SAFETY METHOD AND APPARATUS
The invention relates to a safety method and apparatus. Specifically, the invention relates to a method and apparatus for monitoring a particular zone, for example a hazardous or restricted access zone, and providing a signal when a presence is detected in such hazardous or restricted access zone.
Equipment/zone guarding to prevent the risk of worker injury or to prevent entry by an intruder is regulated according to strict legislative standards in most countries. Application areas are many and various, and include, but are not restricted to, perimeter guarding for industrial robots, machinery, hazardous fluid areas, hot areas, electrical hazard areas, radiation hazard areas and point-of-access guarding for automated machine assemblies. As safety standards in the workplace become ever more stringent, the demand for more technically advanced and more secure products increases.
Typical guards include an assortment of components and systems designed to protect industrial machinery and/or its operators from hazards such as the examples given above. It is a requirement that such guards be relatively simple in terms of activation as deployment time is a
critical factor in assuring operator and/or equipment safety.
Typical technologies used are, for example, photoelectric sensors ("light guards") and pressure-sensitive devices. Light guards typically use a photoelectric transmitter which projects a suitably configured array of synchronised, parallel light beams (a "light curtain") in the vicinity of the equipment or area to be guarded. The beams, which are typically in the infrared region, are normally detected by suitably placed detectors. When one or more of the beams is interrupted, the light-curtain controller activates an emergency condition, for example sending an emergency stop signal to the guarded equipment. Alternatively, pressure-sensitive devices may be similarly employed.
The devices are typically placed in one or more locations around the zone/machine to be monitored, an emergency condition being activated once a device is tripped.
The abovementioned devices have various drawbacks. For example, the monitoring region of a light-curtain device is typically planar ("flat-face monitoring") . Such devices are consequently inflexible, and are difficult to configure for different spatial arrangements.
Furthermore, should it be desired to monitor plural approaches to a region or piece of equipment, a number of separate light curtains must be deployed. In addition, alignment of the devices is complex and difficult. When using pressure-sensitive devices, a large number of devices must be provided in order to provide satisfactory equipment/zone monitoring. Furthermore, it might be possible, for example, to jump or fall into the danger zone without having triggered the sensor. The sensor arrangements must also be robust enough for local deployment in an often-inhospitable industrial environment .
There is therefore a requirement for an improved safety method and apparatus avoiding the above disadvantages. In particular, there is a requirement for such a method and apparatus that is simple, reliable, rapid in its operation, fail-safe, rugged (for example, weatherproof, waterproof and/or environment-proof) , portable and easy to install, deploy and use. There is an additional requirement for such a method and apparatus that is versatile, so that it can be used in guarding equipment, zones or hazards of various shapes and sizes, and in addition has the ability simultaneously to monitor different regions. In addition, there is a requirement for the regions to be capable of rapid redesignation, for
example by programming. There is a further requirement for the equipment to be capable of deployment in a number of different positions and/or orientations with respect to the zone/item to be guarded. Further requirements are the ability to discriminate between objects, for example between persons and products (muting) , and also the ability to recognise different items such as fingers, hands, limbs and bodies. The apparatus/method should also be amenable to automation, networking and centralised monitoring, and should be resistant to removal/avoidance .
It is an object of the present invention to fulfil one or more of the above requirements.
According to one aspect of the present invention there is provided a method for monitoring at least one zone, the method comprising the steps of a) capturing a current image of the at least one zone; b) analysing the at least one current image in order to derive current image data representative of the current image; c) comparing the current image data with historical image data representative of at least one corresponding previous image; and
d) generating an alarm condition dependent on the result of the comparison.
According to another aspect of the present invention there is provided an apparatus for monitoring at least one zone, the apparatus comprising: a) means for capturing a current image of the at least one zone; b) means for analysing the at least one current image in order to derive current image data representative of the current image; c) means for comparing the current image data with historical image data representative of at least one corresponding previous image; and d) means for generating an alarm condition dependent on the result of the comparison.
The image may be captured using a digital or analogue camera. The image may be a visible image, an infra-red image or any other suitable image.
Step c) may include the additional step of analysing any difference that exists between the current image data and historical image data, the alarm condition of step d) being generated according to whether the difference fulfils certain predetermined criteria.
Such difference criteria may be adaptable in real time and will depend on the particular application. The difference criteria may be selected from one or more of object size, object shape, object movement, object temperature, direction of movement, speed, colour information, approach zones around the zone to be monitored and previous history of the intruder (such as where he has come from and how long he has been there) .
The at least one zone to be monitored may contain a hazard such as for example hazardous equipment, material or fluid, temperature levels, gas, radiation or similar. Alternatively, the at least one zone may be an environmentally controlled zone, disturbance of which may jeopardise a sensitive operation, for example a manufacturing operation.
The alarm condition may include the generating of an audible and/or visible alarm signal. Alternatively, or in addition, the alarm condition may include one or more of the steps of disabling or restricting access to equipment or one or more zones, cutting off power to equipment, putting on standby of equipment or the initiation of any other safety or security procedure.
Step c) may include the sub-steps of deriving a present weighted reference image using a plurality of previous historical images of the at least one zone; forming a difference image which represents the difference between the current image and the present weighted reference image; dividing the difference image into a defined number of cells dimensioned such that each cell is more than one pixel; calculating at least one of mean and variance values of the pixel intensity of each cell; and determining whether the calculated values exceed preset corresponding trigger threshold values.
For a better understanding of the present invention and to show more clearly how it may be carried into effect reference will now be made, by way of example, to the accompanying drawings which show schematically various embodiments of the present invention. The figures are not to scale.
Figure 1 is a schematic representation of one embodiment of the present invention;
Figure 2 is a view from above of a piece of hazardous equipment to be monitored, with no obtruding presence;
Figure 3 shows a difference image between Figure 2 and a reference image;
Figure 4 is a processed image showing the calculated difference between Figure 3 and a reference image;
Figure 5 is a view from above of the equipment of Figures 2 to 4, this time with an obtruding presence;
Figure 6 shows a difference image between Figure 5 and a reference image; and
Figure 7 is a processed image showing the calculated difference between Figure 6 and a reference image.
In Figure 1, a camera 1 is placed above the area to be monitored. The zone 2 to be monitored is defined with reference to the image captured by the camera. The area to be monitored may contain, for example, a piece of hazardous equipment 3. The camera 1 captures a current image 4 of the zone. Image processing and analysis means 5 are provided, in which data representative of the current image 4 are compared with historical image data representative of at least one previous image 6. An alarm response 7 is generated depending on the result of the comparison.
The camera 1 may be optical (i.e., responsive to visible radiation) , infrared and/or responsive to one or more further wavelengths in the electromagnetic spectrum and, depending on environmental requirements, may be for example weatherproof or waterproof. The use of a camera allows large zones to be simultaneously monitored, and also allows multiple monitoring zones so that, for example, equipment with rotating, moving or otherwise hazardous parts (such as lathes, presses, conveyor belts, etc.) may be selectively monitored. The camera image is initially analysed and the zones to be monitored are designated by defining a virtual boundary around the zones. The image signal from the camera passes to processing means (such as a suitably programmed computer) , which analyses the image. Data derived from the image are compared with data derived from at least one historical image. From this comparison, the size, shape, direction of movement and/or nature of any obtruding object can be determined. Any difference in image is compared against pre-defined criteria and an alarm response generated as determined. Subject to the defining of relevant criteria, this approach allows, for example, discrimination between an insignificant object (such as an insect) and a significant object (such as a hand) obtruding into the monitored zone. The alarm response may typically be the sounding of an alarm signal, the lighting of an alarm
indicator panel, and/or the shutting down of hazardous equipment, or any other required response.
The difference criteria may be adaptable in real time and will depend on the particular application. The difference criteria may be selected from one or more of object size, object shape, object movement, direction of movement, speed, colour information, approach zones around the zone to be monitored and previous history of the intruder (such as where he has come from and how long he has been there) .
In Figure 1, an object 8 such as a hand is shown obtruding into the monitored zone 2. An image 9 of the obtruding portion appears in the current captured image 4. In a comparison and analysis step 5, data representing this current image are compared with data representing one or more historical images (e.g. 6) and, if the result of this comparison fulfils particular criteria, an alarm response 7 is generated.
A particularly advantageous, but not exclusive, method of comparing current and historical images and determining whether an alarm response is required is described in International Patent Application No. PCT/GB02/00762 "Method of Detecting A Significant Change of Scene", the contents of which are herein incorporated by reference.
In the method according to this application, a "present weighted reference image" is derived using a plurality of previous historical images of a scene. A difference image is formed which represents the difference between the current image and the present weighted reference image. The difference image is divided into a defined number of cells dimensioned such that each cell is more than one pixel, and at least one of mean and variance values of the pixel intensity of each cell are calculated. A significant change of scene is detected when the calculated values exceed preset corresponding trigger threshold values.
Figure 2 is a view from above of a piece of hazardous equipment to be monitored, with no obtruding presence.
Figure 3 is a corresponding difference image showing the monitoring zone 2. Figure 4 is a processed image showing the calculated difference between Figure 3 and a reference image. In this case, the reference image may be substantially identical to that of Figures 2 and 3. In other applications, however, the reference image may be derived from a plurality of previous historical images - this approach is particularly useful when monitoring a sudden obtrusion into a zone which is gradually changing, and is described more fully in International Patent
Application No. PCT/GB02/00762 "Method of Detecting A Significant Change of Scene", previously discussed.
Figure 5 is a view from above of the equipment of Figures 2 to 4, this time with an obtruding presence 4 in the monitored zone 2, and Figure 6 is a corresponding difference image. Figure 7 is a processed image showing the calculated difference between Figure 6 and a reference image. The obtruding presence 4 in the monitored zone 2 can clearly be seen in Figure 7, and this will trigger the required alarm response.