1, the scene train / brand also with the help to look at a color filter. Assuming, of course, that brand and train differ in the color spectrum. If For example, a bright concrete carriageway without a filter is considered more or less the entire color spectrum, that is red, orange, yellow, green and blue. For a pure yellow brand however, only the yellow portion of the spectrum becomes visible. On consideration of brands and railway by a yellow only permeable Filter the yellow mark then hardly in their intensity subside, while from the web now only the yellow Light is visible. The light components red, orange, green and blue is eliminated, and thus the output is at the sensor element correspondingly smaller. The sensor elements that look at the web become a much smaller light intensity with a filter Show as without a filter.

Of the Comparator or evaluator will therefore find that here the railway is considered. If another sensor element no significant change in intensity compared with and without filters indicates is obviously just a brand or considered part of a brand Service.

in the Principle is also with the use of a yellow blocking filter when considering a yellow mark the proof of that mark is possible, because then no light is received from the brand. However, you can the yellow mark then only distinguished from a black mark or course because it looks brighter when viewed without a filter the black mark or train.

The Procedure is also appropriate if the train itself rather dark is, for. B. in asphalt or a certain color spectrum, z. B. yellow and the marks are white or black or too have a different color spectrum, eg. B. blue marks on red train, etc. In the latter case, the use also provides of two color filters, e.g. B. a blue and a red filter at.

Important is that the relative attenuation of light intensity with filter in relation to the measurement without filter is evaluated. Looking at a yellow mark with and without Yellow filter will hardly make a difference here. Of course, the same scene sections should be considered become.

The Deciding if what the sensor element is looking at the web or belongs to a brand hangs in this example from the yellow part in relation to the residual light component. The absolute light intensity plays no role here. In This color example is the contrast by a possible high yellow content in one area and neutral or the lowest possible yellow content in the other area educated.

different Light conditions by z. B. shadowing, wet spots or black tire wear then do not bother or less.

Of the Track sensor is used to detect the optical marks on the web usually use multiple optical sensor elements, each for itself, to different, but usually adjacent scene segments are directed. The sensor elements can be part of a or multi-line camera and thus with each shot the scene depict.

Around to detect the differences in the colors of brands and railways the scene is then captured with a filter that either the brand or track color is disabled. In another recording is omitted this filter or used another filter if necessary, then possibly the other color locks to compare the effect of first filter to increase. The following is generally of Comparison with or without filter spoken, even if two different Filters are used. Then a filter locks the track color and the other filter leaves the brand color or a filter only the brand color through and the other filter only the track color.

These Description uses examples for a better understanding the idealized conditions, eg. B. pure red or pure yellow, etc. and ideal filters. The recognition of the brands on the train, of course, also works with the usual Filter and the usual colors.

To Claim 2 is the track sensor for guiding a vehicle used. The track sensor is attached to the vehicle and on aligned the roadway. It is the position of the brand, z. B. a contrasting strip, in relation to the track sensor or the vehicle detected. This allows the vehicle to be guided automatically if necessary be and / or used for driver assistance. The driver can be warned if it is too far from the lane or removed the linear mark.

The line-shaped mark does not have to be uninterrupted. It can also be broken lines or generally broken Trade lines. The stamps can even be made from individual ones There are points with large gaps in between. However, it is a disadvantage if the interruptions get so big that the leadership quality of the vehicle underneath suffers.

To Claim 3, the track sensor but also for the detection of way marks be used. Such brands may u. a. also as Strokes can be arranged transversely and / or longitudinally to the roadway. she essentially serve to provide information, quasi as a milestone for determining distance, as an indication of speed limits, Turns, stops, etc. Simple waymarks can be over a temporal change of color distribution found become. Waymarks can also be coded two-dimensionally be so that several sensor elements are necessary for detection.

Also Line-shaped marks can by intentional Interrupts coded. The information may be in the Brand length, the length of the interruption and / or the frequency of interruptions or combinations thereof. The track sensor can be used both for detecting linear marks, as well as for the detection of way marks.

The brands can have different color spectra, so z. B. classically red, orange, yellow, green or blue. Thus, a red mark can also be understood as a slow or stop mark, while a green mark is evaluated as an express track. Even brands with unterschiedli chen colors can side by side or in constant change, for. B. red / green / red / green are used for security reasons.

Consequently is detected by detecting the one color mark with simultaneous recognition the other or when recognizing the color change the evaluation with the other evaluation to be verified. With the number of used Colors also increase the need for filters.

To Claim 5, the recognition of the brand u. a. with physical Filter, z. As a color-selective disc or optics performed. When using physical filters can be inexpensive Insert black and white cameras. Suitable is u. a. the Use of two cameras, one of which z. B. a filter has that only lets through the chosen brand color. The other camera essentially looks at the same scene. However, this camera does not need a filter. In another Evaluation is also the use of a filter possible, which locks the brand color and lets only the track color through. It is important that the results with filters with the results be compared without filters. Of course you have to both cameras look at the same scene and the sensor elements assign a camera to the sensor elements of the other camera to let.

in principle It is also necessary to use only one camera. Then you have to the scene one after the other with once and without filters considered become.

Even the use of a 2D camera with a split filter is possible.

physical Filters are available as wide or narrow, as high or as Low pass or as a band or notch filter available. Thus, certain color spectra may preferably be transmitted or also be filtered out. Possibly. are combinations of different Filter suitable for a good filtering effect. physical Filters can be made very narrow band, which also a high filtering effect is achieved, although the mark or web to have a corresponding narrow-band spectrum.

At long last are all combinations regarding white / black or all color variants possible. The only important thing is that brands and train in the color spectrum differ.

To Claim 6 can be used for the track sensor sensor elements different spectral sensitivity, z. B. a color camera deploy. Usually, color cameras have a lot of sensor elements, the one-third red, green or blue sensitive are. With these color cameras, however, it should be noted that their filter effect is relatively weak and the colors partly not clearly assigned can be. Because there are no real ones with these color cameras Yellow filter gives, you can not tell if it is the yellow is a pure yellow or whether it is a mixed color of red and green. Therefore, the color cameras have in the application often a worse quality than black and white cameras, using a color-selective filter disc or optics.

color cameras can then be useful if the main component of Brand color in the spectrum is as far away as possible from the main component of the railway paint. The respective main components should then match the colors for the different ones Sensor elements are sensitive, so for example. Red or green or blue. A suitable example of a color camera would be the use of red marks on blue or vice versa. Here, the filter effect of a color camera may be sufficient be. So that in this application, not the light intensity, but only the relative color share of red or blue is evaluated, can then in the observed scene segment, the entire light of red, green and blue sensor elements together, so to speak be used as unfiltered reference value. In this example But it is also possible that only the ratio Red to blue is used to decide whether it is a railroad or a brand segment.

Of the Advantage of the color cameras lies more in the fast configurability or conversion to other color compositions.

A Filter effect can also be determined according to claim 7, characterized that scene with artificial, colored light is illuminated. In bright sunshine outdoors This can be difficult. The use is therefore more appropriate for shaded areas, indoors or at night meaningful. Often have lights of roads or work surfaces a relatively strong yellowish sting. It would then make sense, with an artificial light, the rather strong blue parts have the scene to be viewed out of orbit and brand to light. The additional bluish Light should be switched on and off alternately, eg. B. as a flash. In this case, then the track sensor, z. B. with a camera, respectively a scene with the yellowish light and the bluish light Record separately and then compare. At a yellow mark On a light or even white concrete track is then at the bluish flash light a stronger brightening occur on the white concrete track, as on the yellow Brands. For other brand and train colors are different accordingly To choose flash colors.

The Filter is not directly in front of the optical in this application Sensor element, but before the artificial lighting or is part of the artificial lighting. advantageously, provides the artificial lighting, physically conditioned, essentially exactly the appropriate color spectrum. The artificial one Lighting is then part of the track sensor, even if they is slightly spaced.

It is also possible a white artificial Lighting such filters to use that one part the scene with filter and the other part of the scene without filters lit up is. When driving a vehicle will be the same Scene part viewed and evaluated one after another with and without filter. In particular, if the general lighting is relatively weak, It makes sense, two differently colored lighting to use, for. B. a reddish and a bluish. If marks and orbit differ in their red and blue values, This difference can be synonymous with a black and white camera be determined.

at Change of the general brightness, z. B. by the lighting with bright sunlight during the day and relatively weak artificial light in During the night, strong light fluctuations occur. The track sensor is according to claim 8 designed so that it with or without sunlight can work. If necessary, the vehicle will be an on-board vehicle Use lighting. Furthermore, the track sensor fits in the Sensitivity, e.g. B. by controlling the aperture and / or exposure time, at. In order to avoid overdriving, it is advantageous in case of a light track corresponding to the sensitivity of the track sensor to regulate the brightness of the roadway.

at Use of multiple cameras or use with and without Filters can have different sensitivities or apertures or Exposure indicators are chosen so that appropriate Signals for comparison and evaluation are available.

Embodiment:

Based In the accompanying drawings, the invention is intended below be explained in more detail:

It shows:

1 the principle of a linear brand recognition with two sensor elements.

2 the same principle with variations in light intensity by shading.

3 a block diagram of the track sensor with filter, camera and evaluator.

In 1 is a brand 2 in the form of a strip on a roadway 1 to see. At the left edge of the mark 2 is the left detection segment 3 a sensor element and on the right side to see the right detection segment of a second sensor element. A detection segment is the area on the roadway 1 or brand 2 which is viewed by a particular sensor element. The area outside this detection segment is not seen by this sensor element.

Below this is the intensity profile for an optical sensor element without filtering. If the sensor element is above the yellow mark, only the relatively low light intensity of the narrow yellow spectrum is used for evaluation. The mark 2 It is ideally made of pure yellow light and not of a red, orange or green color mixture. Right and left of the mark 2 is the higher intensity of the bright reflection from the roadway 1 to recognize. Doing that on the road 1 If the red, orange, green and blue light components are added, the output signal at the sensor element is of course stronger here. Assuming in this example that all colors red, orange, yellow, green and blue contribute 20% each to the output of a sensor element and no filter 7 is used, then the output is for the yellow mark 2 at 20%.

In the transition area between the mark and the carriageway 1 Changes output signal on the sensor element according to the shift.

When the track sensor 6 is shifted to the left with two sensor elements, the left sensor element receives more light, the right sensor element is then weaker. Thus, by comparing the output signals on the left and right sensor elements, one can see the shift to the left.

In 2 is a fault in the right half 5 to recognize. This may be z. B. a shadow, a wet lane on an otherwise light road 1 or even black tire abrasion. If one evaluates only the brightness on the disturbed surface, without consideration of the color this disturbance could 5 with a brand 2 be confused. The scene is from a single-line camera 10 , which has a plurality of sensor elements considered. Shown is almost the projection of the single-line camera 10 or the adjacent scene segments, which are seen by the individual sensor elements of the single-ended camera.

Underneath, the corresponding signal image without filter SOF is shown in the diagram. Without filter 7 the entire light spectrum is transmitted. In the areas 12 and 13 There is a reduction, because here the red, orange, green and blue light components are missing and in this example then only the yellow light is detected with only 20% of the total light intensity. A further weakening takes place in that in the areas 13 and 14 through the disorder 5 once again 75% are filtered out. As a reference signal with the comparison strength 1 applies the unfiltered and undisturbed light road 1 ,

In another, lower-level view, the entire scene is shown with the signal with yellow filters SGF. The white roadway 1 and the yellow mark 2 are in the fields 11 . 12 and 15 equally strong. However, there is a weakening due to the disturbance 5 in the fields of 13 and 14 ,

The effect of the filter 7 can be seen in the filter effect FW diagram. Here it turns out that in the fields 12 and 13 capturing the scene with filters 7 versus detection without filters 7 has brought no change. The filter 7 is neutral in the filter effect FW, ie has the neutral factor 1. This range corresponds to the yellow mark 2 , The areas 11 . 14 and 15 show that the filter 7 let only 20% of the light through. This area is therefore a carriageway 1 ,

A simple block diagram is in 3 shown. track sensor 6 consists in this example among others of a camera 8th that a filter 7 can be switched on. Furthermore, there is a comparator 9 pictured the scene from the lane 1 and brand 2 considered and with and without filters 7 evaluates.

The comparator 9 can a small computer, z. As a microcontroller, be. The camera 8th can send analog or digital signals to the comparator 9 deliver. The physical filter 7 may be a glass or plastic disc with a highly selective action, e.g. B. with a narrow passage in the yellow area.

The track sensor 6 may be attached to a vehicle, not shown, and for tracking over a linear mark 2 walk along.

A deviation to the left or right is thereby from the track sensor 6 detected and possibly used for a steering correction.

Likewise it is possible that it is at the mark 2 is a waymark and with the track sensor 6 this brand 2 to the reference point or as information on a roadway 1 is recognized.

With the track sensor 6 it is possible according to the design brands 2 to recognize the tracking as well as for referencing or information.

QUOTES INCLUDE IN THE DESCRIPTION

This list The documents listed by the applicant have been automated generated and is solely for better information recorded by the reader. The list is not part of the German Patent or utility model application. The DPMA takes over no liability for any errors or omissions.

Cited patent literature

• - DE 10013767 A1 [0002]

Claims (8)

Track sensor for detecting optical marks on a web, characterized in that - the marks have essentially a different color spectrum than the web, - that by means of a color filter, which is selected for the color spectrum of the marks or the web and a comparator, the View and compare the scene with the track sensor with and without filters and - that the filter effect distinguishes the marks from the web and determines their position. Track sensor according to one of the preceding claims, characterized, - that the track sensor for Guide a vehicle is used. Track sensor according to one of the preceding claims, characterized, - that waymarks on the roadway are detected Track sensor according to one of the preceding claims, characterized, - that depending on the meaning of Brands different color spectrums can be used. Track sensor according to one of the preceding claims, characterized, - that as a filter a physical Filter is used. Track sensor according to one of the preceding claims, characterized, - That for filtering sensor elements be used with different spectral sensitivity. Track sensor according to one of the preceding claims, characterized, - that the filter effect by a change in the color spectrum of the lighting is achieved. Track sensor according to one of the preceding claims, characterized, - that the sensitivity is adjusted and thus also in different lighting conditions can be worked.