INVENTORY CONTROL SYSTEM

The present invention relates to an inventory control system and in particular, but not exclusively, to an inventory control system for monitoring the use of hand tools, surgical implements or other critical items. The invention also relates to a storage container for use with the inventory control system, and an inventory control process.

An inventory control system may be useful when it is important to monitor the usage of hand tools, and to ensure that they are returned to storage after use. This can help to ensure that tools are not lost or stolen. Such a system is particularly important when tools are used for repairing or maintaining aircraft engines, as any tools left inside the engine after completion of the job could cause catastrophic damage. Similarly, in the case of surgical implements, it is essential to ensure that no implements are left inside a patient after an operation.

An inventory control system can also be useful by helping to ensure that only the correct tools are used for a particular task. Such a system may also be helpful to track the identity of the person using the tools, so that if a tool goes missing or if a tool is taken that is inappropriate for the task in hand, appropriate remedial action can be taken.

One process of monitoring the use of tools is to store the tools in a container having designated storage compartments for specific tools. For example, the container may be a tool box having drawers with rigid foam liners, with cut-outs matching the size and shape of the tools. Using such a container, it is a relatively simple matter for a supervisor to check that all the tools are present by looking at each drawer in turn, to ensure that all the spaces are filled. However, this is time-consuming and laborious, and it does not allow use of the tools to be monitored on a continuous or real-time basis. In addition, the correct choice of tools for a specific task cannot easily be monitored, and the identity of the user is not automatically recorded.

US 5608193 describes an alternative tool inventory control system, in which a tool box containing tools is weighed before and after use. Discrepancies between the two weights are noted and optionally a database of tool weights can be interrogated to suggest which tool may have been left behind. The system does not however allow the presence of each individual tool to be verified or the choice of tools taken for a particular task to be monitored.

A known inventory control process is disclosed in WO2005/028165. That system uses a series of sensors located in recesses formed in a foam liner to detect the presence tools in each of the recesses. Each recess is individually shaped to receive a specific tool, the shape of the recess being matched to the shape of the tool. Each recess includes a sensor to detect whether or not the tool is present. However, there are several problems with this particular type of technology. Firstly, since there are multiple sensors (at least one for each tool) there are many wires connecting each of the sensors to a processing device. This takes up a lot of space and requires jointed conduits containing the wires to connect them to the processing unit.

Another drawback to this type of system is that the sensor is not arranged to recognise whether the tool located in the recess is the correct tool. For example, a piece of metal may be located in the tool box in place of the tool. The optical sensor will detect its presence and believe that the tool has been replaced. This makes it vulnerable to theft of tools. A further problem is that with sensors being located in each of the tool recesses, the system is essentially bespoke for each new tool layout. Thus when the user of the system wants to introduce a new tool, not only does the user have to redesign the liner to create a suitable recess for the new tool, it is also necessary to connect a new sensor to detect the presence of the tool and to adapt the processing system to accept signals from the new sensor.

Accordingly, the present invention seeks to mitigate at least one of the above-mentioned problems.

According to one aspect of the invention there is provided an inventory control system for monitoring the removal and replacement of inventory items, the system including a camera system and a data processing device arranged to receive signals from the camera system, said data processing device including image recognition means for analysing the signals received from the camera system to determine the presence and/or absence of the inventory items.

There are many benefits associated with the invention. For example, for most inventory control systems it will only be necessary to use a single camera, or a small number of cameras, rather than a multiplicity of sensors as used in prior art systems (see WO2005/028165). This makes installation and maintenance significantly easier. This is particularly advantageous since it enables existing inventory item cabinets to be easily converted to a recognition system without having to make substantial modifications to the cabinets. The camera system can even be located remotely from the inventory items provided that the inventory items are brought into the field of view of the camera system at least some of the time.

The image recognition means is able to identify whether the correct inventory item has been replaced, which is a marked improvement over the prior art system described in WO2005/028165 which cannot discriminate between tools. The data processing means is arranged to issue an alert, when it is determined that an inventory item other than the inventory item associated with the recess is located therein, wherein the alert may include images. It is to be noted that the term image recognition includes, for example, pattern matching, pattern recognition, computer vision and machine vision techniques.

Advantageously the inventory control system includes a container for storing a set of inventory items and the image recognition means is arranged to determine the presence and/or absence of inventory items associated with the container. Advantageously the container can include a plurality of inventory item storage locations each for receiving a specific inventory item. The image recognition means is arranged to detect the presence and/or absence of the inventory items in the inventory item storage locations. The image recognition means is arranged to compare the signals received from the camera system

with stored image related data for at least one of the inventory items being detected. Thus the image recognition means is arranged to determine whether the correct inventory item is located in the correct inventory item storage location. For example, if an incorrect inventory item has been located in a storage location, the image recognition means may recognise the inventory item and determine that it is located in the wrong location or the inventory item may not be recognised as part of the fixed population of tools, which may indicate that the inventory item has been placed in the wrong cabinet. The data processing means can issue alerts, for example to a visual display unit that include images of the missing tool and / or the tool that is located in its place. The combination of individual storage locations for each inventory item, the camera system and image recognition means considerably simplifies image recognition since the orientation of each inventory item is constrained by the storage location, which enables fast and reliable inventory item identification.

Advantageously the image recognition means can be arranged to compare signals received from the camera system with part of an image stored in a database. For example, this enables inventory item identification, in at least some cases, when only part of the inventory items are in the field of view of the camera system such as when a container drawer is not fully open. This can be achieved using measurement data from the received image to determine the extent of opening of the drawer and comparing that part of the image with a corresponding part of an existing image stored in a database.

Preferably each inventory item storage location comprises an individually-shaped recess for receiving a specific inventory item, the shape of the recess being matched to the shape of the inventory item. This ensures that inventory items are always returned to the correct storage locations. This allows a visual check of the inventory items to be completed very easily, by looking for any empty recesses. This provides a manual backup to the automatic system, allowing the full complement of inventory items to be confirmed easily, even in the event of a power failure or other fault. Alternatively, each inventory item storage location may include at least one of a clip, a pin, a strap or similar for locating the inventory item, to achieve a similar effect.

Advantageously the area surrounding the recesses can contrast with the recesses. For example, one of the recesses and the surrounding material may have a bright colour and the other a dark colour. The materials may be dissimilar to produce different reflective qualities. Additionally, or alternatively, the recesses and the surrounding material may be textured differently. For example, one of the recesses and the surrounding material may have a smooth surface and the other a coarse or rippled surface or may include formations that assist the image recognition means to more clearly distinguish between items in the field of view of the camera system.

Advantageously the image recognition means can include a learning routine. The routine can be run when a new inventory item is added to the container to enable the new inventory item to be detected. It can also learn what the image looks like without the inventory item being located therein.

The container may include a plurality of storage compartments, and preferably a plurality of drawers. Advantageously the inventory control system may include means for detecting which of the storage compartments is in the field of view of the camera system. Advantageously the image recognition means can be arranged to do this. For example, the image recognition means can achieve this by comparing measurement data relating to the compartment in the field of view of the camera system with measurement data from image related data stored in a database. In the case of a cabinet having a plurality of drawers, drawers of the same actual size may appear smaller when further away from the camera, the image recognition means can use the leading edge of the drawer as a first datum and one of the sides of the drawer as a second datum to determine the size of the drawer in the image. This can be compared with measurement data taken from image related data for each of the drawers to determine a match. Additionally, or alternatively, each compartment and/or each inventory item can include an identifier, such as indicia, that can be read by the image recognition means. For example, the image recognition means can include character recognition software.

The inventory control system can also include means for reading inventory item identifiers for identifying each item. For example, character recognition software can be

used for reading indicia on the inventory items, such as unique identifying codes or item parameter information, such as a tool size, to assist the recognition process.

In some embodiments the inventory control system may include a system for communicating to the data processing device which of the compartments' inventory items are in the field of view of the camera system. For example, if the container includes a number of drawers, the system is arranged to indicate to the data processing device which of the drawers is open. This assists the image recognition process since it reduces the number of tools that have to be compared with the signals received from the camera system, which speeds up detection.

Advantageously in some embodiments the container can include means that enable the positions of at least some of the inventory items to be adjusted, a system for detecting the extent of movement and a communication system for communicating the extent of movement to the data processing device. This enables the data processing device to determine whether the inventory items are fully within the field of view of the camera system. For example, the system can determine whether a container drawer is fully or partially open. This enables the data processing means to speed up the image recognition process on the part of the drawer that is open, since it is only necessary to compare the signals received from the camera system with data relating to a part of a stored image of a fully opened drawer. While full recognition may not be possible for some inventory items, it may be for others, and it is possible to determine for at least some inventory items whether the inventory item is missing entirely or whether an incorrect inventory item is located in a recess.

A problem for containers including a plurality of storage compartments, such as drawers, is the positioning of the camera system and the ability to obtain clear images of all of the inventory items in each of the drawers. Advantageously the camera system can include an automatic focusing system to enable it to adjust its focus according to which of the storage compartments is in the field of view of the camera system. Additionally, or alternatively, the camera system can be arranged to have a long field of view and/or

include image processing means for adjusting out of focus images to improve their sharpness for image recognition purposes.

Advantageously the camera system can be mounted in or on the container. This ensures that the camera has a fixed vertical position relative to the inventory items and can provide a very compact arrangement. Thus the quality of the images produced by the camera is not determined by the position of the camera system relative to the container. For example, if the camera system is not mounted on the container, and the container is moved, the camera system may have difficulty focusing on the inventory items or the inventory items may be partially obscured from the field of view of the camera system. This makes image recognition more difficult. If the container includes at least one drawer the camera system can be arranged to overhang the container to view the inventory items located in the or each drawer. This provides the camera system with an unobstructed view of the inventory items stored in a drawer, when opened. The relative movement between the camera system and the inventory items enables the camera system to view all of the inventory items located in the drawer as it is opened. Advantageously the camera system can be mounted on an elongate stand or bracket attached to the container to provide sufficient distance between the camera system an the closest drawers to enable all of the inventory items to fall within the field of view of the camera system.

When the camera system is mounted on or very close to the container a particular problem is ensuring that all of the inventory items stored in the drawer closest to the camera system are within the field of view of the camera system.

Advantageously the camera system may include a plurality of cameras. Having at least two cameras increases the width of view of the camera system thereby ensuring that all of the desired inventory items fall within the field of view of the camera system. For example, the cameras can be arranged either side of a container to view across each drawer. It also provides the possibility of producing stereoscopic images.

Advantageously the camera system may include at least one camera and an arrangement of optical devices for increasing the field of view of the camera. Thus when the camera system is mounted very close to the inventory items its width of view can be extended by a system including reflective devices, such as mirrors, to ensure all inventory items are within the field of view of the camera system.

Advantageously the camera system can be moveably mounted and is arranged to scan the inventory items. For example, the camera system can mounted in a plane that is substantially parallel to the plane of the inventory items and is arranged to move in at least one axial direction. This enables the camera system to extend its viewable range by scanning the items in the inventory item plane in one or two dimensions, which can be useful when the camera system is mounted close to the plane of the inventory items.

Advantageously the inventory control system can include means for determining and recording the identity of a user. For example, the container may include a locking mechanism for controlling access to the inventory items. The locking mechanism may include an input device for entering the identity of a user, and the data processing device may include means for recording the identity of a user. This makes it possible to monitor the inventory item usage of individual workers and ensure best working practices. Also, if an inventory item goes missing, it is possible to identify who last had possession of that inventory item.

Advantageously the image recognition means can include means for recognising and/or recording the identity of persons in the field of view of the camera system. This enables the inventory control system to control access to the inventory items according to a comparison of image data received from the camera system with image related data relating to authorised users, which is stored in a database. Advantageously the locking mechanism for controlling access to the inventory items can be arranged to automatically open when the image recognition means identifies an authorised user. Advantageously the locking mechanism can be arranged to automatically lock when the authorised user leaves the field of view of the camera system. Advantageously the locking mechanism can be arranged to automatically lock after a predetermined period

of time has lapsed after the authorised user leaves the field of view of the camera system. The locking mechanism may also include a manual switch for locking the mechanism.

Advantageously the means for recognising and/or recording the identity of persons in the field of view of the camera system can be arranged to record the identity of persons in a format that allows comparison with signals received from the camera system. This enables, for example, the inventory control system to determine whether a person in the field of view of the camera system has previously tried to access the inventory items, for example in the situation where the person has tried to access the container but has been denied access. Advantageously the data processing device can record data such as the time and date that the person was in the field of view of the camera system. If the same person enters the field of view of the camera system at a later time, the image recognition means is able to identify that person by comparison with database of images and record new time and date data in that persons record. Advantageously the data processing means can be arranged to record additional data on that person's record, for example any other available data from other databases, swipe cards, chip and pin cards, identity tags, and RFID tags. Advantageously the database for persons can be made available to other camera systems, for example for an other inventory control system at a different location, such as at a different warehouse, factory, maintenance area or shop, such that similar image recognition means, having identified a person in the database, can add information to the database relating to that particular transaction.

Advantageously the inventory control system may include camera system protection means, wherein the protection means is arranged to protect at least part of the camera system and to automatically expose the protected part(s) when a user successfully logs into the system. This prevents the camera system being tampered with by unidentified persons. For example, the camera system may include a closable lens shutter or a casing, which is automatically openable when an authorised user logs in. The lens shutter or casing is preferably automatically closeable when the user logs out. The lens shutter and casing can be electromechanically driven, for example by an electronically controlled motor.

Advantageously the data processing device is arranged to record the removal and replacement of inventory items according to the received signals. This makes it easy to assess whether all the inventory items taken from the container have been returned after use, thus reducing the risk of any inventory items being inadvertently left behind. Safety in critical situations such as aircraft engine maintenance is thus significantly improved. The invention also reduces the risk of inventory items being lost or stolen, since the fact that an inventory item has gone missing can be indicated immediately. The system also makes it possible to monitor which inventory items are being taken for any specified task, thereby helping to ensure best practice in maintenance operations.

Advantageously, the data processing device includes means for recording the time of removal and replacement of inventory items. The data processing device is preferably a computer. The data processing device may be located remotely from the container and may be connected to the container by a data link, for example a wired link, an optical link (e.g. using infrared light) or a radio link. The data processing device may also be positioned locally, for use in the workplace. Alternatively, the container may include a local indicator device for indicating the presence and/or absence of inventory items in the container. The computer can be connected to a network so that alerts can be sent to and instructions received from another terminal. The inventory control system may include a plurality of cabinets and the data processing device can be arranged to receive signals from each of the containers.

According to another aspect of the invention there is provided an inventory item container for use in an inventory control system, said container having, a camera system, means for sending signals from the camera system to a data processing device, said data processing device including image recognition means for analysing signals received from the camera system for determining the presence and/or absence of the inventory items associated with the container.

Advantageously, the means for sending signals from the camera system to the data processing device is constructed and arranged to transmit signals to a remote data processing device.

Advantageously the container may include any feature or any combination of features of the inventory control system described above.

According to another aspect of the invention there is provided an inventory control process for monitoring the removal and replacement of inventory items, including using a camera system to view the inventory items, using a data processing device to receive signals form the camera system and analysing said signals with image recognition means for determining the presence and/or absence of the inventory items.

Advantageously the inventory items are tools. The term 'tool' is to be understood to include surgical instruments and other similar objects. For example, the process can be used to monitor the use of surgical instruments in an operating theatre. Preferably the tools are manually operable tools.

The process may include storing the inventory items in a container. Preferably the container includes a plurality of inventory item storage locations each for receiving a specific inventory item. Each inventory item storage location may comprise an individually-shaped recess for receiving a specific inventory item, the shape of the recess being matched to the shape of the inventory item.

The process may include recording the removal and replacement of inventory items according to the received signals.

The process may include recording the time of removal and replacement of inventory items.

The process may include detecting and recording the identity of a user of the inventory items. The process may include checking the identity of a prospective user against a list of authorised users and allowing access to the inventory items only if the prospective user is authorised. Advantageously the camera system can be protected by protection means that is arranged to automatically expose at least part of the camera system when a user successfully logs into the system.

The process may include using any of the features of the inventory control system and/or the container described above.

An embodiment of the invention will now be described, by way of example only, with reference to the accompanying drawings, in which:

Figure Ia is a perspective view of a tool cabinet having an integral camera with an open tool drawer;

Figure Ib is a side view of the tool cabinet of Figure Ia with open tool drawers;

Figure Ic is a perspective view of a combination of tool cabinet and a separately mounted camera;

Figure 2 is a perspective view from above of the drawer shown in Figure Ic;

Figure 3 is a perspective view of the tool cabinet of Figure Ic from the rear;

Figure 4 is a plan view showing the layout of another tool drawer;

Figure 5 is a perspective view of a tool cabinet control panel;

Figure 6 is a schematic diagram illustrating the main components of an inventory control system;

Figure 7 is a flow diagram showing the main steps of an inventory control process;

Figure 8 is a diagrammatic view of a static camera and an arrangement of mirrors for viewing tools stored in the cabinet;

Figure 9 is a diagrammatic view of a moveable camera and an arrangement of mirrors for viewing tools stored in the cabinet; and

Figure 10 shows a cabinet with two cameras mounted thereon.

The tool cabinet 2 shown in Figures 1-3 includes a steel cabinet housing 4, which may be static or mobile (for example, it may be mounted on castors). The cabinet includes four tool drawers 6a-6d, the top drawer 6a being shown pulled out. Each drawer includes a liner 8, for example of a rigid foam material, having a number of cut-out compartments (or recesses) 1 Oa-I Of, which are shaped to receive specific tools. The recesses are preferably colour-coded, to provide a simple visual indication that a tool has been removed. For example, the recesses may have a bright colour (e.g. yellow), while the top surface of the drawer liner has a contrasting colour (e.g. black).

The cabinet 2 includes a locking system (shown illustratively as item 62 in Figure Ib), which ensures that only one drawer 6a-6d can be opened at a time.

A camera 12 is arranged to view tools in the recesses 10a- 1Of when each of the drawers 6a-6d are opened. Preferably the camera 12 is mounted on the cabinet 2 (see Figure Ia- b) however it is not strictly necessary to do to this. As an alternative, the camera 12 can be mounted on a separate support, such as a wall, roof beam or stand, adjacent the cabinet 2 (see Figure Ic).

The camera 12 is connected to a computer 24 either directly or via a control unit 16. An input/output unit 14 is mounted on top of the cabinet 2 (see Figures Ia and 5) or is located in each drawer 6a-d (see Figures Ic and 2), which may optionally include a visual display to show which tools are present. The input/output unit 14 is in turn connected to the control unit 16 mounted on the rear of the cabinet. The control unit 16 is connected to the computer 24 via an output cable 18 or alternatively by a wireless connector 19, for example via an infrared, radio or GSM link, and can send data to and receive data from the computer 24.

The computer 24 may be kept locally or at a remote location and includes image recognition software for interpreting the signals received from the camera 12. Thus when one of the drawers 6a-d is in the field of view of the camera 12 the computer 24 is

able to determine which of the tools is present and which is absent. There are a number of image recognition systems that are suitable for implementing the current invention. Preferably the recognition software used is able to detect the presence of tools by identifying the edges of objects in the field of view from a still image derived from the signals received, and compares the outlines with data relating to each of the tools stored in the drawers 6a-6d, for example compares the identified profiles from the received image with a database of previously identified profiles relating to existing images showing the tools in place. If a match is found for a particular tool, the software determines that the tool is present. If a match is not found for a particular tool, the software determines that the tool is not present and sends an alert identifying the tool, the cabinet, the drawer, the user who took the tool from the cabinet 2, the date and time the tool was removed, and if necessary an indication that an alternative tool/object is present. The alert may include a visual representation of the missing tool(s), for example an image of the drawer with the tools in place, with the missing tool(s) visually highlighted by a boarder, or different colour, etc.

Since the system uses image recognition software it is able to determine whether an incorrect tool has been placed into the container. This is particularly advantageous in situations where, for example, a user sets out to steal a tool and tries to fool the system by placing another, dissimilar, tool into the recess, or the user accidentally replaces an incorrect tool into the recess. The image recognition software detects the presence of the object in the recess but when compared with data relating to the tool that should be located in the recess a match is not found. Accordingly the software concludes that an incorrect tool is present and issues an alert.

The computer 24 also includes database software for maintaining an inventory of the tools in the cabinet and a log of tool usage.

The tool cabinet 2 also includes an electrically operated lock for the drawers, which can be actuated using a key, a personal identification number (PIN) via an input pad 15, or a swipe card system 17 containing data identifying the user. This allows the identity of the user to be monitored. This data can be stored in database of tool usage. The input

pad 15 can be used to input other data, such as a part number or to select options on a menu system.

Figure 4 shows an alternative drawer layout, with a number of tool compartments 10 as well as an input/output unit 14 with an LED display located in the drawer. Each of the compartments also includes a finger recess 20, allowing the tool to be easily removed from or replaced in the appropriate recess 10.

The main components of an inventory control system are shown schematically in Figure 6. The system includes the tool cabinet 2 and a computer 24, which is connected to the cabinet 2 via the link 18,19. The cabinet 2 includes the camera 12, the input/output unit 14 and control unit 16. It also includes a power supply 26, an access control system 17(for example a swipe card reader, a PIN reader or a key lock) and a solenoid 30, for operating electrical locks on the tool drawers. The computer 24 includes a central processing unit 32, a power supply 34 and a visual display unit 36. If the camera 12 is mounted separately from the cabinet 2 as in Figure 1 c, it will typically have a separate power supply from the cabinet 2 and is connected directly to the computer 24.

An inventory control process carried out using the system described above will now be described with reference to Figure 7, which shows the steps of a typical control process. First, a user (for example a technician or mechanic) identifies him or herself 40 by entering a PIN or using a swipe card. The identity of the user is checked against a list of authorised users held on the computer 24 and, if the authorisation of the user is valid, the locks of the tool box are unlocked, allowing the user to gain access to the tools. At the same time, the identity of the user and the time are recorded 42 in a database on the computer. If the identity of the user is not validated as that of an authorised user, the tool box remains locked, preventing access to the tools. Optionally, a warning may be sent to the computer 24, to indicate that an unauthorised person has attempted to gain access.

Assuming that the identity of the user has been validated, the user then opens one of the drawers 6a-d. As this takes place, the tools in the drawer 6a-6d are pulled into the field

of view of the camera 12. The camera 12 sends signals to the computer 24 and the image recognition software determines which of the tools are present in the drawer 6a- d.

The user then removes the required tools from the cabinet 44 in the field of view of the camera 12, the image recognition software automatically detects 46 which tools have been removed and this information is recorded 46, together with the time of removal and a user identifier. The user then closes the cabinet and, after a short delay, the box then re-locks automatically 50, and the registered user is signed off.

After completing the assigned task, the user re-enters his or her ID 40, and once this has been verified the box unlocks 42 and the identity of the user is registered on the computer 24. The user opens the appropriate drawer 6a-d which brings the tools into the field of view of the camera 12. The user then replaces the tools 52 in the appropriate recesses, the image recognition software determines the identity of the tools and the computer 24 records that the tools have been replaced 54, together with the time of replacement and the identity of the user. If the recognition system determines that an incorrect tool has been placed in a recess it issues an alert 56, which enables a supervisor to investigate the incident. The user then again closes the cabinet and, after a short delay, the box re-locks and the user is logged off 58.

The computer 24 therefore records which tools are in use, who has taken them and the time at which the tools are removed and returned. Using this information, it is a simple matter for a supervisor to check whether all the tools are present in the cabinet and, if any are missing, who has taken them and when. The supervisor can also check that the tools taken for a particular task are appropriate for that task. Checks can be carried out by the supervisor whenever required, or they can be instigated automatically, for example whenever the cabinet is opened or at predetermined intervals. In addition, the tool cabinet can be checked visually at regular intervals, to ensure that the full complement of tools is present and that the automatic system is operating correctly.

The computer can also keep a continuous log of how long each tool has been in use, which may be useful for tools and measuring instruments such as torque wrenches that have to be recalibrated at preset intervals. It can also be programmed to disregard the absence of tools that have been removed deliberately for repair or replacement.

Various other modifications of the invention are of course possible. For example, the cabinet 2 can include a motorised drawer opening and closing mechanism in order to control the rate at which the drawer opens and / or closes. This can assist the image recognition process.

The cabinet 2 can include an arrangement of mirrors 64, and/or other optical devices, for extending the field of view of the camera 12 (see Figures 8 and 9). Reference number 66 shows diagrammatically the field of view of the camera 12 with the mirrors 64. The mirrors can be of any suitable type and may include planar 64a, concave or convex 64b mirrors. This is particularly useful for viewing tools located in the top drawer 6a when the distance between the camera 12 and the tools is small.

The camera 12 can be mounted on a drive system 68 so that it can scan tools located in a drawer. Figure 9 shows the camera 12 mounted on the drive system and used in conjunction with an arrangement of mirrors 64. Reference number 70 shows diagrammatically the field of view of the camera 12 with the mirrors 64. It should be noted that the arrangement of mirrors 64 is not strictly necessary when the camera 12 is mounted on a drive system 68.

The cabinet 2 can include more than one camera 12 to ensure that all of the tools are in the field of view of at least one of the cameras 12. Figure 10 shows two cameras 12 mounted at the sides of the cabinet directed to look diagonally across the drawers 6a-d.

The cabinet 2 can include an openable casing or moveable plate that is arranged to block access to the or each camera 12, and in particular the or each lens. The camera 12 can be protected in a static position or can be arranged such that it moves into a recess or casing formed in the cabinet body. In the closed position, the plate or casing prevents

at least the lens from being tampered with, for example an unscrupulous person wanting to overcome the inventory control system may try to smear the camera lens or attach an opaque material thereto in order to provide himself with an opportunity to steal tools. The openable casing or moveable plate is actuated by an electromechanical drive system when a user successfully logs in. Thus the lens is not exposed when the cabinet 2 is not in use and becomes exposed only in the presence of a known user (the user's details are recorded at log in). When the user logs out, the actuator system closes the casing or moves the plate across the lens to prevent access thereto.

The cabinet 2 can include a sensing system 60 for determining which of the drawers has been opened. The sensing system can also be arranged to determine the extent to which the drawer has been opened. The sensing system can be connected to the computer 24 to enable the computer 24 to improve the image recognition processing. For example, knowing which drawer is open reduces the tool data set for comparison with the signals received from the camera 12, which can speed up the recognition process. Knowing the extent to which a drawer is open enables a map of the drawer to be produced such that the image recognition software is able to identify a part of a tool by comparison with a part image so that recognition can take place when the drawer is not fully opened, and thus at least some of the tools may not be fully within the field of view of the camera 12. Determining the extent to which a drawer is opened can be achieved by using a rack and pinion arrangement with an indexer for determining the extent of rotation of the pinion. Knowing the extent of opening can enable a map of the tools to be produced in each drawer.

The image recognition system can be an intelligent system that is arranged to learn new tooling layouts for example if the foam layers for storing the tools are modified at a later date. The system can be arranged to automatically learn new tool layouts by running a learning program on the computer.

The tool container may also take different forms: for example it may consist of a box with a single layer of tools, or with tools in removable trays, in clips or straps, or it may take the form of a cupboard or a board on or within which the tools are hung, or any

other suitable form. The tools may also of course be of any kind, including engineering tools, surgical tools and so on. The invention may also be adapted to other non-tool applications where an inventory control system is required, and references within this specification to tools should be construed accordingly to include equivalent items in suitable non-tool applications. The invention may for example be used for inventory taking or stock taking/control purposes.

The system may also include image recognition software for recognizing people. This enables the inventory control system to control access to the cabinet 2 by face recognition techniques, by comparing persons attempting to gain access to the tools with a database of images of authorized users. The locking mechanism can be arranged to automatically unlock when an authorized user is recognized, but remains locked if the user is not recognized.

The inventory control system can also be used in a shop, warehouse or library to monitor the usage of inventory items such as consumer goods and books. For embodiments that include face recognition software, a database of user images can be generated. For example, in the context of a shop, a database of images of customers can be maintained together with other data such as name, address, credit card details, customer preferences, details of customer purchases. Each time the customer enters the shop and enters the field of view of the camera, the image recognition software can compare the image received from the camera with the database to thereby identify the customer. The image recognition software is also able to identify the goods that the customer picks up from an inventory item database. Having identified the customer, details of the new transaction can be added to the customer's record. The database can be made available to a computer network such that if the customer enters a different shop that has access to the customer database, the customer can be recognised and additional transaction data can be added to the database. The system also helps to prevent thefts by sending alerts if known shop lifters enter the shop.

Similar systems can be applied in the context of libraries to monitor users of books and in warehouses to monitor the movement of goods.