JP3329494B2 - Use condition setting input control device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a use condition setting input control device for setting and inputting use conditions to an operation unit provided in various apparatus and equipment.
The present invention relates to an image forming apparatus having a function of determining user replacement and performing an automatic reset based on the result.

[0002]

2. Description of the Related Art In recent years, in order to improve the convenience of users, automation of equipment has been increasingly promoted. In particular, the field focused on in the present invention is to enhance the various functions of the equipment by detecting the approach of the user to the equipment and the separation of the user from the equipment by using a sensor capable of detecting the human body. This is the field I want to make.

For example, in an automatic door that detects the approach of a person and automatically opens and closes the door, and in a cash dispenser of a bank or the like that detects an approach of a person and automatically turns on / off an operation display unit (LCD). It is widely put into practical use. In addition, in an image forming apparatus such as a copying machine, an automatic preheating function for automatically turning on / off an operation display unit (LCD) by detecting approach of a person, and various types input from the operation unit Copying conditions
An automatic reset function for automatically returning the copy mode to the initial condition is being put to practical use.

In the image forming apparatus, after a copying operation is performed under special conditions and mode settings (double-sided, combining, sorting, binding adjustment, scaling, etc.), the setting is made when the next user operates the apparatus. If the copying operation is executed as it is without carefully confirming the copying conditions, the copying is often executed under the previously set special conditions, resulting in erroneous copying. In order to prevent such a problem, an automatic reset function for automatically returning the copy conditions and the mode to the initial conditions when a predetermined condition (elapse of a predetermined time after completion of the copy operation, separation from the user's apparatus) is satisfied is satisfied. Widespread.

[0005]

As described above, the auto-reset function itself is an effective means for preventing a malfunction due to an erroneous setting input. When such a predetermined condition is not satisfied, this function does not operate, and the conventional problem has not been solved.

More specifically, the auto-reset function does not operate unless a predetermined condition (elapse of a predetermined time after completion of the copying operation, separation from the user's apparatus) is satisfied, and the use condition does not automatically return to the initial condition. If the next user handles the device before the predetermined time has elapsed from the end of the copying operation or before the previous user has separated from the device by a predetermined distance or more, the use condition is still the same as the previous copy condition, and However, there has been a problem that a malfunction occurs due to the same wrong setting input.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems of the prior art. The present invention has been made to solve the above-mentioned problems. An object of the present invention is to provide a use condition setting input control device in which a malfunction due to a setting mistake is reliably prevented.

[0008]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention discriminates and detects whether an operation unit for setting and inputting use conditions when using the apparatus and a person who uses the apparatus are replaced. Discriminating means, and control means for automatically switching a use condition set and input from the operation unit to a predetermined initial use condition when replacement of a person using the device is detected by the discriminating means. It has a door,
After the use condition of the apparatus is automatically switched to the predetermined initial use condition by the control means, the use condition at the time of immediately preceding use is stored in the storage means, and the use condition is stored in the storage means. call has been the use conditions in one operation input is obtained by said and Turkey automatically sets the apparatus to said use condition.

[0009]

According to the present invention, the switching of the user is discriminated and detected. When the switching is detected, the automatic reset function is activated regardless of the operating condition (elapse of a predetermined time after completion of the copying operation, separation from the user's apparatus by a predetermined distance or more). Activate the reset function.

FIG. 1 is a block diagram for explaining a configuration concept when an automatic response system is applied to OA equipment. The automatic response system includes a sensor unit 2 composed of one or more sensors.
And a recognition providing device 3 and a service providing device 4, which are mounted on various OA devices 1 to realize an automatic response function, and which are appropriate regardless of active work by a user of the OA devices 1. It responds and provides various services.

The sensor unit 2 is mounted on the surface or inside the OA device 1 to generate various data for recognizing an object (for example, an approaching human) to be automatically responded, and recognizes and determines the obtained data. Send to device 3. The type of the sensor, the mounting position and the number thereof are determined by the parameters to be detected from the data from the sensor, the direction to be detected (for example, the direction in which the operation panel is facing = the front of the OA device), and the size of the object to be responded to. (Width, height), detection accuracy (resolution, detection interval) and the like are appropriately determined.

FIG. 2 is a diagram showing an example of mounting a sensor unit including a plurality of distance sensors 6 on an image forming apparatus 5 which is a copying machine, a facsimile, or a laser beam printer. Overall perspective view of the device, (b)
The figure is a plan view. In the example of FIG. 2, the response target object is a human (operable person), the data obtained by measurement is the distance from the image forming apparatus 5 to the measured object, and the directions to be measured are the front and side surfaces of the apparatus 5. . In addition, in order to obtain the direction of the object to be measured, the distance is measured in a plurality of directions by a distance sensor having a relatively sharp directivity. For example, a sensor of this type irradiates infrared light from a light emitting unit in a target direction and measures the amount of reflected light at a receiving unit to measure a distance, or transmits an ultrasonic wave of a certain frequency from a transmitter. In some cases, the reflected wave is received by a receiver and the distance is measured based on the phase difference.

In the example shown in FIG. 2, a plurality of sensors are mounted in order to obtain a high resolution in a relatively short detection period, and the distance sensors are operated in parallel for measurement. In order to obtain the direction of the object to be measured, each sensor emits light little by little (10 degrees interval).
The transmission / reception / reception directions are shifted. Also, since vertical direction data (height etc.) is not required, the direction of light emission / transmission / reception / reception is developed and mounted only on a horizontal plane. As the configuration of the sensor unit, an image input device using a CCD, for example, may be considered in addition to the distance sensor. In this case, the image data captured from the image input device is sent to the recognition determination device.

The recognition judging device 3 is built in or external to the OA device 1 and makes a judgment based on data sent from the sensor unit 2. For example, as shown in the example shown in FIG. 2, based on the distance to the object and its direction data, a stationary object and a moving object as a response target are recognized, or the response target object (human) recognizes the OA device. An action decision such as whether or not to use (or whether or not use has been completed) is made. For a device having a sensor unit composed of an image input device and using image data, the characteristics of an object (human) to be responded to are extracted, the individual is identified by the extracted characteristics, and the individual is identified. ID (for example, name,
Number, etc.), and sends the generated judgment data to the service providing apparatus.

The service providing device 4 has a function of driving each unit of the OA device 1 and implements a service by various automatic responses. For example, when there is a response target object approaching the OA device 1 and determination data indicating that the object uses the OA device is sent, the preheating mode is automatically released, or conversely, the use termination determination is made. A service that automatically transitions to the residual heat mode when data is sent corresponds to this. In the case of an apparatus having a configuration in which a personal ID is transmitted as data, the operation unit is optimized for each user (key layout change, switching of instruction screens,
Provide an environment that is easier to use. This service providing apparatus may be provided with dedicated hardware, but it is also possible to substitute functions as software by a central processing unit of the OA equipment.

FIG. 3 is a block diagram showing a basic configuration of the recognition / judgment device.
The operation of each unit will be described. The OA device on which the automatic response system is mounted is an image forming apparatus, and the configuration of the sensor unit that sends data to the recognition / determination apparatus is such that a plurality of distance sensors having high directivity emit light as shown in FIG. The following description will be made assuming that the direction of reception / light reception / reception is expanded on a horizontal plane and mounted. The recognition determining device 3 includes a sensor driving unit 7, a parameter extracting unit 8, a recognition determining unit 9, and a post-processing unit 1.
0, a control unit 11, an input / output management unit 12, a storage device 13, a data line 14, a control line 15, and an external I / F (interface) line 16.

The sensor driver 7 drives the distance sensor 6 and receives measured distance data. Control unit 1
Each distance sensor 6 is driven based on the sampling signal from 1 to measure the distance to the object. After that, the measurement result data is sent to the parameter extracting unit 8. The parameter extraction unit 8 extracts and calculates feature amount parameters required for recognition of an object to be a response target and various determinations from the distance data to each object from the measured distance data. The generated parameters and the additional information thereof are sent to the recognition determination unit 9 and written into the storage device 13 as appropriate.
Read from other blocks as needed.

The recognition determining section 9 makes a determination relating to the response target object based on a request signal from the control section 11. The parameters generated by the parameter extracting unit 8 are received directly or via the storage device 13 and, for example, whether the response target object is a user of the image forming apparatus (“use” or “do not use” the image forming apparatus) )), Whether the use of the image forming apparatus has been completed (“in use” or “completed”),
And so on.

The post-processing unit 10 puts together the results of the determination and finally prepares the format for external output. For example, the processing when there are a plurality of objects to be responded is performed in this block. The control unit 11 controls the entire recognition determination device. Communication is performed with the outside (image forming apparatus) via the input / output management unit 12, and control is performed by sending a control signal to each block.

The input / output management unit 12 controls an interface with the outside (image forming apparatus) through the external I / F line 16. It also functions as a buffer for synchronizing with the outside. The post-processing unit 10 includes
In addition to the determination data generated in step (1) and sent to the service providing apparatus, control signals such as various requests and timing signals between the recognition determining apparatus and the image forming apparatus are also included.

The storage device 13 is composed of a RAM for storing the data generated in each block as necessary, and a ROM for storing programs and data necessary for operating each block. Data read / write is performed. Data line 14
Is used for transmission of each data. The control line 15 is used for transmitting a control signal. The external I / F line 16 is used for transmitting control signals and data for interfacing with the outside (image forming apparatus).

Sensor Drive Unit The sensor drive unit 7 has a cycle T (T must be a cycle sufficiently shorter than the moving speed of the object to be recognized and responded) from the control unit 11 via the control line 15. The distance sensor 6 mounted on the image forming apparatus is driven according to the sent sampling signal. Each distance sensor is driven simultaneously (parallel) and measures the distance once in one sampling cycle (time interval T). The measured data is converted from analog data to digital data in the sensor driver 7 and the parameters are determined by a method capable of identifying which sensor has measured the data (for example, adding an identification number of the sensor to each data). It is sent to the extraction unit 8.

Parameter Extraction Unit The parameter extraction unit 8 extracts parameters necessary for recognition from the distance data sent from the sensor driving unit 7. As shown in FIG. 4, the distance sensors are arranged at intervals of 10 degrees around the center of the image forming apparatus 5 (however, for the tenth to nineteenth, # 10, # 11,
The measurement is performed in 19 directions (note that it is described as $ 19) (however, the measurement direction is only the measurement surface and the front direction of the image forming apparatus, and the rear surface is not considered here). , The measurement of the distance to the object is repeated at the same time interval T. Each arrow in FIG. 4 indicates the light emission / transmission / light reception / reception direction of the sensor.

The parameter extraction unit performs the following processing at time intervals T (every time measurement is performed). Position Detection (1) Store of Measurement Results As shown in FIG. 4, the image forming apparatus 5 is now installed in a place where there is a stationary object 17 such as a wall in one of a forward direction and a lateral direction. It is assumed that the maximum distance at which the distance to the object can be measured by the distance sensor is Rmax.
The measurable distance is, in addition to the measurable distance of the distance sensor itself, a gap between the measurable directions (a range where no measurement is performed) is sufficiently smaller than the size of an object (human) as a response target. It is in such a range. In the example of FIG.
Among the measurement directions, there is some stationary object (in this case, a wall, hereinafter referred to as an obstacle) 17 within the measurable distance of the distance sensor with respect to the direction indicated by the outlined characters. The parameter extractor 8 stores the distance data in the storage device (memory) 13 together with the measurement direction (distance sensor number) for each measurement. FIG. 5 schematically shows an example of distance data written in the storage device 13 in the case of FIG. Here, in FIG. 5, r _d represents the measurement results for the direction d (distance to the object), also, ∞ symbol object was not anything closer than the object was not measured (Rmax)) It is shown that. The parameter extractor 8 writes the measurement result to a predetermined address in the storage device 13 every time the measurement is performed at the time interval T.

(2) Update of Environmental Information Objects (= obstacles: for example, walls, etc.) that are within the measurable range of the distance sensor and do not actively act on the image forming apparatus 5
The position information of a desk, a chair, etc.) is called environmental information. The parameter extracting unit 8 distinguishes these objects from objects that may actively act on the image forming apparatus 5 (= moving objects: for example, humans, etc .; hereinafter, referred to as target objects). Next, the environment information in the storage device 5 is referred to. The environment information is conceptually in the form of a table of distances in each direction shown in FIG. 5, and indicates at which distance an obstacle is present in each direction.

The environmental information selects the maximum distance measured in each direction during a period (for example, after power-on) sufficiently longer than a period in which the target object is active in the measurable range of the distance sensor in the image forming apparatus 5. It is created by doing. When the maximum distance for the measured direction d within the period and r _d max, environmental information is shown conceptually as in FIG. Each making measurements at time intervals T, and the distance rd is measured for each direction and the environmental information rdmax are compared, If Naritate the relationship r _d> r _d max, replacing environmental information r _d. In this way, after a sufficiently long time has passed, the environmental information is created. For example, after a sufficiently long time (it is sufficient if a moving object such as a human does not exist within the measurement range of the distance sensor) in the image forming apparatus installed in the environment as shown in FIG. The environmental information is as shown in FIG. However, r _d is the distance to the obstacle in the direction d in the case of FIG.

(3) Object Detection After updating the environment information, the parameter extraction unit 8 performs object detection by the following method. For example, the state at time t0 is as shown in FIG.
Consider when approaching. FIG. 9 shows the distance data for the case of FIG. 8 written in the storage device in the procedure of (1). Here, r _dt is the distance to the object in the direction d at time t. The parameter extracting unit 8 stores the distance data and the storage device 5 created by the procedure of (2).
The target object is detected by comparing it with the environmental information stored therein. Specifically, for example, the difference between the distance written in the environment information and the distance data is obtained for each direction. FIG. 10 shows a result obtained by taking a difference between the environmental information and the distance data in the case of FIG. According to FIG. 10, there is a difference between the distance data and the environment information in the direction, and the target object 18 is recognized in the direction based on this (environment / distance data). Now, considering a polar coordinate system in which the origin of the image forming apparatus is the origin and the direction of 〇10 is the angle θ = 0, the position of the target object 18 in the example of FIG. 8 is (r _5t0 , 50 °). expressed. When the object is recognized,
The position (distance and direction) is written to a predetermined position in the storage device 13.

As the moving object 18 approaches the image forming apparatus, the same object may be measured by a plurality of distance sensors. In this case, the position is calculated by the following method. In the example of FIG. 11, the same object is measured by the sensor in and directions. According to the above-mentioned means, two positions (r ₆ , θ ₆ (= 40 °)) and (r ₇ ,
θ ₇ (= 30 °)) is detected. Therefore, when two or more positions are detected, the distance between each of them is calculated, and all the distances are set to a predetermined value Lmin (where Lmin is a response target object (= human) Points that are smaller than) determined from the size of) and whose detection directions are adjacent to each other are collected as one position. In the case of two points, they are put together at the position of the midpoint and 3
If the number of points is equal to or more than the point, the positions are combined at the position of the center of gravity to generate one position. In the example of FIG. 11, the distance l between two points detected is

[0029]

(Equation 1)

When l <Lmin, the two points are combined into one and the middle point is newly adopted as the position.
In the case of three or more points, as shown in FIG.
3 only when <Lmin and l2 <Lmin and l3 <Lmin
The points are collected, and the three centers of gravity G are adopted as the positions of the target object and written to the storage device.

Object Tracking (1) Tracking One Object Once the target object is recognized within the measurable range of the distance sensor, the target object is tracked. For example, at time t1 (= t0 + T), the target object 18 shown in the example of FIG.
_Assuming that the object has moved as shown in FIG. 3, the object position (r _6t1 , 40 °) is detected by the above-described method. Here, if the position information of the target object is stored in the storage device 13 one measurement interval time before (time T before), the moving speed v and the moving direction φ are calculated. For the target object 18 shown in the example of FIG. 13, since the position information calculated in the example of FIG. 8 is already stored in the storage device 13, the calculation of the moving speed v and the moving direction φ is performed.

The calculation method will be described below with reference to FIGS. 8 and 13. The movement distance of the target object 18 between time t0 and t1 is l _t1 , the average speed is v _t1 , a line connecting the coordinate origin (center of the image forming apparatus) and the position of the target object 18 at time t0, and If the angle (moving direction) between the position of the target object 18 and the line connecting the position of the target object 18 at time t1 is defined as _φt1 , the amounts represented by the respective parameters are as shown in FIG. In FIG. 14, _it1 is

[0033]

(Equation 2)

## EQU1 ## The moving speed v and the moving direction φ calculated by Expressions 3, 5, and 6 are written in the storage device 13 together with the position (r, θ) calculated previously. By repeating the above operation at every time interval t,
In the storage device 13, the position information r, θ, and if
If there is position information measured before the current time, the moving speed v and the moving direction φ are sequentially stored as the trajectory information of the object at each time interval t. The trajectory information of the object 18 is stored in the storage device 13 in a data format such as a list or a ring buffer. However, the trajectory information may be conceptually regarded as a table. Target object 18 after time T (= t2) in FIG.
FIG. 15 shows the state of the movement of FIG.
FIG. 16 shows the state of the movement of the object 3), and a conceptual diagram of the trajectory information obtained when the object moves as shown in FIGS. 8, 13, 15, and 16 between times t0 and t3. Are shown in FIG.

(2) Tracking of a plurality of objects When a plurality of target objects exist within the measurement range of the distance sensor 6, a plurality of trajectory information is generated for each target object in the storage device and tracking is performed. For example, consider a case where two target objects, a target object A and a target object B, exist as shown in FIG. Two pieces of trajectory information are generated for the target object A and the target object B in the storage device. FIG.
In the state of (1), the position of the target object A is (r ₆ , θ ₆ (= 4
0 °)), and the position of the target object B is (r ₁₂ , θ ₁₂ (= −
20 °)) is detected, and the respective trajectory information is written. FIG. 19 shows a state one sampling cycle (time interval T) after the state of FIG. Position 1 (r ₇ , θ ₇ (= 30 °)) and position 2 by object detection
Two positions of (r ₁₁ , θ ₁₁ (= −10 °)) are detected. As a possibility, as shown in FIG.
There are two cases in which the target object B has moved to the position 1 and the target object B has moved to the position 2 (case A), or the target object A has moved to the position 2 and the target object B has moved to the position 1 (case B). The following method is used to determine which trajectory information each position is to be written into.

For the above two cases, the moving direction φ and the velocity v are calculated by equations 5 and 6, respectively. The moving direction of the target object A calculated for the case A is φ _A2 , the moving speed is v _A2 , the moving direction of the target object B is φ _B4 , the moving speed is v _B4 , and the target object A calculated for the case B is The moving direction is φ _Ab , the moving speed is v _Ab , the moving direction of the target object B is φ _Bb , and the moving speed is v _Bb . Further, one sampling period (time interval T) previous state (the moving direction of the object A in the state of FIG. 18 phi _Apre, the moving velocity v _Apre, when the moving direction of the target object B and phi _BPRE, the object A movement direction variation .delta..phi _a, the change in the moving speed amount .delta.v _a, the movement direction variation .delta..phi _B of the target object B, and the amount of change in the moving speed δ of
v _B is

[0037]

(Equation 3)

Is represented by Here, defined by Equation 9 below variation e _i about the target object i, further case is defined by equation 10 below the total variation En for n.

[0039]

(Equation 4)

Where α and β are the changes δ in the moving direction.
It is a constant for weighting the change amount δv _i of φ _i and the moving speed. As the combination of the target object and the detection position, a combination of the target object and the detection position in the case where the total change amount is the smallest is adopted. The total changes E _a and E _{b in} case A and case B are

[0041]

(Equation 5)

When E _a <E _b , a combination of the target object and the detection position for case A is adopted,
The trajectory information of the target object A includes position 1 (r ₇ , θ ₇ (= 30
°)), the moving direction φ _Aa and the moving speed v _Aa , and the trajectory information of the target object B includes position 2 (r ₁₁ , θ ₁₁ (= −10 °)), the moving direction φ _Ba and the moving speed v _Ba Are written respectively.

Similarly, in all cases where three or more target objects are present within the range of measurement of the distance sensor, a combination of the target object and the detection position is created.
The data to be written into the trajectory information is determined by calculating the total change amount E in each case. When a target object for which the total variation E cannot be calculated is included, for example, the moving distance 1
Is performed (correspondence between detected positions of objects that are closer to each other compared to one sampling period (time interval T) before) so as to minimize the target object and the detected position.

When a plurality of target objects overlap as seen from the image forming apparatus (exist in the same sensor measurement direction) as in the example of target object A and target object B in FIG.
The number of detected target objects temporarily decreases. In such a case, tracking is performed as follows, and trajectory information is generated. In the example of FIG. 21, the object is detected in the direction, and the position (r ₉ , θ ₉ (= 10 °)) is obtained. Here, the target object A before one sampling cycle (time interval T)
And the position of the target object B, respectively (r _Apre , θ _Apre )
And (r _Bpre , θ _Bpre ), and as shown in FIG. 22, the moving direction and moving speed when the target object A and the target object B are considered to have reached the detected positions are φ _A , Assuming v _A and φ _B , v _B , from equations 5 and 6,

[0045]

(Equation 6)

Is as follows. Further, the moving directions and moving speeds of the target object A and the target object B one sampling cycle (time interval T) are set as φ _Apre , v _Apre and φ _Bpre , v _Bpre
Then, the respective amounts of change e _A and e _B are given by Equation 10,

[0047]

(Equation 7)

Is as follows. Assuming that the detected position is the position of the target object having the smaller change amount, in the example of FIG. 22, the detected position is included in the trajectory information of the target object B from e _B <e _A. Written. In this case, the writing of the trajectory information is suspended for the target object A whose position is not determined, and the trajectory information is written retroactively at the time when the position is determined. For example, _assuming that the position of the target object A is determined to be (r _Apost , θ _Apost ) after one sampling cycle (time interval T), two points (r _Apre , θ _Apre ) and (r _Apost , θ _Apost ) The point is assigned to the reserved position (r _A , θ _A ). Hereinafter, similarly, when the position is determined after n sampling periods (time interval nT), points that internally divide the two points into n points are respectively assigned to the reserved positions and written into the trajectory information.

The same applies to the case where three or more target objects exist within the measurement range of the distance sensor, and the target objects overlap when viewed from the image forming apparatus (exist in the same sensor measurement direction). Then, a change amount e is calculated for each target object, and the target object and the detected position are correlated by comparing them.

Generation of Recognition Judgment Trigger The distance r to the approaching target object 17 is a certain value Lmi
If n or less, the parameter extraction unit 8 sends a recognition determination trigger to the control unit 11 to make a distance determination on the target object 17. When a plurality of target objects exist within the measurement range, this operation is performed every time the distance r to any one of the target objects becomes equal to or less than a certain value Lmin. The recognition determination trigger is converted into a recognition determination request signal by the control unit 11 and sent to the recognition determination unit 9. Similarly, when the target object using the image forming apparatus moves away, the distance r to the target object becomes
When the value exceeds a certain value Lmin, the parameter extracting unit 8 sends a recognition determination trigger to the control unit 11 in order to perform recognition determination on the target object. The case where a plurality of target objects exist within the measurement range is the same as the case where they approach. FIG. 23 shows how a judgment recognition trigger is generated.

The value of Lmin is usually set to a distance at which the recognition determination device must output a recognition determination result (for example, whether or not the target object acts on the image forming apparatus) to the service providing apparatus. This distance is appropriately determined by the type of service provided by the output result from the recognition determination device, the time required for the service, the moving speed of the target object, and the like. The value of Lmax is appropriately set to a distance within the maximum measurement distance (Rmax in FIG. 4) of the distance sensor. After outputting the recognition determination trigger, data relating to the trajectory information of the target object for which the recognition determination needs to be performed (the address in the storage device where the trajectory information is stored, the size of the trajectory data, whether the target object is approaching or not) The recognition determination section 9 sends data required for access to the trajectory information, such as whether the user is moving away, to the recognition determination section 9.

Recognition judging section The recognition judging section 9 makes a judgment relating to the response target object in response to a request from the control section 11. After receiving the recognition determination request signal from the control unit 11, data relating to the trajectory information of the target object that needs to be recognized and determined by the parameter extraction unit 8 (the address in the storage device 13 where the trajectory information is stored, the trajectory data From the parameter extraction unit 8, such as the size of the target object, whether the target object is approaching or moving away, etc. Thereafter, by accessing the trajectory information of the target object in the storage device 13 which needs to perform the recognition determination and performing a predetermined process, for example, when the target object approaches, the user is the user of the image forming apparatus. ("Use" or "do not use" the image forming apparatus), and if it goes away, whether the use of the image forming apparatus has been completed ("in use" or "has been used"), And outputs the result to the post-processing unit 10. In the following, some examples of processing for determining whether to use the image forming apparatus or not to use the image forming apparatus when the target object approaches in the recognition determination processing will be described.

Preprocessing In order to determine the recognition of the target object, the recognition determining unit 9 generates some parameters from the trajectory information. One or more reference distances are provided in the measurement range of the distance sensor. For example, as shown in FIG. 24, n reference distances L0, L1, L2,..., L (n-1) from the center of the image forming apparatus (where L0 <L1 <L2 <... L (n-1)) Is established). Here, the reference distance of the closest distance L0 from the image forming apparatus is Lmin
(Distance at which the recognition determination trigger is generated by the parameter extraction unit). The trajectory information at the reference distance Lm is time _tLm , distance _rLm (= Lm), direction _θLm , moving speed _vLm , moving direction _φLm, and the distance L (m−
The change amount of the moving speed between 1) and Lm is represented by Δv _Lm (= v _Lm
−v _L ( _m−1 )), and the amount of change in the moving direction is Δ _Lm (= φ _Lm −φ _L
( _m-1 )), and a parameter is generated for each distance (however, the change amount Δv _L0 of the moving speed and the change amount Δφ _{0 of the} moving direction at the distance L0 are excluded). FIG. 25 conceptually shows the parameters generated for each distance. The generated parameters are further processed so that they can be used in the next stage. For example, the data is grouped according to a value range and converted into the group number, or a normalization process for a certain value range is performed.

Recognition Judgment Processing Among the methods for performing recognition judgment using the parameters generated in the preprocessing stage, the method using a neural network will be described below. For example, the following neural network 21 is prepared for determining whether the target object uses or does not use the image forming apparatus. FIG. 26 is a schematic diagram of a neural network 21 for recognition determination by a hierarchical neural network including three layers, an input layer 22, a middle layer 23, and an output layer 24.
The input layer 22 corresponds to each parameter to be input, and the output layer 24 corresponds to each determination (in this case, “use” and “do not use”). The neural network used here learns by a predetermined method (for example, back propagation) using characteristic patterns collected in advance through experiments and the like as teacher data. That is, each parameter is input, and it is learned whether the target object "uses" or "does not use" the image forming apparatus at that time.

At the time of the recognition judgment processing, the parameters generated from the trajectory information by the control signal from the control unit are input to the neural network for recognition judgment, and among the output layers 24 of the neural network at that time. The result corresponding to the unit that fired strongly (having the largest output value) is output to the post-processing unit 10 as a recognition determination result. With the same specifications, a neural network for determining whether the target object is “in use” or “finished” of the image forming apparatus is prepared, and the recognition determination result is output to the post-processing unit 10.

Post-Processing Unit The post-processing unit 10 collects the results of the recognition judgment and arranges the final output to the outside.
, And sends recognition determination data to the input / output management unit 12. For example, consider a case where a plurality of target objects exist in the measurement range of the distance sensor as shown in FIG. In the case of the example (a), a target object B who is going to newly use the image forming apparatus comes to a place where the target object A is already using the image forming apparatus. in this case,
When the target object B approaches the distance Lmin from the image forming apparatus, a recognition determination trigger is generated, and the recognition determination unit 9 determines that the target object B "uses" the image forming apparatus. If not suspended, a problem occurs that the service for the target object B is started even though the target object A is in use. Conversely, as in the case of the example (b), it is assumed that the target object A as the previous user has left while the target object B is using the image forming apparatus. In this case, when the target object A approaches the distance Lmax from the image forming apparatus, a recognition determination trigger is generated, and the recognition determination unit 9 determines that the target object A has used the image forming apparatus. If the target object B is not masked or held, a problem occurs that the service is started even though the target object B is in use (transition to the residual heat mode, automatic reset, etc.). The post-processing unit 10 fetches external (image forming apparatus) information (for example, whether or not copying is in progress) through the control unit 11 if necessary, and masks or suspends determinations regarding a plurality of target objects. Only when it is necessary to change the state, the recognition determination data is sent to the input / output management unit 12.

Control Unit The control unit 11 controls the entire recognition judging device. The main control contents are as follows.

(1) Timing Processing of Each Block Data of each block (sensor driving unit 7, parameter extracting unit 8, recognition determining unit 9, post-processing unit 10, input / output managing unit 12) in the recognition determining unit is Performs synchronous processing for delivery. For example, when data is transferred from a certain block A to a certain block B, first, a data transmission request is issued from the block A to the control unit 11. The control unit 11 sends a data reception request signal from the block A to the block B, and after confirming that the reception preparation of the block B is completed, a data transmission request signal is sent to the block A. In block A, after receiving the data transmission request signal from the control unit 11, data transfer is started.

(2) Request processing for each block A request sent from each block in the recognition / judgment device (for example, a recognition judgment trigger generated in the parameter extraction unit, a recovery request for an error generated in each block,
Etc.).

(3) Storage Device Access Management Arbitration is performed so that a plurality of blocks do not access the storage device at the same time (so that read / write does not collide). Each block in the recognition / judgment device can access the storage device only by requesting an access right from the control unit and only when permitted.

(4) Control of Interface with External The clock signal and the reset signal from the external (image forming apparatus) are received through the input / output management unit 12, and the external and the respective blocks in the recognition judging device are synchronized. Also,
External information (the state of the image forming apparatus, for example, “operating / not operating”, information on the operation panel, etc.) is fetched, and those necessary for recognition determination are transferred to the corresponding blocks.

Various requests from the outside are sent to the input / output management unit 12.
Process that responds to the request.
For example, when the trajectory information of the target object is requested from outside,
This corresponds to processing such as reading out locus information from a predetermined area in the storage device and outputting it through the input / output management unit 12. In addition, a request to the outside when an error or the like occurs in the recognition determination device (for example, displaying an error on an operation panel, etc.) is also generated by the control unit 11, and the input / output management unit 12
Output to the outside through

Input / Output Management Unit The input / output management unit 12 controls an interface with the outside (image forming apparatus) through the external I / F line 16.
Further, it also functions as a buffer for synchronizing with the outside, and each input / output data / signal is latched and synchronized by the input / output management unit 12. The input / output signals include, in addition to the determination data on the target object generated by the post-processing unit 10 and sent to the service providing device, various control signals between the recognition determination device and the outside, control signals such as timing signals, and the like. included.

Storage Device The storage device 13 is a RAM for storing data generated in each block as needed, and a ROM for storing programs and data necessary for operating each block.
And data reading / reading by each block.
Writing is performed.

Service Providing Apparatus Outline of Service Providing Apparatus The service providing apparatus 4 has a function of driving each section of the OA equipment, receives various judgment results generated by the recognition judging apparatus 3, and provides a service by an automatic response corresponding thereto. To embody. Further, if necessary, it requests the recognition determination unit for more detailed information on the target object. The service providing device 4 is a conceptual device, and several methods can be considered for its realization. In particular, it is better to realize such a service that does not specify the provision of dedicated hardware and that can perform functions as software by a central processing unit of an OA device.

Service for People When there is a response target object (person) approaching the OA device, and when the judgment data indicating that the object uses the OA device is sent, the following service is provided. Execute. (1) AI (Artificial Intelligence) residual heat function (2) AI automatic power switching function (3) AI pressure plate automatic opening / closing function (4) AI automatic interrupt function (5) Height variable function

When a person who has been using the OA device ends the operation, leaves the OA device, and receives determination data indicating that the use has been completed, the following service is executed. (1) AI residual heat function (2) AI automatic power switching function (3) AI pressure plate automatic opening / closing function (4) AI automatic interrupt function (5) AI automatic reset function (6) AI original / copy forgetting check function

Service for Individuals For an apparatus having a configuration in which there is an object (person) to be responded to approaching the OA device and the ID of the individual is transmitted as data, the following service is executed. (1) Operation panel AI automatic setting function (2) AI guidance function (3) Management function (4) AI automatic interruption function (5) AI auto reset function (6) Message notification function

Next, user recognition / personal recognition based on image data will be described. User Recognition In the above-described automatic response system, an example in which a distance sensor such as an infrared sensor is used as a sensor has been described. Next, based on information obtained by processing the image data, not the distance information, such as the body and face orientation when the target object (person) approaches, is the person a "person to use"? ,
A method of recognizing / determining whether a person is a “mere passer” will be described. FIG. 1, which is a configuration example using a distance sensor, FIG. 2, which is an implementation example, and FIG. 3, which is a basic configuration of a recognition determination device, are completely the same in the case of a recognition method using image data described below. . However, the distance sensor 2
It must be a sensor for inputting some kind of image data using a CCD, MOS, image pickup tube or the like. The parameter extracting unit 8 in FIG. 3 extracts necessary information from image data or the like by image processing or the like.

Now, consider the operation of a person approaching to use the image forming apparatus 5 such as a copying machine provided with the present recognition determining apparatus. In particular, in the absence of obstacles, one can usually think of a person walking straight to the device he wants to use. In other words, the user approaches the device he or she wants to use while keeping his body straight. Also, even if there is some obstacle, etc., if you can not approach straight, if you come within the range where the device can be seen by the time you approach the position of the target device, you can look at the device at least several times You can think. In other words, the direction of the face is directed several times straight toward the device to be used. Such behaviors in which humans behave naturally are stored as rules, and these behaviors are extracted / determined by image processing or the like to determine whether the target person is a “user” or a mere “passer”. I do.

Hereinafter, a simple example of a specific determination method will be described with reference to FIG. In the photographed image data, it must first be determined whether a person has been photographed (step 1). This can be achieved, for example, by holding image data in which no one is shown and calculating the difference from the captured image data. In other words, by removing the background portion from the currently captured image data, the remaining portion becomes a moving object other than the background,
In other words, it can be seen that the subject person.

Next, in the image data of the target person, the orientation of the body or face must be detected (step
2). Speaking of detection, it is only necessary to judge whether or not this is facing. As a very simple example, if the body or face of the photographed person is facing this direction, the person may be determined as a user. However, it's usually hard to judge just because you're facing him only once. In other words, this alone does not provide sufficient accuracy. Therefore,
For example, a counter may be provided, and the number of times that it is observed that the user is facing the user may be stored by the counter. .

Instead of yes / no, for example, the orientation may be converted into a numerical value based on the number of angles and the like, observed, and a determination may be made based on the data sequence. In this case, the data string pattern of “user” may be stored in advance, and the determination may be made by referring to the stored table, or an appropriate weight parameter may be added using time and direction as parameters. Judgment may be made by an evaluation function, or the behavior pattern of the “user” may be described by natural language rules (for example, if you turn here many times in a short time, you are the user.) Judgment may be made by fuzzy inference, or the behavior pattern of the "user" may be stored in the neural network in advance, and the observed data sequence may be used as input to output from the neural network whether or not the user is "user". You may.

The configuration for carrying out the present recognition processing is basically not significantly different from that described above, but the distance sensor 2 in FIG. 1 is replaced with an image input sensor such as a CCD. In addition, image data of the approaching target object (target person) is processed by the parameter extraction unit 8 in FIG. The feature quantity (data / information) to be extracted here is the face or body orientation of the target person described above. Further, the recognition determining unit 9 refers to the table described above,
Judgment processing is performed by an evaluation function, fuzzy inference, a neural network, or the like.

Personal Recognition Recognizing that a target object that is a person is a specific individual is called personal recognition. Personal recognition is divided into two larger recognitions. One is to identify individuals who have been registered in advance (identification that has already been individualized), and the other is not to register individuals in particular, but to target individuals observed at a certain point in time. , To determine whether they are the same person (sequentially identifying the same person).
In the above-described automatic response system, an example in which a distance sensor is provided to measure the position of a target object as a sensor has been described, and an infrared or ultrasonic sensor has been described as a specific example. Here, an example relating to personal recognition of a target person is shown. In this case, a specific example of the sensor is a CCD.
An example in which image data is processed after an image is input by, for example, will be mainly described.

In this case, the recognition judging device 3 shown in FIG.
, And the distance sensor 2 is merely a sensor capable of inputting image data, but the basic configuration is not changed at all. Hereinafter, one image input sensor will be described as an example for simplicity of description, but a plurality of sensors may be provided as shown in FIGS. In addition,
Even if an image input device such as a CCD is used as a sensor, it is possible to observe the position up to the target object described in the aforementioned automatic response system without any problem using current image processing techniques ([[Robot "Bit extra number, pp
711/724, July 1976 issue], ["Stereoscopic vision", Journal of the Robotics Society of Japan, Vol. 1, pp30 / 35, 1983]. In other words, the image input sensor is only a specific example of the distance sensor. For example, in order to obtain the position of the target object from image data captured from one fixed sensor, an image of the position of the foot of the target person is used. Data coordinates may be extracted and mapped to real world coordinate data. If there are a plurality of sensors, the position to the target object can be observed by applying the principle of triangulation (binocular stereovision). Therefore, even if the sensor is changed from the above-mentioned infrared sensor or the like to an image sensor, the above-described processing is similarly possible, and the sensor is also useful in personal recognition described below.

Personalized Specific Identification The processing method of the personalized specific identification for specifying the registered individual from the image data will be described.
Identification technology has been around for a long time, and has been put to practical use, for example, for character recognition. The same technique / principle can be used for personalized identification. To explain the principle briefly, we prepare a dictionary called a dictionary that describes the characteristics of what is to be specified, extract the characteristics of the observed input data, compare them with the contents of the dictionary, Is to be identified. In addition to the discriminant approach, a table reference method, an evaluation function method, a certainty calculation method, a fuzzy inference method, a neural network method, and the like are also known.

Information about an individual, which must be registered in advance, as data for identifying the individual, that is, a characteristic, may be any characteristic obtained by image processing or the like, but must be good for identifying the individual. Must. For example, height, weight (volume), gender, size and shape of body and face, presence or absence of glasses, and the like can be considered. The height is obtained by converting the number of pixels from the head to the toe from the image data to the actual length. Since the distance to the target object must be known, the conversion is easy. Finding the head and toe can be realized by ordinary image processing techniques such as binarization and filter processing and knowledge engineering. The gender can be determined by obtaining the shape of the hair, the covering, the presence or absence of makeup, and the like by image processing. Facial features are effective and often used to identify individuals by image data. For example, the relationship between the components and the shape of the individual components, such as the degree of eye separation, the shape of the eyes, the contour of the chin, and the like, are often used.

With the personal recognition of the already personalized identification,
It is possible to improve the operability of the user, save labor in managing the apparatus, and the like. For example, when this personal recognition device is attached to an image forming apparatus such as a copying machine,
The operation panel AI automatic setting function, AI guidance function, AI automatic interrupt, and the like can be realized.

Next, as described above, instead of registering an individual in advance, the target person observed at a certain point in time is recognized as the same person or not. The same person determination will be described. In this method, the characteristics of the target person observed at a certain point in time are stored, and it is determined whether the characteristics match with the characteristics of the target person observed at another point in time. It is not necessary to register information including the characteristics of the individual you want to identify as a dictionary in advance, but instead, for example, if someone does not artificially enter the device, such as the name of the observed person, you can not understand No information can not be obtained. The method for successively identifying the same person is basically not significantly different from the method in the already-personalized identification. In individual identification, based on observed / extracted features, the characteristics of each individual, which were previously described in a dictionary, were compared. The difference is that they compare with the observed features. In the same person determination, the characteristics of the target person observed at a certain point in time are stored as a dictionary and used when comparing with the characteristics of the person observed later, but only the immediately preceding observation object is stored in the dictionary. Alternatively, a plurality of storages may be made.

Since the same person can be sequentially discriminated, for example, it is known that the user of the apparatus has been changed, so that "in use / completed use" can be distinguished at that time. When a person is identified as the same person, it can be recognized as “in use” within a certain period of time, and if a person who is not the same person comes, the previous person is recognized as having used it. it can. If the apparatus is a copying machine, for example, the parameters can be automatically reset (AI auto reset function) to prevent the next user from erroneous operation. Further, even if the AI automatic interrupt function cannot be specified even to an individual, it can be realized by the real recognition. Also,
The sensor for observing the feature in the same person determination sequentially is not particularly limited to the CCD. For example, any device that can measure the characteristics of a person, such as a color sensor, a weight scale, and a sound sensor, may be used. In addition, there are sensors that can be distinguished by themselves, such as the above-described image input sensor using a CCD or the like. However, even sensors that are difficult to distinguish by themselves can be distinguished with higher accuracy by using them in combination. .
The configuration in this case is only required to connect these sensors in parallel with the distance sensor 2 shown in FIG.

Method of processing parameters for recognition judgment: Initial condition setting method for predetermined measurement detection area A person approaching is a person who comes to operate a system (apparatus) provided with the present recognition apparatus. When determining whether a person is just passing by or not, it is desirable to recognize and store the surrounding situation in order to make the decision early or more reliably depending on the situation in which the system is set. The "situation" mentioned here is, for example, a location of an obstacle such as a surrounding wall. For example, in the case where the system is installed at the end of a dead-end corridor, it is considered in most cases that the person approaching the system is the person who uses the system. here,
A method for automatically recognizing situations such as "the system is installed at the end of the dead end corridor" will be described. The information to be detected is, for example, a fixed object such as a wall, which hinders / obstructs human walking. What is necessary is just to find these objects at a certain timing. For example, a service person or a user who installs a system to which the present recognition device is attached can directly input to the system by some method.
It has to be reset, which is troublesome. Here, a method of automatically recognizing these will be described.

The following conditions (a) and (b) are required for recognizing data (area data) such as an obstacle as an initial condition in a measurable detection area by the sensor or the like described in the above-described recognition device. , (C). (A) Area data is stored from detected values detected after a predetermined time has elapsed since the power was turned on (blocking obstacle data based on regular light reception / reception data) A predetermined time predetermined from the time when the power was turned on This is a method in which the sensor is activated later, and the location where some object is detected by the sensor is recognized as a fixed object (obstacle) and stored. Immediately after the power is turned on, there is a possibility that the person who turned on the power is nearby and recognizes this as a fixed object. Then, after the power is turned on, after a while (for example, one minute), the sensor is activated, and the place where something is detected is recognized as a fixed object and stored.

As one of the embodiments, the case where the initial conditions are set by the above-described sensor will be described. 1 and 2 show the configuration of a system (OA device 1 or image forming apparatus 5) to which the present recognition device is attached and the configuration of the sensor 2. FIG. Here, the sensor is mounted on the system in a plurality of directions. When this system is turned on by the user, the timing of turning on the power can be easily known by any means, for example, the power of the recognition device is shared with the power of the system. A timer is installed in the device or the like, and each sensor is activated after a predetermined time set in advance, for example, one minute, and a predetermined measurement is performed. These sensors are
At least some object can be measured in what direction and at what distance. For example, FIGS. 4 and 5
In such a case, a certain object is indicated by a combination of the observed direction and distance, and at ~ 〇13, 〇15 to 〇19,
It recognizes that there is an obstacle and stores it.

(B) Renewal of area data stored at predetermined timings from detected values detected at predetermined timings after power-on The sensors are activated at predetermined time intervals after power-on, This is a method of comparing the detection status of the sensor at the time of the previous measurement with the current detection status, recognizing where the fixed object is, and updating the storage. The fixed object should be detected at the same place no matter how many times the sensor performs detection, except when the situation is not artificially changed due to a layout change or the like. On the other hand, if the object is not originally an obstacle, such as a person who happened to be near by chance, it is first observed by the sensor, but at the next measurement, it has already left and should not be observed. So, for example,
This is a method in which the detection status is obtained from the sensor every five minutes, the past detection status is remembered, and the one that is always detected at the same place is recognized as a fixed object. The predetermined timing does not need to be observed at all times, that is, at minimum intervals.
However, if you only want to get area data,
There is no particular need for constant observation, and measurement may be performed at certain intervals. For example, if you always activate the sensor to make some kind of judgment, use that,
It is more reliable if the area data is constantly updated, but if there is no particular need for constant observation, it is better to measure at some intervals to save power.

In realizing this method, it is necessary to store the past detection status at least once. However, a plurality of times may be stored, which increases the recognition accuracy but increases the cost. When the past detection status is stored only once, the stored previous detection status is, for example, θ = 15 degrees, distance = 3 m and θ = 8.
It is assumed that the position is 0 degree and the distance = 2 m, and in the detection state measured this time after a predetermined time, 30 degrees, 4 m and 80 degrees, and 2 degrees
If it is m, it is recognized / determined that there is some fixed object (obstacle) in the 80 degree, 2m part detected at the same place. And 30 degrees and 4 meters as the past detection situation;
80 degrees and 2 m are stored.

Also, when a plurality of detection situations are stored, the determination may be made in substantially the same manner. For example, when the situation for five times is stored, it is determined that there is a fixed object in a place that has been detected at the same place three or more times in succession recently,
Although it is not continuous, it is 4
It is determined that there is an obstacle that can be an obstacle in places where the number of times has been detected. By doing so, it is possible to more accurately determine whether or not the obstacle is an obstacle. As a result, besides a truly stationary object such as a wall, for example, a word processor, for example, is installed near the system to which the recognition device is attached, and a human using the word processor is detected by the sensor. In other words, it is possible to deal with a case where the system is originally a moving object, but temporarily becomes an obstacle to the system for some reason, and a correct installation environment, that is, an initial condition can be obtained.

(C) Storing area data from detected values detected under predetermined conditions (updating storage of area data for each predetermined condition) Timing for recognizing area data is performed only when predetermined conditions are satisfied. Is the way. The predetermined condition is
Basically, it is a situation where no one is around. The sensor itself may be activated constantly or irregularly,
Whether or not the area data should be recognized is determined by referring to predetermined conditions and stored. As the predetermined condition, the following four cases can be considered.

(I) When not in use Measurement / recognition is performed when the system to which the present recognition device is attached is not operating, that is, when no one is using it.
When the system is not running, it is basically considered that there are no nearby people. That is, if the measurement is performed at that time, it may be determined that there is an obstacle at a location where some detection is made.
In this case, in one measurement, a location detected at that time may be recognized as an obstacle, or a number of locations may be measured several times and determined as an obstacle. .

(Ii) Night with no people: timer In a normal office or the like, there are no people at night. Therefore,
At this time, measurement / recognition is performed. When a sensor is activated by a timer or the like at a time when no person is present, such as at night, and if any object is detected, it is assumed that the object is an obstacle such as a wall. Further, it is possible to more reliably determine that there is no person by using not only the timer but also the brightness of the lighting. For example, as a specific time setting, the default is 3 o'clock in the middle of the night, and if it is desired to change the time, it may be set by the user,
It may be set in conjunction with an ON / OFF timer incorporated in the system for another use, at the time of OFF, immediately before being turned off, immediately after being automatically turned on, or the like. In this case, at least one measurement is sufficient, and the place where something was detected at that time may be identified as an obstacle, or several times during a time when nobody is present just in case. Measure
The locations where many are detected may be obstacles.

(Iii) Holiday: Weekly timer In a normal office, there are no people on holidays. Therefore, measurement / recognition is performed at this time. When a sensor is activated by a calendar (weekly timer) or the like on a day when no one is present, such as Sunday, if any object is detected, it is assumed that the object is an obstacle such as a wall. The specific time may be set by default to Sunday, and may be changed by the user if it is desired to change the setting. In this case, at least one measurement may be performed, and a location detected at that time may be recognized as an obstacle. Is also good.

(Iv) Initial condition setting means ON Here, means for setting initial conditions are released to the user. For example, an operation button such as a "initial structure condition setting button" is attached to the apparatus. When a service engineer, user / administrator, or the like determines the layout, etc. and determines that initial conditions should be set, the user can explicitly press this button to recognize / set the initial conditions. is there. However, immediately after pressing the button, the person who pressed it is in front of the device, so it is set in advance so that it will be set a few seconds later, like a self-timer often used for cameras etc. The operator of the button may be caused to retreat outside the detection range of the sensor. Alternatively, the means may be operated from a remote place using an infrared remote controller or the like.

Next, a method of early recognition of “use / not use” or “used / finished” according to the machine installation environment will be described. If it is recognized from the machine installation environment data (area data) based on the initial condition settings that the machine (equipment) has been installed in a corner, the object (moving object) just enters the set detection measurement area. It can be recognized as "use". That is, depending on the installation environment of the machine, it is conceivable that if a person simply comes, that person is always the person who uses it. For example, if a machine is installed at the end of a dead-end aisle, those who come nearby may be considered to be those who use it.

The installation environment (area data) of the machine can be obtained by the above-described method. At this time, basically, it is a problem whether a person can physically pass / overpass the front of the machine. This is judged from the area data, and if it is impossible for a person to physically pass by the front of the machine, simply use a distance sensor to determine whether a person is near. Look, and if they come close, they will be judged as "users". Also, if the sensor is configured to have a plurality of sensors oriented in multiple directions, it is possible to save power by not operating the sensor that measures the direction determined as an obstacle. There is also an effect.

The condition in which most of the approaching people can be determined to be “users”, that is, a situation where a person cannot physically pass / exceed the front of the machine is that the machine is installed at a dead end. If you are. The situation where the machine is in a dead end may be, for example, a case where the end of a dead end passage or two surfaces are placed at the corners of a wall. Is only the front of the machine accessible?
Alternatively, such recognition / judgment is performed in a case where the vehicle can pass only in a horizontal (right or left) direction or the like. Due to this, although there are certain restrictions on the installation environment, there is no particular difficulty, and the "user" can be determined quickly and easily easily.

In the above, the example of recognizing and judging "use / do not use" for the approaching person has been described, but the method of recognizing "in use / finished" for the leaving person is exactly the same. Needless to say, it can be realized with a simple concept.

Next, a method of recognizing "used / not used" or "used / used" of the measurement object (moving object) will be described. (A) The orientation of the face or body is determined based on the image data obtained by processing the image signal from the CCD, and it is recognized whether "used / not used" or "used / finished". When a person approaches to use a system (device) placed in a hallway or room, it usually tries to reach the system as straight as possible. In other words, from the viewpoint of the system, a person who comes to use may think that the body walks while facing the front. However, depending on the size / narrowness of the space before walking, it is possible that the body does not face the system straight and comes closer, but at least at some point, the system is seen unless it is very unconscious. Therefore, if this is detected, it can be assumed whether the person is approaching for use or not. In order to detect body orientation and face orientation, CCD
And the like may be processed and recognized.

Whether the body is facing front or not
For example, a shoulder line is extracted, and if it is close to horizontal, it is known that it is facing the front, and if it is inclined, it is known that it is not facing the front. Also, in the image of the person facing sideways, the front part of the body such as the chest appears very small, so that it can be determined whether the person is facing the front.

Whether the face is facing the front is
Judgment can be made based on the shape of the nose, the line connecting both eyes / ears, and the degree of reflection. The resulting orientation is
For example, it can be extracted as numerical data such as an angle, and it is also possible to output two values of “facing / not facing” in comparison with a preset threshold value. . These data extraction processes can be realized without any problem by using a conventionally used general image processing technique.

If it is observed that the body / face direction of a simply photographed person is facing the present recognition device, it is possible to immediately determine that the person is a “user”. . However, it is not enough accuracy. Therefore, for example, a counter may be provided to store the number of times that the user is facing the object is stored, and it may be determined that the device is “in use” when the number of observations is equal to the preset number of times. In order to further increase the accuracy, it is necessary to handle numerical data (columns) represented by angles or the like, instead of binary values of “facing / not facing”.

As a method of processing the parameter, the following method can be considered. (I) Table reference method Observing possible behavior patterns of “in use” in advance, finding a data sequence observed within a certain time,
Remember it. For example, if the data sequence is [5,3,0] or [−2,1, −1] in the data sequence observed three times, it is stored as a “user” pattern. As a storage method, for example, since it is indicated by three sets of data in the present example, it is stored as a three-dimensional table such as (x = 5, y = 3, z = 0) = “user” flag You should keep it. At the time of recognition determination, it is only necessary to determine that the user is a "user" if the "user" flag is set by referring to a table in which the data observed every moment is stored.

(Ii) Evaluation function method An action pattern that the "user" can take is observed in advance, and an evaluation function to which a weight function is added is set using some data in the data sequence as a parameter. For example, an evaluation function such as k1 * d1 + k2 * d2 + k3 * d3 + k4 * d4 is set for a data sequence obtained in order as d1, d2, d3, and d4. For example, k1 = 0.5,
When k2 = 0.6, k3 = 0.8, and k4 = 1, d1 = 80, d2 = 80, d3
= 85 and d4 = 90, 246 is obtained from the evaluation function equation. D1 = 80, d2 = 9
Assuming that 0, d3 = 80, and d4 = 75, 233 is obtained from the evaluation function equation. When the observation data indicating the direction is such that 0 is obtained when the target person is facing sideways, and 90 when the target person is facing the front, the former is an example in which the front is gradually facing the front, The latter is an example of changing direction or trying to cross. Here, the numerical value obtained by the evaluation function formula is compared with an appropriately set threshold value in advance and "use"
Is determined. For example, if the threshold value is 240, it is determined that the user is “user” in the former example, and “not a user” in the latter example. In the above example,
Although it is a simple example such as k * d, d does not have to be the direction of the observed angle or the like, for example, d2-
It is also possible to use d1 as one data.

(Iii) Fuzzy Reasoning Method As described in the example of the above (ii) evaluation function method, for example, a case in which "facing gradually toward the front" or a case in which "facing constantly in the front" Is a "user." Also,
On the other hand, in the case where the direction suddenly deviates, for example, it can be understood that the person is a “non-use person”. By preparing a number of such natural language rules and expressing these ambiguous languages such as "gradually" and "frontal" by membership functions of fuzzy logic, recognition judgment by fuzzy inference can be performed. Become.

(Iv) Neural network method The features obtained from the data string of "users" are input as input, and the neural network is trained. At the time of recognition judgment, the observed data string is input / unused. The recognition is determined by the output of. In the above, an example of recognizing and judging "use / do not use" for an approaching person has been described.
It is needless to say that the method of recognizing is also possible with the completely same concept.

(B) It is recognized from the image data obtained by processing the image signal from the CCD whether or not the user is "in use / completed" based on whether or not the user is the same person. Performs image processing based on image information (data) obtained from a CCD or the like,
Various characteristics of the user are collected / stored and then compared with the measured characteristics of the user in the same manner to determine whether or not the user is the same person. If there is, the previous user judges that it has already been used.

First, as a method of determining whether or not two persons are the same person, the following method may be used. Image data around a specific portion of a person in use is photographed and held. Depending on how the CCD is attached to the system, an image near the face of the user in use may be captured, or the entire image may be captured. It is also possible to take a picture before approaching. Various features are extracted from the image data thus photographed. For example, when taking a picture near the face, the portrait of the face
The horizontal length, the degree of eye separation, the distance between the nose and the mouth, the size of the hair, the presence or absence of glasses, and the like are extracted as feature amounts. It is natural that parameters related to size, length, etc. are normalized at the image level, or to be normalized to the parameter determination formula when extracting as parameters, and technically usually Used. In other words, when an image is taken, the same person may appear small or large depending on the time. For example, if you shoot while you are far away, it will appear small.
Processing such as affine transformation may be performed on the image data by calculating the distance to the subject at that time, or the distance may be corrected for the number of pixels of the face width obtained from the image, for example. If the distance to the subject is not known, an invariable feature amount may be used for the distance such as the roundness of the face (vertical / horizontal of the face).

When it is determined whether or not the person is the same person, how to determine “in use / completed use” can be determined by a person who is determined to be the same person within a certain time. Comes, it is recognized that the person is in use, and when a person who is determined not to be the same person comes, it recognizes that the previous person has already used it.

Next, a method for recognizing a solid object to be measured (moving object) will be described. Height, weight, gender, face (size, shape, hairstyle, etc.) are obtained from the characteristic data of the measurement target object, for example, image data obtained by processing an image signal from a CCD.
The glasses (e.g., glasses), clothes (shape, color), and the like are determined, and a solid object to be measured (moving object) is recognized. This is intended to identify who is using the system provided with the recognition device. The image information (data) obtained by an image input device such as a CCD is processed, The characteristic parameters of the person are extracted, and reference is made to a standard characteristic dictionary for each individual, which has been input / stored in advance, to recognize / determine which person is among them. Identification technology has been around for a long time, and has been put to practical use, for example, for character recognition. The same technique / principle can be used for personalized identification.

In brief, the principle is described. A dictionary called a dictionary, which describes the characteristics of the object to be specified, is prepared in advance. The characteristics of the observed input data are extracted and compared with the contents of the dictionary. The one that matches is specified. In addition to the discriminant approach, a table reference method, an evaluation function method, a certainty calculation method, a fuzzy inference method, a neural network method, and the like are also known.

Information about an individual, which must be registered in advance as data for identifying the individual, that is, a feature may be any feature obtained by image processing or the like, but it must be good for identifying the individual. Must. For example, height, weight (volume), gender, size and shape of body and face, presence or absence of glasses, and the like can be considered. The height is obtained by converting the number of pixels from the head to the toe from the image data to the actual length. Since the distance to the target object must be known, the conversion is easy. Finding out the head and toes can also be realized by ordinary image processing techniques such as binarization and filter processing and knowledge engineering. The gender can be determined by obtaining the shape of the hair, the covering, the presence or absence of makeup, and the like by image processing. Facial features are effective and often used to identify individuals by image data. For example, the relationship between the components and the shape of the individual components, such as the degree of eye separation, the shape of the eyes, the contour of the chin, and the like, are often used.

A method of recognizing "use / not use" when a plurality of measurement objects (moving objects) exist in the detection measurement area will be described. Here, there are cases where the determination is made only with the distance data and cases where the determination is made with the distance data and the direction data. First, the case where the determination is made only with the distance data will be described. In this example, a case where there are two moving objects will be described for the sake of simplicity.

First, consider a case where two moving objects both approach the apparatus. First, when two moving objects approach from the same direction, measurement is impossible because the longer distance is behind the shorter distance when viewed from the recognition device. Therefore, in this case, recognition is performed using only the data of the shorter distance. In addition, when approaching from each other direction, infrared light pulsatingly emits light, and the amount of light returned by the infrared light is captured by a light meter, so that the pulsating light emission of infrared light returned from two or more objects in other directions is obtained. The change in quantity in one cycle is
As shown in FIG. 29, the mountain becomes a composite mountain of two. By judging this using a judging means such as a neuro, a plurality of moving objects can be identified. Using this data, it is determined whether the user is a user.

Next, consider the case where one approaches and the other passes by. In this case, it is conceivable that a passing moving object may be a shadow of an approaching object, or be cast as a shadow. In other cases, two moving objects can be identified by the above-described method, and a processing method at this time will be described. First, when another moving object crosses the approaching moving object, the moving object that has crossed is judged as it is, and for the approaching moving object, based on the data before entering the shadow, as long as it is in the shadow Predict its movement. After that, when it comes out of the shadow, tracking is started again. Next, in the case where another moving object passes behind the approaching moving object, the judgment is made as it is for the approaching moving object, and the moving object that traversed enters the shadow based on the data before entering the shadow. Predict that movement as long as you have. After that, the tracking is started again when the moving object comes out of the shadow. When there is a moving object other than the user, the user who has prioritized the judgment on whether to use the device from the time when it is determined that the user and the other moving object have been switched in terms of distance has been used. From to a moving object.

Next, a case where the determination is made based on the distance data and the direction data will be described. First, when two moving objects come closer, two moving objects from different directions
However, when viewed from the same direction, one becomes in the shadow of the other and cannot be detected by the sensor. Therefore, the determination is made based on the distance and direction data of the moving object close to the image forming apparatus. Also, if it is detected out of the shadow on the way, it is determined as a newly appearing object. Next, when approaching from different directions, since different sensors detect each of them, they can be regarded as separate moving objects, so it is determined whether or not to use the device from each of the movements.
In this case, the one closer in distance to the image forming apparatus is determined first.

Next, consider the case where one approaches and the other passes by. In this case, it is conceivable that a passing moving object may be a shadow of an approaching object, or be cast as a shadow. In other cases, two moving objects can be identified by the above-described method, and a processing method at this time will be described. First, when another moving object crosses the approaching moving object, the moving object that has crossed is judged as it is, and for the approaching moving object, based on the data before entering the shadow, as long as it is in the shadow Predict its movement. After that, when it comes out of the shadow, tracking is started again. Next, in the case where another moving object passes behind the approaching moving object, the judgment is made as it is for the approaching moving object, and the moving object that traversed enters the shadow based on the data before entering the shadow. Predict that movement as long as you have. After that, the tracking is started again when the moving object comes out of the shadow.

In this case, the priority order for the determination is set to the object closest to the apparatus, and if the distance is the same,
Let's start with the one approaching fast. By using the above method as a basic idea when judging whether or not an apparatus is going to be used, it is possible to make a reliable judgment.

Next, a description will be given of a method of determining when the object to be measured (moving object) recognizes the user of the image forming apparatus during use. In this example, the image forming apparatus has an infrared light emitting unit and an infrared light receiving unit, emits infrared light emitted by the infrared light emitting unit, and receives infrared light reflected by an object by the infrared light receiving unit. When distance data is obtained by using a detecting unit that detects a distance to an object near the device, and when the user is recognized based on the obtained distance data, use the device at the time of starting use. Since the user is considered to be at the closest distance, the moving object at the closest distance is always marked and tracked to determine whether or not the moving object is being used.

When the determination is made based on the distance data and the direction data, the same determination method as that based on only the distance data may be used to reduce the number of sensors to be detected by reducing the number of sensors to be detected, or to lengthen the detection timing to increase the timing of the recognition device. It is also possible to reduce the burden. Further, the following is another determination method.

FIG. 30 is a view of an apparatus for making a judgment of a user by using data of the distance and direction of the image forming apparatus.
5b is a group of a plurality of ultrasonic sensors. Infrared sensor 5a
Has, in the detection range, a portion close to the apparatus from the opposite end direction of the paper discharge port toward the paper discharge port of the apparatus.

Until the time when the image forming apparatus is used, the infrared sensor 5a is not operating, and the ultrasonic sensor 5b determines whether the moving object is a user. Next, when the image forming apparatus is used, the detection and determination by the ultrasonic sensor 5b are stopped, and the detection of the user is started by the infrared sensor 5a. That is, as long as the moving object is detected, it is determined that the moving object is in use. Next, when it comes out of the detection range of the infrared sensor,
The operation of the infrared sensor 5a is stopped, and at the same time, the operation of the ultrasonic sensor 5b is started to determine the end of use. With the above method, it is possible to accurately determine whether the user uses the image forming apparatus continuously.

[0121]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention relates to a device having a discriminating means for discriminating and detecting a change of a person using the above-mentioned apparatus, for example, in an image forming apparatus, discriminating a change of a user and performing an automatic reset according to the result. AI (Artifi
cial Intelligence) Auto reset function.

In the conventional auto reset function, the screen is automatically reset to the initial setting screen according to the time setting from the end of the user's operation. Therefore, even when the user does not want to reset, the reset is often performed. In order not to perform such unnecessary auto-reset, the present invention determines that the user has been replaced, so that unnecessary auto-reset is not performed unless the user is replaced.

FIG. 2 is a diagram showing an example of an image forming apparatus to which the present invention is applied. The image forming apparatus 5 is provided with the recognition judging device as described above, and is capable of discriminating between individuals or the same person sequentially. In order to perform this operation, an object in which the distance sensor 6 is replaced with a sensor such as a CCD for inputting image data is used.

FIG. 31 is a flowchart showing the flow of processing when performing an AI auto-reset according to the present invention. First, before the image forming apparatus is used, personal recognition (feature recognition in the case of sequential same person determination) is performed. , By the recognition determination device shown in FIG. Next, the recognition determining device recognizes and determines the next user from the time when the copying operation is completed. Here, if it is determined that a person other than the previous user tried to use the automatic reset function, the operation panel is changed to the initial setting screen. The initial setting screen at this time can be customized for the user by using data in the storage device 13 when the individual is performing recognition.

[0125]

Turnover was determined detection used for persons according to the present invention, when the turnover is detected, (the course of the copying operation is finished after a predetermined time, a predetermined distance or more away from the device user's) auto reset function enable conditions Regardless, by operating the auto-reset function, a malfunction due to an erroneous use condition setting of the device is reliably prevented, and a use condition setting input control device with extremely excellent operability and handleability can be realized. Furthermore ,
Since the use conditions at the time of use immediately before are memorized and can be called up by one operation, it is excellent in convenience even when the former user once leaves the device and immediately uses the device under the same use conditions. A use condition setting input control device can be realized.

[Brief description of the drawings]

FIG. 1 is a block diagram for explaining a configuration concept when an automatic response system is applied to an OA device.

FIG. 2 is a diagram illustrating an example of a case where a sensor is mounted on an image forming apparatus.

FIG. 3 is a block diagram showing a basic configuration of a recognition device.

FIG. 4 is a diagram illustrating a relationship between an image forming apparatus and an obstacle.

FIG. 5 is a diagram schematically illustrating distance data between an image forming apparatus and an obstacle.

FIG. 6 is a diagram illustrating an example of environment information for an image forming apparatus.

FIG. 7 is a diagram showing environment information in FIG. 4;

FIG. 8 is a diagram showing a state of change of environmental information.

FIG. 9 is a diagram showing distance data in FIG. 8;

FIG. 10 is a diagram showing a result obtained by taking a difference between environment information and distance data in the case of FIG. 9;

FIG. 11 is a diagram showing a state of change of environmental information.

FIG. 12 is a detailed view of a main part of FIG. 11;

FIG. 13 is a diagram showing a state of change of environmental information.

FIG. 14 is a detailed view of a main part of FIG. 13;

FIG. 15 is a diagram showing a state of change of environmental information.

FIG. 16 is a diagram showing a state of change of environmental information.

FIG. 17 is a diagram illustrating a concept of movement trajectory information of a target object.

FIG. 18 is a diagram illustrating a state of change of environmental information.

FIG. 19 is a diagram showing a state of change of environmental information.

FIG. 20 is a diagram showing a state of change of environmental information.

FIG. 21 is a diagram showing a state of change of environmental information.

FIG. 22 is a diagram showing a state of change of environmental information.

FIG. 23 is a diagram for describing recognition determination for a target object.

FIG. 24 is a diagram for describing an example of recognition determination when a target object approaches an image forming apparatus.

FIG. 25 is a diagram conceptually showing parameters generated for each distance in the example of FIG. 24;

FIG. 26 is a diagram illustrating an example of a hierarchical neural network.

FIG. 27 is a diagram for describing an example in which a plurality of target objects exist within a measurement range of a distance sensor.

FIG. 28 shows whether the target person is a user of the image forming apparatus,
It is a figure showing an example of the way of judging whether or not.

FIG. 29 is a diagram for describing an example in which two or more objects are identified by pulsating emission of infrared light.

FIG. 30 is a diagram illustrating an example in which a user is determined using an infrared sensor and an ultrasonic sensor.

FIG. 31 is a flowchart of a process when performing an AI auto reset.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... OA equipment, 2 ... Sensor part, 3 ... Recognition determination device, 4 ...
Service providing device, 5: image forming device, 5a: infrared sensor, 5b: supersonic sensor, 6: distance sensor, 7: sensor driving unit, 8: parameter extracting unit, 9: recognition determining unit, 10
... Post-processing unit, 11 ... Control unit, 12 ... Input / output management unit, 13
... Storage device, 14 data line, 15 control line, 16 external I / F, 17 obstacle, 18 target object, 21 hierarchical neural network, 22 input layer, 23 intermediate layer, 24 ... the output layer.

────────────────────────────────────────────────── ─── Continued on the front page (51) Int.Cl. ⁷ Identification code FI G01S 17/88 B41J 29/00 G03G 15/00 G01S 17/88 B G06F 1/24 G06F 1/00 350B (72) Inventor Ichinomiya Takashi 1-3-6 Nakamagome, Ota-ku, Tokyo Ricoh Co., Ltd. (72) Inventor Fukuo Ichikawa 1-3-6 Nakamagome, Ota-ku, Tokyo Examiner at Ricoh Co., Ltd. Yuji Tsuruya (56) Reference Reference JP-A-59-123858 (JP, A) JP-A-1-160448 (JP, U) (58) Fields investigated (Int. Cl. ⁷ , DB name) H04N 1/00 G03G 15/00

Claims (1)

(57) [Claims] 1. An operation unit for setting and inputting use conditions when using an apparatus, determining means for determining and detecting replacement of a person using the apparatus, and switching of a person using the apparatus by the determination means. If but detected, that have a and automatically switching control means the use conditions set input from the operation unit to a predetermined initial operating conditions
In the condition setting input control device used in the control means
The operating conditions of the device to the predetermined initial operating conditions.
Use when used immediately after switching automatically
The condition is stored in the storage means and before the condition is stored in the storage means.
The operating conditions are called by one operation input, and the device is operated under the operating conditions.
Control means for automatically setting the case
Use condition setting input control device.