WO2010094832A2

WO2010094832A2 - Method and arrangement for monitoring sparking

Info

Publication number: WO2010094832A2
Application number: PCT/FI2010/050067
Authority: WO
Inventors: Jani PÖNKKÖ; Kari PÖNKKÖ
Original assignee: Sesca Group Oy
Priority date: 2009-02-06
Filing date: 2010-02-05
Publication date: 2010-08-26
Also published as: WO2010094832A3; FI121443B; FI20095115A0; FI20095115A

Abstract

In the invention sparking in the object is monitored with a camera (201) and an image processing unit (203). The object is imaged and from the images thus obtained an image pair is formed in the image processing unit, in which image pair there is a first and a second image. Images of the image pair are compared to each other and those pixels that have values that are substantially the same in the corresponding pixels of both images and those pixels that have a value lower than the set threshold value are removed from the first image. The number of pixels remaining gives the sparking value that represents the sparking of the object. To compare the pixels of the image pairs there is a comparator (204) of pixels values in the image processing unit. The pixels that do not fulfil the comparison values are removed with the image reducer (205) from the first image. The number of pixels remaining in the first image is determined with the counter (206), which number is the sparking value, which is proportionate to the sparking occurring in the object at the time when the image was taken. The sparking value is delivered to the system (210) monitoring the object, which system undertakes the necessary actions if sparking exceeding a certain limit is detected in the object.

Description

Method and arrangement for monitoring sparking

The invention refers to a method for monitoring sparking occurring in an object, in which method the object is imaged and the intensity and quantity of the sparks in the image is determined. The invention refers also to an arrangement for monitoring sparking occurring in an object, which arrangement also comprises a camera for imaging the object and an image processing unit.

A spark is a small airborne hot glowing object. These are usually generated in processes of combustion or as the source of the spark heats up enough for exam- pie as a result of friction. Also when substances in different temperatures and states get in contact with each other sparking can occur. The spark itself can be either solid, molten or partially molten material. In electric sparking spark-overs occur between different potentials, which spark-overs produce a flash of light. Sparking is generally a chaotic and uncontrolled process, which is in many envi- ronments a unwanted event. For example in metal industry there are processes, in which sparking of the material being handled is a malfunction that should be detected quickly. One example is continuous casting machine, in which the molten metal is guided through a water-cooled mould.

The patent publication US6903357 presents a counter for electrical sparks. There is a light detector that produces electric current, when light hits it. The amount of sparks is estimated by measuring this electric current. The device is meant to measure an object that is located quite near, which object has been placed in a chamber isolated from outside light. This limits the use of the invention.

Optical fire detectors are known, that aim to identify flames from images captured by a camera. In general image recognition is used in these, such as in patent publication US5937077. These cannot be applied to monitoring sparking, because sparks are small, fast moving objects and a flame is a substantially stationery object. In these detectors flames presented in the image are formed to patterns that are aimed to be recognised from consecutive images. This requires quite a lot of image processing and mathematical procedures. Furthermore they generally use infrared cameras, that are quite expensive and are not suitable for all environments, such as for example for such an environment in which there are a lot of hot objects to begin with. An objective of the invention is a solution, with which the disadvantages associated with the prior art can be significantly reduced. The goals according to the invention are achieved with a method and an arrangement, which are characterised by what is set forth in the independent claims. Some preferred embodiments of the invention are presented in the dependent claims.

The basic idea of the invention is as follows: The object is imaged and an image pair is formed from the images thus obtained. Images of the image pair are compared to each other and those pixels that have values that are substantially the same in the corresponding pixels of both images and those pixels that have a value lower than the set threshold value are removed from the second image and the number of pixels remaining gives the sparking value that represents the sparking of the object.

In the method according to the invention an object is imaged and the intensity and quantity of the sparks in the image is determined. Imaging of the object produces digital images, in which the value of each dot, or pixel gives the brightness in the corresponding spot of the object. In the method an image pair is formed, in which there is a first and a second raw image. A raw image is an image selected from the images produced by the camera that is used to count the sparking value. Both raw images of the image pair are in their own files, and in these raw images the value of each dot, or pixel is a binary digit, that gives the brightness in a corresponding area of the object. The two raw images of the image pair, the first and the second raw image, are compared to each other and those pixels in the first raw image, that have substantially the same value as the corresponding pixels in the second raw image, are removed from the first raw image. Also those pixels that have a value lower than the set threshold value are removed from the first image. From the reduced image obtained above the number of the pixels, or the spark value is counted, and monitoring data is formed from the sparking values to be delivered to the system monitoring the object.

In an embodiment of the method according to the invention an image pair is formed in the image processing unit from two consecutively captured raw images. In another embodiment the image pair is formed so that the latest raw image is the first raw image and the former first raw image is the second raw image. In a third embodiment the image pair is formed so that the latest raw image is the first raw image and some former raw image is the second raw image. Then the second raw image, to which the first raw image is compared to, stays the same in several image pairs. This is advantageously such an image representing the object, in which there is either very little sparking or none at all. This second raw image can be changed if there are changes in the object. These can be lighting changes or changes in the function of the object. The interval between the capturing moments of the raw images is constant and the interval between image pairs is constant. These constant intervals do not need to be the same.

The image pairs can be formed for example in the following ways. Image pairs are captured at five seconds intervals, the difference of the capturing moments of which images is half a second. The image pair is saved in the image processing unit as a first and second raw image. To count the sparking value the necessary operations are performed to the first image and the second image is used as a reference image. In the next image pair there are new raw images that are used to count the next sparking value. The forming of the image pairs is done so that the sparking values obtained from different image pairs are comparable with each other and with the given threshold value of the sparking value, exceeding of which will trigger an alarm or a contact to the system monitoring the object.

In an embodiment of the method according to the invention the raw images are converted to gray scale before any other processing. This diminishes substantially the size of the image files and facilitates their comparison and other processing. Furthermore as each raw image is formed pixels corresponding to a preset mask can be removed. With masking immobile parts or regions of the object are chosen in which substantially no sparking occurs. Thus the size of the images is diminished and their processing accelerated. In an embodiment of the method according to the invention removing pixels means that their value is set to some constant. A pixel can also be removed by removing the bits indicating its value from the im- age file so that the size of this file correspondingly is diminished.

In the method according to the invention the parameters relating to the imaging can be adjusted, the parameters being the properties and settings of the imaging device and the location of the imaging device in relation to the object. A sparking value is counted after the adjustment only after there are two raw images captured with the same settings.

Said monitoring data can be the sparking values as such. The monitoring data can be formed also by calculating mathematically from consecutive sparking values a balanced sparking value. The range of fluctuation of the sparking value can also be divided in a few ranges, and the monitoring data is the number of the range to which the sparking value falls on. The monitoring data can also be an alarm as the sparking value exceeds the set upper limit. In forming monitoring data also combinations of the means mentioned above can be used.

The arrangement according to the invention for monitoring sparking occurring in an object, which arrangement comprises a camera for imaging the object and an image processing unit, is characterised in that the raw images produced by the camera are digital consisting of pixels, and that there is a memory in the image processing unit for storing the values of the raw image pixels, a comparator of the pixels values, an image reducer for removing the pixels not fulfilling the set comparison conditions, counter for determining the sparking value, that is for counting the number of the remaining pixels and a sparking value processing unit for forming monitoring data from the sparking values.

In an embodiment of the arrangement according to the invention the image reducer of the image processing unit is arranged to remove pixels corresponding to the preset mask when the raw image is formed, which mask covers at least a part of the permanent structures and areas of the object, in which no significant sparking occurs. This is done before other image processing. In an embodiment of the arrangement according to the invention there is further in the image processing unit an arrangement to change the image to gray scale before any other processing of the image. In another embodiment of the arrangement according to the invention the comparator of the images' values compares the corresponding pixels of the image pair. The image reducer is arranged to remove those pixels from the first raw image, that have substantially the same values as in the second raw image and those pixels that have a value lower than the preset threshold value. The counter counts from the first image the sparking value from the remaining pixels of the first image. In the image processing unit there is a communication device for communicating with the system monitoring the object, and the communication device is arranged to send the monitoring data. The monitoring data is in such a form that it can be understood by the monitoring system monitoring the object or it can be sent to the system monitoring the object through the communication device of the arrangement. If for example the connection to the system monitoring the object is arranged via a computer network, the communication device in question places the monitoring data to a message according to the communications protocol used. If on the other hand the connection is arranged for example via a mobile telephone network the monitoring data is placed for example in a text message. The arrangement according to the invention may have been arranged to process the sparking values mathematically. Thus the error interpretations possibly occurring in the calculation of the sparking values or alarms caused by individual sparking events can be decreased. The sparking values obtained are thus for example averaged from some certain period.

The comparison conditions of the image processing unit can be adjustable according to the lighting or the function of the object. The image processing unit can perform this automatically according to the instructions in the image processing unit or the new comparison conditions are sent to the image processing unit by hand.

Image processing unit and the communication device for communicating with the system monitoring the object can be implemented with a computer and the camera can be a surveillance camera.

The image processing unit can be substantially implemented with the resources of the system monitoring the object. If for example the system monitoring the object comprises a computer, the image processing unit can be implemented with its resources. In one embodiment the processing unit is implemented with one or several signal processors.

An advantage of the invention is that with it several kinds of objects can be monitored and it can be used in different locations. The objects can be different in size and the location can be indoors or outdoors.

Further advantage of the invention is that it produces results fast. Also a further advantage is that no powerful equipment is needed for using it. An advantage is also that the invention does not need to be very close to the object it monitors. An advantage of the invention is also that the object monitored or the monitoring di- rection of the object can be changed easily.

Further the invention facilitates an arrangement to measure sparking in locations where traditional fire or spark detectors measuring for example infrared radiation or ultraviolet light do not function reliably, such as in smelting plants or foundries, in which the hot metal processed causes powerful background radiation in these wavelength ranges.

An advantage of the invention is also that it enables in some cases usage of solutions already existing as a part of the sparking monitoring. Next the invention will be described in detail. In the description reference is made to the accompanying drawings, in which

Figure 1 presents as a flow chart an example of the method according to the invention for monitoring sparking occurring in the object,

Figure 2 presents an example of an arrangement according to the invention for monitoring sparking and

Figure 3 presents an example of the usage of the arrangement according to the invention.

In Fig. 1 there is as a flow chart an example of the method according to the inven- tion. The method has some object in which sparking may occur. In the method the object is imaged with a camera and this imaging produces images that can be saved and processed. The images either are or can be converted into digital form.

Advantageously the imaging produces images continuously. The sparking value representing the intensity of the sparking is calculated in the image processing unit.

In the method an image pair is formed, in which there is at least two raw images. A raw image is an image selected from the images produced by the camera. Advantageously the selection is made by the image processing unit. The camera can also be instructed to send raw images to the image processing unit on a certain interval. No operations have yet been done to the raw image with the help of which the sparking of the object can be estimated from the image. Some operations that speed up the evaluation may have been done to the raw image before the evaluation has been started. These can be for example converting the image to gray scale or masking the image, in which regions or parts of the object, in which no sparking is supposed to occur, are removed from the image. The first raw image of the image pair is that image, for capturing moment of which the sparking value is calculated. The second raw image is used as a point of comparison in calculation.

In phase 101 the method for monitoring sparking is started. In phase 102 of the chart an image pair is formed consisting of two raw images. The raw image is in digital form. Then the raw image consists of pixels, the value of which is proportional to the brightness of the object in the area in question. The raw image is advantageously in gray scale, therefore containing only brightness information of the object. The image in grey scale contains less data than an image containing colour information, and thus its processing is faster. In phase 103 in this example immobile parts and regions, where no significant sparking is expected to occur, of the object are removed from the raw image. This procedure is called masking. These parts can be different structures belonging to the object or its surroundings. This removal is done advantageously by setting the pixels connected to the mentioned immobile parts or regions to some constant value. These areas to be removed from the raw image can be preset when sparking monitoring is arranged ready for use. In this stage also other operations facilitating the further processing of the image can be taken, such as for example converting the image to grey scale. If needed copies of the images can be saved to the memory of the image processing unit. This can be done if for example one wants to study how the object has behaved in certain moments. Phase 103 can be done also before the image pair is formed.

In phase 104 the first and the second raw image of the image pair are compared. At that time those pixels that have substantially the same value as the correspond- ing pixels in the second raw image are removed from the first raw image. The immobile parts of the object are thus substantially removed from the image, leaving those pixels that have a changed value. In phase 104 also those pixels, that have a value that does not exceed the preset threshold value set for the arrangement are removed, that is those pixels, that correspond to the areas that are not bright enough to be sparks. Thus only those pixels, that have a value that has changed between the two raw images of the image pair and the value of which exceeds the given threshold, remain. These pixels represent the sparking of the object.

In phase 105 the sparking value is calculated from the image. This is done by counting those pixels of the image that have not been removed in phases 103 and 104. The number of the pixels gives the sparking value, which is proportionate to the sparking occurring in the object at the capturing moment of the first raw image. The removal of pixels from the image can be done in several ways. The removed pixel can be for example set to some preset value or they can be removed from the bitmap, whereby this map is condensed.

In phase 106 a monitoring data is formed from the sparking value. This is forwarded to the system monitoring the object, which system is for example in the control room. Also an alarm can be made when the sparking value exceeds some set threshold value. The monitoring data can be a calculated sparking value or it can be processed to a form that is easier to handle. If the spark measurement of the object is to be continued, in phase 107 one returns to phase 102, in which forming of a new image pair is started. The images of the image pair can be entirely new or the first raw image, from which no pixels have been removed to calculate the sparking value, of the former image pair is used as the second raw image of the new image pair. Advantageously the interval between the capturing moments the first raw image and the second raw image remains substantially constant. It is also possible that as the second raw image a reference image, that stays the same during several calculations of the sparking value, is used. The reference image can be changed either at certain intervals or as the lighting conditions or imaging parameters change. Substantially little sparking occurs in the reference image. If the calculation of the sparking values of the object is stopped, one moves to phase 108.

Time series can be formed of the collected sparking values, which time series can be processed mathematically, whereby the occurrence of sparking value alarms caused by individual sparking events can be reduced.

Figure 2 presents an example of an arrangement according to the invention for monitoring sparking. The arrangement 200 comprises a camera 201 , an image processing unit 203 and a communications device 207 for communicating with the system 210 monitoring the object. The system 210 monitoring the object can be an automated equipment or for example a person in the control room observing the readings of the gages showing the state and function of the objects.

The camera 201 is placed so that at least a part of the sparking of the object is in its field of view. The camera can produce individual images or image stream, which are advantageously digital. From the camera data can be extracted or it can send to the image processing unit 203 data, from which the image processing unit can choose raw images in order to calculate sparking values. The imaging and image forming settings of the camera can be adjusted. The sparking values to be compared are calculated from the raw images that are captured with the same camera settings. In some embodiments as the settings of the camera change also the settings of the image processing unit are changed in order to count the sparks, whereby the sparking values obtained with different camera settings are comparable to each other. Advantageously the camera functions in the visible region. In the arrangement 200 there is a connection 202 between the camera 201 and the image processing unit 203. The connection can be implemented by wire or wirelessly. Through the connection image information is transferred from the camera to the image processing unit and information of the camera adjustments and settings. The camera can also be controlled through this connection. The image processing unit can give the camera new directions for example as the functional state of the object changes, for example when the probability of sparking or its significance for the function of the object diminishes. The direction can be for example a command to take images less frequently or to change the imaging set- tings, such as exposure.

The image processing unit 203 processes the images produced by the camera, calculated the sparking value from them, processed the sparking values mathematically if needed, converts sparking values to monitoring data, forwards sparking value data or monitoring data to the system monitoring the object or it can when necessary receive new directions and threshold values and deliver directions to the camera. The image processing unit is comprised of a memory 209, a comparator 204 of the values of the pixels, an image reducer 205, calculator 206 and a sparking value processing unit 211 to give the sparking value data a form readable by the monitoring system. There is also an arrangement in the image processing unit to perform other calculations and an arrangement to arrange connections to other devices. The image processing unit 203 can perform some preliminary processes to the raw images. These can be for example removing unnecessary regions of the images by masking in those areas where no sparking occurs or it is so rare that it does not impact the sparking value significantly.

Images produced by the camera 201 and other data needed by the image processing unit are saved to the memory 209. These can be unprocessed images, raw images, sparking values or images produced when counting the sparking. First and foremost the values of the pixels from those raw images that are needed for calculating the sparking value are saved there. When the image processing unit requires a certain image, it is retrieved from the memory. Also images, the calculated sparking value of which, at the time they were captured, has exceeded a certain threshold value, can be saved to the memory. This has to be done for example if one wants to examine whereabouts in the object sparking occurs and how it behaves.

Comparator 204 of the values of pixels compares the values of corresponding pixels of two images to each other or pixels of some image to some given value. When calculating the sparking value pixels whose value has changed compared to the image to be compared to and whose brightness exceeds the given threshold value are searched from the raw image.

The image reducer 205 removes those pixels from the raw image that do not fulfil the given conditions. When calculating the sparking value it removes from the first raw image those pixels, the value of which the comparator of values of pixels has observed to stay substantially the same compared to the reference image, and those pixels whose brightness does not exceed the given threshold value. The remaining pixels represent the sparking of the object. The removal of the pixels is done for example by setting the removed pixels to a certain constant value or by removing the information corresponding to the pixels in question from the image file or bitmap.

Counter 206 calculates the amount of pixels in the processed image or those pixels that have a value that differs for some constant value. When calculating the sparking value the counter counts those pixels from the image that the image reducer 205 has not removed. The figure received by the counter is the sparking value, which is proportionate to the sparking of the object at the capturing moment of the image.

The sparking value processing unit 211 forms monitoring data from the sparking values. Thus it converts the sparking values to such a form when needed that the system 210 monitoring the object understands the sparking values or they are easier to process. The communications device 207 can send a sparking value or information of it to the system monitoring the object. Then the monitoring data is included in a file or a message, the form of which depends on how the connection 208 has been implemented. The sparking value can be adapted for example to a form more illustrative to a human being. For example the representation of the sparking value can be adapted when forming the monitoring data from the number of the pixels to some reference number that represents the amount of sparking in the object. These reference numbers can be for example 0 that means that no sparking occurs in the object, 1 that means minor sparking, 2 that means significantly increased sparking and 3 that means sparking that requires immediate action. Thus the person managing the system monitoring the object directly observes from the monitoring data which kind of sparking is in question.

Different parts of the image processing unit 203 can be implemented with one or several microcircuits meant for digital signal processing. The image processing unit can be implemented also programmatically, whereby functions of its parts are implemented with computer resources. The image processing unit can be implemented with the same computer as the system monitoring the object. The image processing unit functions in this case as an independent component. Thus the im- ages produced by the camera are sent directly to the system monitoring the object.

The image processing unit 203 is in contact with the system 210 monitoring the object with the help of communication device 207. The device is a in itself known transceiver. The connection 208 between the communications device 207 and the system monitoring the object can be fixed or wireless. Via the connection the system controlling the object can send commands to the sparking monitoring arrangement 200. These commands are for example a command to start and stop the spark counting or to set new threshold values or a request to the image processing unit to send some saved image to the system monitoring the object.

Arrangement 200 for monitoring sparking can also be implemented so that all its components are in the camera. Another possible embodiment is that there are several cameras that image either the same object from different directions or different objects, and one image processing unit that processes the images produced by the cameras and forms separate sparking values for each camera. By using several cameras for the same object very specific information can be obtained about the behaviour of the sparking.

Figure 3 presents an example of the usage of the invention. There is a camera 302 and a computer 301. The camera is arranged to image the object 304, in which sparking 303 occurs. The object can be for example some industrial proc- ess, in which molten or partially molten material is handled. Such is for example a water-cooled mould in a continuous casting device, in which molten metal, for example copper, is casted through a mould and continuous copper bar is obtained. In a malfunction situation sparking may occur, in which the sparks are molten copper. The sparking should be detected as soon as possible, for if it continues it can damage the equipment nearby, such as the rolls forming the copper bar. The sparking generally occurs at the lower end of the water-cooled mould, from where the solid crusted metal bar exits. If a hole is punctured to the hardened shell, the still molten material inside it gets out and is exposed to air and water, whereby sparking is created. If the material continues to seep out the hole can grow and the molten insides of the bar flows out through the growing hole on the rolls and equipment below. This can cause great financial losses. The traditional infrared or ultraviolet radiation measuring spark detectors do not work in this environment, for the equipment and the hot metal cause background radiation through which detecting sparking is very difficult.

Camera 302 produces images of the object 304, which images are transferred to the computer 301. In the computer there is an image processing unit and means for communicating with the system monitoring the object. This computer, with which the image processing unit is implemented, can also be the system monitoring the object or a part of it. The image processing unit chooses from the images produced by the camera the images needed, from which image pairs are formed. Each image pair formed produces a sparking value which is proportionate to the sparking occurring in the object. The computer delivers the sparking value or information of the sparking value as it exceeds a given threshold value to the system monitoring the object. The system monitoring the object performs actions, which can be for example an alarm to the controller or changes to the function of the ob- ject.

In the foregoing some advantageous embodiments of the invention are described. The invention is not limited to the solutions just described, but the inventive idea can be applied in numerous ways within the limits set by the claims.

Claims

1. Method for monitoring sparking occurring in an object, in which method the object is imaged and the intensity and quantity of the sparks in the image is determined, characterised in that - an image pair (102) is formed, in which there is a first and second raw image both in their own files, and in these raw images the value of each dot, or pixel is a binary digit, that gives the brightness in a corresponding area of the object

- those pixels that have substantially the same value as the corresponding pixels in the second raw image are removed (104) from the first raw image

- furthermore those pixels that have a value lower than the given threshold value are removed (104) from the remaining part of the first raw image

- the number of pixels or the sparking value is counted (105) from the re- duced image obtained in the foregoing

- monitoring data is formed (106) from the sparking values to be delivered to the system monitoring the object.

2. Method according to Claim 1 , characterised in that an image pair is formed (102) from two consecutive raw images.

3. Method according to Claim 1 , characterised in that said image pair is formed (102) so that the latest raw image is the first raw image and the former first raw image is the second raw image.

4. Method according to Claim 1 , characterised in that said image pair is formed (102) so that the latest raw image is the first raw image and some former raw im- age is the second raw image.

5. Method according to any of Claims 1-3, characterised in that raw images are chosen so that the interval between the capturing moments of the raw images of the image pair is constant and the interval between image pairs is constant.

6. Method according to Claim 1 , characterised in that furthermore when each raw image is formed pixels corresponding to a preset mask are removed (103) from the image.

7. Method according to Claim 1 , characterised in that removing a pixel means that its value is set to some constant.

8. Method according to Claim 1 , characterised in that removing a pixel means removing the bits indicating its value from the image file so that the size of this file diminishes correspondingly.

9. Method according to Claim 1 , characterised in that the said monitoring data is the sparking values as such.

10. Method according to Claim 1 , characterised in that a balanced sparking value is calculated mathematically from the consecutive sparking values to form monitoring data.

11. Method according to Claim 1 , characterised in that the range of fluctuation of the sparking value is divided into a few ranges and the monitoring data is the number of the range that the sparking value falls on.

12. Method according to Claim 1 , characterised in that the said monitoring data is an alarm as the sparking value exceeds the given upper limit.

13. Arrangement (200) for monitoring sparking occurring in the object, which arrangement comprises a camera (201 ; 302) for imaging the object and a image processing unit (203), characterised in that raw images produced by the camera are digital being comprised of pixels, and that in the image processing unit there is - a memory (209) for storing the values of the pixels in the raw images

- a comparator (204) of values of the pixels

- an image reducer (205) for removing the pixels not fulfilling the given comparison conditions

- a counter (206) for determining the sparking value, or in other words for counting the number of remaining pixels and

- a sparking value processing unit (211) for forming monitoring data from the sparking values.

14. Arrangement (200) according to Claim 13, characterised in that the image reducer (205) of the image processing unit (203) is arranged to remove pixels cor- responding to the preset mask when the raw image is formed, which mask covers at least a part of the permanent structures and areas of the object, in which no significant sparking occurs.

15. Arrangement (200) according to Claim 13, characterised in that the comparator (204) of values of pixels is arranged to compare the corresponding pixels of raw images of an image pair and the image reducer (205) is arranged to remove those pixels that have values substantially the same as in the second raw image from the first raw image, and those pixels that have a value lower than the given threshold value, and the counter (206) is arranged to count the sparking value from the remaining pixels of the first image.

16. Arrangement (200) according to Claim 13, characterised in that there is an arrangement in the image processing unit (203) for converting the image to grey scale before further image processing.

17. Arrangement (200) according to Claim 13, characterised in that there is a communication device (207) in the image processing unit for communicating with the system (210) monitoring the object, and the communication device is arranged to send the monitoring data.

18. Arrangement (200) according to Claim 13, characterised in that the image processing unit (203) is arranged to adjust the comparison conditions or receive new comparison conditions according to the lighting or function of the object (304).

19. Method according to any of Claims 13-18, characterised in that the image processing unit (203) and device (207) for communicating with the system (210) monitoring the object is implemented with a computer (301) and the camera (302) is a surveillance camera.

20. Method according to any of Claims 13-18, characterised in that the image processing unit (203) is arranged to be implemented with the resources of the system (210) monitoring the object.

21. Method according to any of the Claims 13-18, characterised in that the image processing unit (203) is implemented with signal processors.