JPH0471069A - Camera system - Google Patents

Abstract

PURPOSE:To easily retrieve an image relating to a specific speaker out of speaker's voices recorded together with photographing images by providing the camera system with a camera capable of recording voices at the time of photographing and a reproducing machine capable of retrieving and reproducing a required photographed image based upon the collation of the voice. CONSTITUTION:The reproducing machine is provided with a voice registering means for registering voices. Namely, units from a microphone 10 to a memory 13 are constitution for voice input to input the voice of a photographed person or the like. The system is also provided with a collating means for collating a voice registered in a voice registering means with a voice recorded in a recording medium and an output control means 20 for reproducing and outputting an image corresponding to a coincident recorded voice obtained by the collating means. The voice inputted at the time of recording the photographed image is recorded correspondingly to the image, and at the time of reproducing the photographed image, the previously registered voice is collated with the voice obtained at the time of photographing by voice recognition prior to the reproducing of the image and the coincident recorded voice obtained as the collated result is retrieved. Consequently, a required photographed image is retrieved based upon the voice recorded at the time of photographing the image by the camera.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention consists of a camera that can record audio when shooting, and a playback device that can search for and play back desired captured images based on audio verification. Regarding the camera system.

[Conventional technology]

BACKGROUND ART Conventionally, a methyl video camera has been provided in which a camera is provided with an audio input section, and the audio at the time of recording the VA shadow image is also recorded.

Furthermore, a data search device has been proposed in which a key customer is registered in advance and the input key voice is compared with the registered key voice at the time of a subsequent data search to search for the corresponding voice data (Japanese Patent Laid-Open No. 1-114898 Publication No.).

(Problems to be Solved by the Invention) The still video camera that records the above-mentioned sound together with the photographed image is merely a device that allows the user to simply enjoy the sound playback when the photographed image is played back, and uses the sound to perform the desired operation. It is not intended to extract photographed images.

Furthermore, the data retrieval device described in Japanese Unexamined Patent Publication No. 1-114898 searches for audio data using a simple key voice recorded in the data retrieval device itself, and uses the audio recorded in the recording means of the camera. It does not search for a corresponding photographed image based on a comparison with the above.

The present invention has been made in view of the above, and an object of the present invention is to provide a camera system in which a desired photographed image can be searched based on the audio recorded during photographing with a camera.

[Means to solve the problem]

A camera system according to the present invention includes a camera having a recording medium that records photographed images and audio in association with each other, and a playback device that reproduces the photographed images recorded on the recording medium, The playback device includes a voice registration means for registering voice, a collation means for collating the voice registered in the voice registration means with the voice recorded on the recording medium, and a matching recorded voice by the collation means. and output control means for outputting images for reproduction.

[Effect]

According to the present invention, the recorded image is recorded in association with the sound input at that time, and when the captured image is played back, the sound registered in advance is compared with the sound at the time of the shooting using voice recognition. be done. Then, when the matching recorded voice is retrieved as a result of this verification, a photographed image recorded in association with the searched voice is selected and output,
It is played by a playback machine.

(Embodiment) FIG. 1 shows an example of a block diagram of a camera that captures images.

In the figure, a lens 1 to an encoder 8 are configured for photographing images.

2 is an image sensor (
(hereinafter referred to as COD), which transports the object image obtained through the lens 1. The amplifier 3 amplifies the output image signal from the C0D 2 at a predetermined amplification factor and outputs it to the next stage AD converter 4.

The AD converter 4 converts the input analog image signal into a digital image signal. White balance correction circuit 5
The digital image signal is subjected to white balance correction based on color temperature data from a white balance sensor 14, which will be described later, and the γ correction circuit 6 further performs γ correction. Furthermore, the matrix processing circuit 7 performs predetermined gradation correction on the digital image signal. The encoder 8 encodes the input digital image signal into an NTSC signal for reproduction images.
Output the signal to memory card I/F9.

The microphone 10 to the memory 13 are configured for audio input.

A microphone 10 is provided at a suitable location on the camera body and captures the voices of the photographer, the person to be photographed, and the like.

The amplifier 11 amplifies the audio signal from the microphone 10 at a predetermined amplification factor and outputs it to the next stage AD converter 12. The AD converter 12 converts the input analog audio signal into a digital audio signal. The memory 13 is used for voice recording (for speaker recognition), and temporarily stores input digital voice data, and then stores the input digital voice data in the memory card I/O.
Output to F9.

Next, a configuration for importing various other information will be explained.

The white balance sensor 14 detects the color temperature of the subject, and determines from the obtained color temperature data whether the photograph was taken using fluorescent light correction or under sunlight, and determines the appropriate white color. Apply balance. The AF sensor 15 is used to calculate the distance to the subject using a known distance measurement method, and for example, separates the subject image that has passed through the lens 1 into two directions and sends the image to each light receiving section (not shown). It is configured to guide the image of the subject. The subject images obtained by the two light receiving sections are input to the AFCPU 16, and the AFCPU 16 calculates the phase difference from the subject images of the two light receiving sections, and calculates the distance to the subject image from the phase difference. calculate.

Furthermore, the AFCPU 16 drives a lens drive system (not shown) based on the calculated distance data to move the lens 1 to the in-focus position. When the movement to the in-focus position is completed, focal length data f from the lens 1 and subject distance data D from the AFCPU 16 are output to a control unit (hereinafter referred to as CPU) 20, which will be described later.

A temperature sensor 17, a humidity sensor 18, and an atmospheric pressure sensor 19 are used to respectively measure the temperature (air temperature), humidity, and atmospheric pressure at the carving machine. Each measured data is processed by CPU20
Output to. Each of the above-mentioned sensors automatically measures each data during photographing and guides it to the CPU 20.

The reception tuning section 21, the demodulation section 22, and the GPS receiver 23 are configured to obtain position information. The reception tuning unit 21 generates a code (
The demodulator 22 receives FM radio waves containing codes indicating regions, tourist spots, events, etc. (hereinafter referred to as location codes), and the demodulator 22 converts the received FM signals into I! The location code data is reproduced and output to the CPtJ20. Note that details will be explained with reference to FIG. 2. On the other hand, the GPS receiver 23 receives radio waves from a satellite (not shown) with an antenna 23a, and outputs the latitude, longitude, and altitude of the current position to the CPU 20.
The details are explained in FIG. 3.

The clock section 24 is for clocking the date and time built into the camera body.

The operation and lead control system is comprised of the CPLI 20 and an operation and display section 25 that outputs the operation status of each operation switch 81 to S3 to the CPU 20.

The CPU 20 is a microcomputer that centrally controls the operations of the various parts of the camera described above. The photographing preparation switch S1 starts the operation of each of the sensors. The exposure start switch S2 is a so-called release button that starts the exposure operation of the camera. Preferably, the photography preparation switch S1 is activated when the release button is pressed halfway, and the exposure start switch S2 is activated when the release button is pressed fully. Further, the operation and display unit 25 outputs the switching states of the switches $1 to S3 to the CPU 20, and displays various data from the CPU 20 at the time of photographing, such as frame numbers, as necessary.

Under the control of the CPU 20, the memory card I/F 9 stores the image data, audio data, and various photographic information in a memory card 26 made of SRAM or E2PROM that can be attached to and detached from the camera body.
6 letters of use! II (presence or absence of storage area a, etc.) is output to the CPU 20.

The photographic information stored in the memory card 26 is used as search information during filing, as will be described later.

Figure 2 explains the location code transmitting and receiving device.
In the same figure (A> indicates a transmitter that is installed in each region, tourist spot, event venue, etc. and transmits the location code, and in the same figure (B)
> indicates a receiver installed in the camera body that receives the location code.

In FIG. 2A, a ROM 201 has a location code, for example, a location (event) identification code, a Kanji JIS code string of a location (event) name, or both previously written therein. The DPSK modulation section 202
The output code data from OM201 causes a differential phase shift It (DPSK: [)differential
phase 5hift keyin (N).The FM modulation aFg203 applies FM modulation to the carrier wave according to the input serial data, and furthermore, the amplifier 2
After being amplified to a predetermined power in step 04, the signal is transmitted from the antenna 205 as an FMIi wave including a location code.

The timing generation circuit 206 controls the timing of each block on the transmitter side, and outputs read address data to RO to 1201 and clock pulse GK to the DPSK modulation section 202 and FM modulation section 203.

Note that the DPSK modulation is performed by the timing generation circuit 20.
6 clock pulses CK are used to extract phase difference data from digital data by AFM in the audible range, MFM in medium waves, PLL method, etc., and convert it into serial data.

Next, in FIG.
It receives i-waves, and the received signal is amplified by an amplifier 207 and input to an FM demodulator 208 . Received signal is FM
A demodulator 208 removes the carrier frequency, and a next-stage DPSK demodulator 209 demodulates the signal to the original code data. The CPU 20 takes in the vI-toned code data and stores it in the memory card 26. That is, C.P.
U20 writes the code data to the memory card 26 in association with the writing of the photographed image at that time. The timing generation circuit 210 controls the timing of each block on the receiver side, and controls the FM demodulation section 208 and DPSK'jaW4 according to the inheritance control signal from the CPtJ20.
A clock pulse CK is output to section 209.

Therefore, when a camera image is taken at a location where the transmitter is installed, such as a certain area, tourist spot, or event venue, the location code is automatically stored in the memory card.

The ROM 201 may be one that stores only a location code specific to that location, or one that has multiple location codes written in advance and can be switched to output the location code depending on the installation location. But that's fine.

FIG. 3 is a detailed block diagram of the GPS receiver 23.

The antenna 23a is of, for example, a quadrifilar helix (quadrifilar heX) type, which receives transmitted radio waves from a NAVASTAR IIjl (not shown).

RF signal mixer 23 received by this antenna 23a
It is input to 0. On the other hand, the modulator 231 is the local oscillator 23
The local oscillation signal CK+ from 2 is sent to the PN code generator 233.
The spread modulation signal is input to the mixer 230. As a result, the IF signal, which is the intermediate frequency of the RF multiplied signal, is converted into an IF low signal and input to the data demodulation circuit 234.

This data demodulation circuit 234 demodulates data including the time when the satellite transmits the signal from the input signal. The demodulated data is input to a data processing circuit 235 and a delay measurement circuit 236.

The delay measurement circuit 236 first outputs a timing signal to the PN code generator 233 when demodulated data is input. The PN code generator 233 always generates PN by the clock pulse CK2 from the PN code clock generator 237.
A code is generated, and when the timing signal is input, the generated PN code is sent to the delay measurement circuit 236. And data demodulation circuit 234
PN code from and PN from PN code generator 233
The code is led to a delay measuring circuit 236, and the delay time of the PN code necessary for correlation between both PN codes is measured. The delay time of this PN code is determined by the measurement clock generator 238.
The counted value is outputted from the delay measuring circuit 236 to the data processing circuit 235 as delay data necessary for correlating both PN codes.

The data processing circuit 235 is composed of a microprocessor and is driven by the clock pulse CK4 from the data processing clock generation circuit 239, and includes transmission time data included in the demodulated data from the data demodulation circuit 234 and a built-in GPS receiver. From the reception time data obtained from a cesium and rubidium vapor atomic clock (not shown), the radio wave propagation time from the satellite to the GPS reception power camera is determined,
The distance from the satellite to the GPS receiver II (camera) is calculated from the time. Then, the data processing circuit 235 calculates the position information regarding the latitude, longitude, and altitude of the camera (II!Actor) from the distance information from each satellite and the position information of each satellite itself included in the 1m data. and outputs it to the CPU 20.

As is well known, GPS is a system that covers the entire world using 18 NAVASTARs, but currently not all satellites have been launched, and there are areas where reception is not possible or where reception is unavailable depending on the time of day. The CPU 20 automatically stores location code data from the location code receiver in areas where reception is not possible, while storing data from the satellite when no location code transmitter is installed and no location code is received. Both receivers are controlled so that they receive data, take in the received data as position information, and interpolate each other.

FIG. 4 is a flowchart explaining the operation of the camera.

In step #1, wait for the power (not shown) to be turned on,
When the power is turned on (YES in step #1), the presence or absence of the memory card 26 is determined (step #2). If there is no memory card 26, a message indicating that there is no card will be displayed and a warning that photography (recording) is not possible will be displayed and the process will return to step #1 (step #3, step #4). The camera may have a memory capable of recording multiple frames.

If the memory card 26 is installed, "card present" is displayed (step #5), and then the usage status of the memory card 26, that is, the presence or absence of recordable free space is detected (step #6). ). If there is no free space (
(YES in step #7), a warning is issued indicating that recording is not possible, and the process returns to step #1. Note that, as described above, when a built-in memory for a plurality of frames is provided, whether or not recording is not possible may be determined in consideration of the presence or absence of free space in the built-in memory.

On the other hand, if recording is possible (NO in step #7), I
It is determined whether the shadow preparation switch s1 is on or not (step #9). If the switch S1 is not on (No in step #9), it is determined that the camera is not taking pictures and the process returns to step #1; if it is on (YES in step #9), the process moves to step #10. Then, various operations and processes for preparing for shooting are performed. Specifically, the AF sensor 15 measures the distance to the subject, the WB sensor 14 measures the color temperature of the object, the temperature sensor 17 measures temperature, the humidity sensor 18 measures humidity, the air pressure sensor 19 measures atmospheric pressure, and the GPS receiver 23 measures the temperature. Positioning and reception of FM radio waves including the location code by location code reception 121 and 22 are performed sequentially (step #10 to step #
16). In the above steps, all photographic information from each sensor etc. is input, but a switch not shown may be used to select desired photographic information so that only desired photographic information can be input.

When the input operation of shooting information from each of the above sensors etc. is completed,
Next, it is determined whether or not the exposure start switch $2 is on (step #17). Above switch S2
If the switch S1 is not on (No in step #17), the process proceeds to step #18, and the state of the switch S1 is checked again.Here, if the switch S1 is on (step #1
If the switch S1 is off (NO in step #18), the process returns to step #10 and sensing is performed again using the sensors, etc., in order to obtain more accurate photographic information immediately before the photograph is taken. ), it is determined that photography has been canceled and the process returns to step #1.

If switch S2 is on (YES in step #17)
), the distance measurement data is locked, and then focal length information f from the lens 1 and object distance information from the AFC CPU 16 are stored (step #19 and step #20).

Then, the colorimetric data obtained by the WB sensor 14 is locked, and then the voice input operation using the microphone 10 is started (Step #21. Step #22). In addition, exposure is performed in parallel with the darning operation (step #23), and when the exposure is completed (YES in step #24), predetermined signal processing for capturing the photographed image is started (step #24). 25).

Next, in step #26, the audio input selector switch S3 is turned on.
It is determined whether or not the switch S3 is on, and if the switch S3 is on (YES in step #26), all audio is captured (step #27), and if it is not on (no in step #26), For example, audio is captured for one second as a reference (step #28).

After the above operations, image data, audio data, and search information (1@shadow information) obtained from each sensor, etc. are written into the memory card 26 in a corresponding format (see FIG. 27) (Step #29> , the free space information on the memory card 26 is updated, and the frame number is counted up by 1 (step #3o).

Step #31), prepare for the next shooting.

FIG. 5 is a block diagram showing an example of a photographed image playback device equipped with a search function.

This reproducing machine is centrally controlled by a CPU 50. This CPU 50 is programmed in advance to perform various control functions in accordance with the content of each search and reproduction process described later.

The memory card I/F 51 is used to guide recorded images, search information, and data regarding the usage status of the memory card 26 from the memory card 26 to the playback device, and also to guide read image selection data from the playback device to the memory card 26. This is the interface. The special reproduction processing section 52 performs special processing on the image read from the memory card 26 as will be described later, and the image data processed by the special reproduction processing section 52 is temporarily stored in the image frame memory 53.

The composition section 54 composes image data and a superimposed image, which will be described later, and is switched and controlled by an on-screen display 0N10FF switching control signal from the CPU 50. The combined image signal is sent to the DA converter 55
The signal is converted into an analog signal at , and further amplified at the next-stage amplifier 56 , and then guided to a TV monitor 57 for reproduction and display.

The superimpose memory 58 stores various search information and on-screen display patterns such as scales. The read clock generation circuit 59 supplies clock pulses for read timing to the image frame memory 53, the synthesis section 54, the DA converter 55, and the superimpose memory 58, respectively.

The character generator 6o stores the character font of the Japanese Industrial Standard JIS code string, and outputs the corresponding character code to the CPU 50 when the JIS code string is adopted as the location code. The filing device 61 is composed of a recording medium such as an optical disk and its driving circuit, and is a so-called album that appropriately records recorded images in the memory card 26. This filing device 6
1 is controlled by the CPU 50 for reading and writing, and images;
It records the registration search information and outputs the registration search information and disk management information to the CPU 50.

The keyboard 62 is used for inputting and modifying search information, and also for specifying the position of the search information recorded in the filing device 161, especially maps and place names. It is connected. The voice input section 64 is used to input voice when registering voice for search, and is connected to the E2 PROM 65 for speaker data.
encodes the input voice at the time of voice registration and records it as registered voice.

In addition, the audio data memory 66 is connected to the memory card I/F 51.
It temporarily stores the audio recorded on the memory card 26 that is input via the IDA converter 67.
, and is guided to a speaker 69 via an amplifier 68 and reproduced as audio. The speaker recognition unit 70 recognizes the voice data input via the memory card I/F 51 and the E2PRO for the speaker data.
This is to determine whose voice the voice is by comparing it with speaker data registered in the M65. The judgment result for this speaker is CP
It is also output to U50 and used as an index during a search.

Note that 71 to 73 are printers, FAXs, and TVIs for transmitting and outputting stored images and search images.

In the above configuration, next, the processing on the playback machine side is shown in Figures 6 to 2.
This will be explained using the flowchart shown in FIG.

First, a procedure for registering images in the memory card 26 in the filing device H61 will be explained with reference to FIG.

When the memory card 26 is inserted into the player, step #
41), the recorded images in the memory card 26 are sequentially read out (step #42). The recorded image that has been read out is subjected to reproduction processing for multi-screen use in the special reproduction processing section 52, and then
Multi-display is performed on the monitor 57 (step #43). Multi-display is performed by allocating a predetermined number of frames to one frame, or a number of frames set according to the number of recorded images in the memory card 26.

Next, the CPU 50 accepts the selection of an image desired to be registered in the filing device 61 from among the images in the multi-screen,
Next, the frame number of the corresponding image is input using the keyboard 62 or the like to select it (step #44), and the multi-image display is performed again for the selected image (step #44).
Step #45). Next, in step #45, it is determined whether the search information input on the camera side is to be modified or added. Examples of such corrections and additions include correcting a person's name when the result of speaker identification is displayed incorrectly, correcting position information, and adding comments.

If there is no modification of the search information (N in step #46)
o) By performing a registration operation with the keyboard 62, the image and search information are recorded in the filing device 61, and the recorded image and search information are erased from the memory card 26 (step #47, Step #48). This erasing operation creates a free space in the memory card 26, making it possible to record new captured images.

On the other hand, if there are corrections, etc. (YES in step #46)
, When an image that requires correction or the like is selected using the keyboard 62 or the position indicating member 63, the image and search information are displayed on the TV monitor screen (step #49, step #
50). The user makes necessary corrections and additions while looking at the screen (step #51). Every time the correction, etc. of one image is completed, the other images are asked whether or not the correction, etc. is necessary (steps #51 to #4).
(Return to 6). When the necessary corrections and the like are completed, the process moves to step #47 and the registration process ends.

Next, the procedure for registering audio will be explained with reference to FIGS. 7 and 8.

Figure 7 shows “Voice Registration
'', and FIG. 8 shows the procedure of ``audio registration ■'' in which only audio is registered. Both steps are performed in step #51 above.
It may be executed in , or it may be provided as a separate voice registration mode.

In FIG. 7, first, after the CPtJ 50 is enabled to accept voice input, voice is input from the microphone of the voice input unit 64 (step #61), and the input voice is A.
D-converted and encoded (steps #62, #63)
. This encoded audio data is stored in E2 for recording speaker data.
It is registered in the PROM 65 (step #64). Next, the CPU 50 makes it possible to accept data from the keyboard 62, and when the speaker's name is input from the keyboard 62 (step #65), it is encoded (step #66).
, further associated with the voice data of the speaker E2
It is registered in the PROM 65 (step #67).

Therefore, as will be described later, when searching by voice, the speaker's name is superimposed on the TV monitor 57 along with the search image by superimposition.

In FIG. 8, first, after the CPU 50 is enabled to accept audio input, audio is input from the microphone of the audio input unit 64 (step #71), and the input audio is AD converted and encoded. (Steps #72, #73). This encoded voice data is used as an EP for recording speaker data.
It is registered in the ROM 65 (step #74).

Therefore, only the searched image will be displayed on the TV monitor 57 when searching by voice.

Next, various search procedures will be explained with reference to FIGS. 9 to 21.

FIG. 9 shows the main flow of the search process.

First, select which item (condition) to search using the keyboard 62.
etc., the CPU 50 collates the search item with each image or search information (step #81), sequentially reads matching images from the memory card 26, and displays them on the TV monitor 57 in multiple formats (step #82). . Then, the user selects the desired one from among the multi-displayed images.
Or multiple images are selected (step #83)
.

When the selection for the current display screen is completed, it is determined whether all other multi-screens that display images that match the search items have been displayed (step #84), and if there are other multi-screens, the current screen is will be updated to the next screen (step #
85). Then, when image selection is completed for the multi-screen to be updated (Y in steps #83 and #84)
ES), the selected image is output in a desired output format as described below (step #86).

FIG. 10 shows a flowchart of "item search" for the search items shown in step #81.

In this example, the search items are "location", "date and time", "
Weather,'' ``Names of People,'' ``Indoors and Outdoors,'' ``Scenery of People,''``Events,'' and ``Sound,'' etc.

When “Item Search” starts, first, “
An inquiry is made as to whether or not to select "location" (step #91). When the "location" is selected, the "location search" flow shown in FIG. 11 is executed (step #92
). If it is not selected, then an inquiry is made as to whether or not to select "date and time" as a search item (step #9).
3).

When "Date and time" is selected, "Date and time search" shown in Figure 12
The flow is executed (step #94).

If it is not selected, then an inquiry is made as to whether or not to select "weather" as a search item (step #95).
. When "weather" is selected, the "weather search" flow shown in FIG. 13 is executed (step #96). If it is not selected, then an inquiry is made as to whether or not to select "Person's name" as a search item (step #97). If "Person's name" is selected, the "Person name search" flow shown in FIG. 14 is performed. is executed (step #98). If it is not selected, then an inquiry is made as to whether or not to select "indoor/outdoor" as a search item (step #99).If "indoor/outdoor" is selected, "indoor/outdoor search" shown in FIG. 16 is performed. The flow is executed (step #100). If not selected,
Next, an inquiry is made as to whether "portrait scenery" should be selected as a search item (step #101). When "portrait scenery" is selected, the flow of "portrait scenery search" shown in FIG. 17 is executed (step #102). If it is not selected, then an inquiry is made as to whether or not to select "event" as a search item (step #103). "
When "event" is selected, the "event search" flow shown in FIG. 19 is executed (step #104). If it is not selected, then an inquiry is made as to whether or not to select "audio" as a search item (step #105).
.

When "Voice" is selected, the [Voice search] shown in Figure 20
Step #106 is executed. continue,
Re-selection is possible in consideration of changes or additions to search items (NO in step #1o7). Thereafter, the "item search" process is ended and the process returns to step #82.

These search items depend on each sensor on the camera side, and it is also possible to select two or more items at the same time.

The details of each of the above search processes will be explained below with reference to FIGS. 11 to 21.

In the flow of "location search" shown in FIG. 11, the location code and/or GPS positioning data serve as an index of location information.

First, it is determined whether a map is required to be displayed on the TV monitor 57 (step #111). When displaying a map, that is, when indicating the shooting location on the map, use the filing device 6.
1 or the map stored in the built-in memory (not shown) of the CPLI 50 or the removable map-dedicated memory (also not shown) is guided to the TV monitor 57 and displayed (step #112), and the position of the mouse, etc. is indicated. Specify the position on the display map using member 63 (step #113)
. The indicated position data (latitude, longitude) is calculated based on the displayed map and the output position signal of the position indicating member 63.

Alternatively, the detection may be performed using a tablet or the like, for example. On the other hand, if map display is not required, the place names and places recorded in association with the recorded images are read out and displayed, and the desired place name etc. from among the place names etc. that are not displayed is specified by inputting it from the keyboard 62. (Step #1
14, #115). Subsequently, the CPU 50 searches for an image matching the specified position or input place name, etc. (step #116). This search is performed by scanning search information recorded in association with each image. When the search is completed, the number of matching items is displayed as the search result (step #117). At this time, the frame number of the corresponding image may be displayed. After this, it is determined whether a re-search is necessary (step #118
). Examples of re-search include a case where the number of matching items is large or, conversely, a case where the number of matching items is zero. In the case of a re-search, if another search condition is input or specified, the search by the CPU 50 is performed in the same manner as above (steps #111 to #1).
17). Then, when the search ends, step #118
(No), return to step #93.

In the flow of "date and time search" shown in FIG. 12, the date information from the timer 24 serves as an index.

Step #121), C
A search for images that match the search conditions input into the PU 50 is executed (step #122). This search is performed by scanning the search information (date and time data) recorded in association with each image, as described above. In addition, CPU50
There is a storage means in which the season and date and time data are associated with each other, and when the search condition is season, the input season is converted into the corresponding date and time data in the storage means, and the converted Search is performed using date and time data.

When the search is completed, the number of matching items is displayed as the search result (step #123). Note that, as described above, the frame number of the corresponding image may be displayed. After this, it is determined whether a re-search is necessary (step #124). In the case of re-searching, if other search conditions are input, the search will be performed by the CPU 50 in the same manner as above (
Step #121 to Step #123). Then, when the search is completed (No in step #124), the process returns to step #95.

In the flow of "weather search" shown in FIG. 13, the temperature and humidity information from the temperature sensor 17 and humidity sensor 18 serve as indexes.

Conditions related to weather, such as "sunny", "rainy", "I want"
, "It's cold", etc. are input from the keyboard 62 (step #131), the CPU 50 searches for images matching the input search conditions by scanning the search information (step #132). The CPU 50 is provided with a storage means that stores in advance the correspondence between the above-mentioned weather-related conditions and weather data consisting of temperature and humidity. For example, if you enter "@i", the temperature (air m) is 30℃.
The search may be performed based on a certain predetermined temperature and humidity, such as those listed above or those with a humidity of 80% or higher.

Furthermore, the reference temperature and humidity may be changed depending on the location, date and time (and season). For example, if you want to search for "summer + station", the temperature is 30 degrees Celsius or higher, and the temperature is 7.8 degrees.
For September, on the other hand, for "winter + hot" the temperature is 25℃.
The above is the one for December, January, and February.

When searching, the input weather-related conditions are converted into corresponding weather data in the storage means, and the search is performed using the converted weather data.

When the search is completed, the number of matching items etc. are displayed as the search results (step #133). After this, it is determined whether a re-search is necessary (step #134)D.If you want to re-search, enter other search conditions, and the CPU
50 is performed (step #131 to step #133).

Then, when the search is completed (No in step #134)
, return to step #97.

In the flow of "person name search" shown in FIG. 14, 1. "
"Voice" or "Person's name J serves as an index.

When a "person's name" is input from the keyboard 62 (step #141), the CPU 50 searches for images matching the input search conditions (step #142). This process is performed according to "search process (■)" shown in FIG. That is, first, it is determined whether the search is by "voice J" or "person name J" (step #151). In the case of a search by "voice", the input person's name is encoded and converted into a person name code; The voice code corresponding to the person name code is extracted from the E2 PROM 65 (step #
153, #154). Next, the audio data recorded in association with the pre-registered image is extracted from the filing device 61 and encoded (step #1
55, #156).

Then, the audio code from the E2PROM 65 is compared with the audio code from the filing device 61 (step #157). This verification is in filing VR161
This is performed for all images by sequentially scanning the audio codes of the images registered in the image (NO in steps #156 to #158, loop of step #159).

When the matching for all images is completed (YES in step #158), the process returns to step #143. On the other hand, if the search by person's name is selected in step #151, the search is performed by scanning the person's name code input at the time of image registration to the filing device 61 (step #152), and when the search is completed, step #143 Return to.

Returning to FIG. 14, the number of matching items etc. are displayed as the search results (step #143). After this, it is determined whether a re-search is necessary (step #144). If you want to search again, enter other search conditions and click CP in the same way as above.
A search is performed by U50 (steps #141 to #143).

Then, when the search is completed (No in step #144)
, return to step #99.

In the flow of “indoor/outdoor search” shown in Figure 16, W
The output from the B sensor 14 becomes an index.

When either "indoors" or "outdoors" is selectively input from the keyboard 62 (step #161), the CPU 50 searches for images that match the input search conditions by scanning the search information. (Step #”162).C
In the PU 50, there is a correspondence relationship between the above search conditions and the output of the WB sensor (color temperature data), for example, "indoor" for color temperature data corresponding to fluorescent lighting, and color temperature data corresponding to sunlight. A storage means is provided for storing the location as "outdoors". When searching, one of the input "indoors" and "outdoors" is converted into corresponding color temperature data in the storage means, and the search is performed using the converted color temperature data.

When the search is completed, the number of matching items etc. are displayed as the search results (step #163). After this, it is determined whether a re-search is necessary (step #164). If you want to search again, enter other search conditions and click CP in the same way as above.
A search is performed by U50 (step #161 to step #163>).

Then, when the search is completed (NO in step #164)
, return to step #101.

In the flow of "person ffl scene search" shown in FIG. 17, the information of "focal length f" and "subject distance D" (image magnification β-f-D) serves as an index.

First, when either "person" or rffllllJ is input from the keyboard 62 (step #171), the CPU
50 executes a search for images matching the input search conditions (step #172). This process is performed according to "Search Process ■" shown in FIG. That is, first, all focal lengths f and subject distances recorded in correspondence with images registered in advance in the filing device 61 are read out, and each image magnification β-f-D for all images is calculated (
Step #181). Next, each obtained image magnification β is β
It is determined whether the upper 1/100 is higher (step #182),
If it is 1/100 above β, it is determined that it is a II scene (step #184), and in the opposite case, it is determined that it is a person (step #183). Then, when "person" is input as a search condition, the CPU 50 extracts the result in step #183, and when "landscape" is input as a search condition, the CPU 50 extracts the result in step #183.
184 results are extracted.

Note that as a matching method, discrimination may be performed for each image as described above, and such discrimination may be sequentially repeated.

When the determination for all images is completed, the process returns to step #173.

Returning to FIG. 17, the number of matching items etc. are displayed as the search results (step #173). Thereafter, it is determined whether a re-search is necessary (step #174). If you want to search again, enter the other search condition, and click C in the same way as above.
A search is performed by the PU 50 (steps #171 to #173). When the search is completed (NO in step #174), the process returns to step #103.

In addition, in the above embodiment, the image magnification β, "person", "landscape"
Although the association is stored as a program, a storage means may be provided in the CPU 50 in which the 1 magnification β, "person", and "landscape" are stored in association with each other. In addition, as another search method, the image magnification β is obtained from the recorded focal length f and subject distance, and on the other hand, the image magnification β is input as "person", r! The search may be performed by predetermining a range of image magnifications β for ill, and determining which range includes each image magnification β to be searched.

In the “event search” flow shown in Figure 19,
The location code (event code) becomes the index.

When an event name is input from the keyboard 62 (step #191), the CPU 50 performs a search for images that match the input search conditions using an event code recorded in association with an image registered in advance in the filing device [61]. This is done by scanning (step #192). When the search is completed, the number of matching items etc. are displayed as the search results (step #193). Thereafter, it is determined whether a re-search is necessary (step #194). In the case of re-search, if other search conditions are input, the search by the CPU 50 is performed in the same manner as above (steps #191 to #193).
). Then, when the search is completed (step #194
O), return to step #105.

In addition, if the location code is an identification code for the location or a JIs code string of characters that displays information about the location, it will be possible to search directly using the code, and the content of the location code will be recorded as is when shooting without having to distinguish it. It becomes possible to do so. Furthermore, since no special determining means is required to determine the content of the location code, the configuration of the search side, for example, can be simplified accordingly.

In the flow of "voice search" shown in FIG. 20, the voice recorded in association with the image is compared with the voice input during the search.

When voice is input from the microphone of the voice input unit 64 (step #201), the CPU 50 searches for an image that matches the input voice (step #192). Processing is performed according to "Processing ■". That is, first, the input audio is AD converted, encoded, and then stored in a reference section (not shown) (steps #211 to #213).Next, filing is performed. The audio data recorded in association with the pre-registered image is extracted from the device 61 and encoded (steps #214, #215>. Then, the audio code from the reference section and the audio data from the filing device 61 are extracted and encoded. (Step #216>. This comparison is performed for all images by sequentially scanning the audio codes of the images registered in the filing device 161 (Step #215 to Step #216). NO1 in #217 (loop of step #218). When verification for all images is completed (YES in step #217), the process returns to step #203.

Returning to FIG. 20, the number of matching items is displayed as the search result (step #203). After this, it is determined whether a re-search is necessary (step #204). In the case of re-searching, if another word is input by voice, a search within the CPU 50 is performed in the same manner as above (steps #201 to #203). When the search is completed (NO in step #204), the process returns to step #107.

Next, returning to FIG. 9, the execution process of step #86 will be explained using FIG. 22.

In this execution mode, first, the output format is selected. Table 1 shows this output format.

(Hereinafter, blank space) Table 1 In other words, after the search is completed, one output format is selected from the output formats A to F in Table 1 above, and the keyboard 6
2, for example, when a corresponding character is input (step #221), the number of output images that are the search results is set to N (step #222), and then the images to be output are sent to the filling device in ascending order of frame number. 61 and sequentially transferred to the frame memory 53 (step #22
3). Next, the image is output according to the output format selected in step #221. If output format A is selected (
YES in step #224), then TV monitor 57
will be displayed. If the output type WAB is selected (YES in step #225), data regarding the size of the subject on the imaging surface, that is, the scale size, is determined from the information on the focal length f and subject distance of the image, and the image and The scale size is combined (steps #226, #227>
.

That is, the scale size is superimposed with the image in the composition section 54 and output to the TV monitor 57.

For example, when displaying a unit scale (see Fig. 24 (B)), the scale size marked on the unit scale SC is the length per unit scale from the image magnification β calculated from the focal length f and the subject distance. Here, regarding the calculation of the length of the unit scale SC, the 23rd
This will be specifically explained using the principle diagram in the image formation state shown in the figure.

In the figure, L is the shooting distance, from the subject 100 to the imaging plane 200.
f is the focal length of the photographic lens 300, H is the interval between principal points of the photographic lens 300, X is the distance from the subject 100 to the front focal point of the Yingchang lens 300, and x' is the imaging surface 20.
0 to the rear focus of the photographing lens 3°, y is the length of the object 100, and y' is the distance of the object 1 at the imaging plane 200.
00 image length, and each of the above variables 1. f, H, x,
Between xy and y-, y:y=x:f-f:x-...(1)x=
-L -2f -H-x ----...-(
There is a relationship of 2>. From these equations (1) and (2), x2
- (L-2f-H)x+f2=O=・----(3)
The equation holds true. Finding the roots of this equation (3), we get: X-((L-2f-H)+ (L-2f-H) -4f)/2...
(4) Furthermore, using the above formula (2), x''-((L-2f-H)--2f-H)-4f)/2... (5) , from the above formula (1), y=y−・X/f (
6) V"'V -f/'X-... (7) Therefore, based on x and x' in equations (4) and (5) above, this equation (6) or (7) ), the length y of the object corresponding to the length y- of the image on the imaging surface can be found. Here, the length y' of the image on the imaging surface is expressed as the length of the unit scale SC. When treated as a unit scale, the actual length (on the imaging plane) is known in advance, so the above (6)
The scale size of the unit scale SC is obtained by calculating the equation or (7).

FIG. 24 shows an example of displaying the scale size.

4A is a display example in which the image magnification is displayed as is, and FIG. 2B is a display example in which the unit scale SC corresponds to 10 cm.

By displaying such a scale size, the actual size of the image can be easily grasped. Furthermore, the display timing of the scale SC is controlled by the read clock generator 5 CPt, 150.
9, the scale SC can be moved to a desired position. For example, if the scale SC can be moved to a desired position using a trackball (position specifying member 63), operability will be excellent.

Furthermore, it is easy to turn on and off the scale display, and the display formats shown in FIG.

In addition, by using the technique described in JP-A-58-158507 to imprint the scale onto the silver halide film, it is possible to make the display of the scale size applicable to the silver halide film as well. I can do it.

Next, if output form C is selected (step #228
YES>, the size of the display screen (for example, CRT inches, etc.) is input from the keyboard 62 (step #
229). Subsequently, the image size is changed based on the image magnification determined from the focal length f and the subject distance and the display screen size (step #230). That is, the display magnification (image size) of the size of the image finally displayed on the TV monitor 57 relative to the actual size of the subject is displayed, or the display magnification is fixed to a predetermined size regardless of the display screen size. The size of the displayed image is changed and displayed.

That is, the CPU 50 determines the focal length f and the subject path 111.
It has display magnification calculation means for calculating the display magnification of the reproduced image from the image magnification determined from D and the display screen size,
This display magnification calculation means determines the size of the image displayed on the TV monitor 57 relative to the actual size of the subject. In addition, if the display magnification is set in advance, the ratio between the display magnification calculated by the display magnification calculating means and the set magnification is calculated, and the displayed image is enlarged or reduced according to this ratio. Image processing is performed as necessary.

In the above, the final display magnification is displayed at an appropriate location on the TV monitor 57 as necessary.

By doing so, it is possible to easily grasp the actual size of the image displayed on the TV monitor 57.

Next, if the output format is selected (step #231
(YES), the image desired to be displayed in a composite manner is extracted by the item search shown in FIG. 10 (step #232). Then, a desired image is selected from among the images that match the search conditions (step #233). Subsequently, the image magnification of the first selected image, the image magnification of the second selected image, and both images displayed together are led to the special reproduction processing section 52, where, for example, a specific subject in both images is The image size is appropriately changed so that the image size is the same size, is written to the image frame memory 53, and is output to the TV monitor 57 (step #2
34). By doing so, the sizes of the objects in both images finally displayed together on the TV monitor 57 can be displayed as if they were the same size. Alternatively, if the image sizes of both images that are displayed together are made to match, the actual size of the subject in both images can be easily compared. Furthermore, as in the case of scale display, by moving and superimposing both display panes with a trackball (position specifying member 63) or the like, you can enjoy a composite photograph or the like.

Note that it is also possible to combine the currently taken image and a previously taken photo using the double magnification, so that, for example, a cut-off photo can be reproduced as a completed photo.

Next, if output form E is selected (step #235
(Yes), the displayed image is guided to the special reproduction processing section 52, where various special processing such as mosaic, negative/positive inversion, etc. are applied to the displayed image, and then output to the TV monitor 57 ( Step #236).

Next, if output form F is selected (step #235
(No), a predetermined number of images, such as 4, 9, or 16 images, are reproduced on one screen. The predetermined number of sheets may be selected in advance or as needed, or a desired number such as (2×3) sheets may be selected.

First, the frame number and number of images for which multiple reproduction is desired are stored (step #237).

Next, it is determined whether or not the number of images has reached the predetermined number (step #238). If the number of images is the predetermined number, multi-processing is performed (step #239), and step #240
to move to. On the other hand, if the number of sheets is not the predetermined number, step #2
47. Note that details of the processing in the case of output type HF will be described later.

Now, when one of the output formats A to F is selected and the image obtained under the selection is displayed on the TV monitor 57 (step #240), next, audio is recorded for the displayed image. It is determined whether or not (step #24
1). If audio is recorded, the audio corresponding to the displayed image is played back (step #242>a).Next, it is determined whether or not to print the displayed image and whether to transmit it (step #243). , #245).When a print instruction is issued, the displayed image is printed out by the printer 71, and when a transmission instruction is issued, the FAX 72
The display image is transmitted using the TV'li story 73 or the like (steps #244, #246).

When printing or transmission of one sheet is completed, the number of output images N set in step #221 is decremented by 1 (unless output type BF is selected),
Subsequently, it is determined whether the number N of output images is 0 (step #248).

If N=O, the process returns to step #221, the next image is displayed, and after the displayed image is printed or transmitted,
The bad value is decremented (steps #221 to #247). This process is repeated for the number of output images, and after that, when N=0 (Y in step #248)
ES), except for output form F (step #24
9), the process returns to the main flow shown in FIG.

On the other hand, when output format F is selected, since a predetermined number of images are captured within one screen, processing that is somewhat different from that for output formats A to F is performed. That is,
Each time an image is read out, the number of images is decremented from the output image number N (NO in step #238, step #
247), and each time a predetermined number is reached (Y in step #238).
ES), multi-processing is performed (step #239).

After a predetermined number of multi-processed images are displayed on the TV monitor 57, they are printed or transmitted in units of a predetermined number of images as described above. At this time, by decrementing the number of images when the predetermined number has been reached in step #247 after step #246, the number N of output images can be accurately decremented. In addition,
In the case of output form F, if the number of output images N is not an integral multiple of the predetermined number of images, even if it becomes N-0 in step #248,
Since there are remaining images (No. 1 in step #238)
Step #248r YES, Step #249 NO
), the processing from step #240 onward is also performed on the remaining images.

FIG. 25 is an example of a table of contents display. When the memory card 26 is inserted into the playback device, "Date and time" and "Time" are displayed in the records (or in the order of photographed frame numbers) as shown in the figure.
Contents such as "location", "audio presence/absence", "person's name", etc. are displayed on the TV monitor 57. Based on this table of contents display, it becomes easy to select the method for reproducing and displaying images or recording them in the filing device 61.

FIG. 26 shows an example of a screen for inputting search conditions for a search. The search conditions include "Date and time", "Time", "Place of mouth", "Speaker", "Weather" [Others are prepared, and in each condition column, as shown in the figure, "Summer" is set as "Date" and "Summer" is set as "Time". "Evening" is "Location" and "Flower Expo" is "Location".
"M wave" is input from the keyboard 62 as "Speaker". The CPU 50 executes the above-described search process based on the input search conditions. On the other hand, at the bottom of the screen are various operation instruction fields! For example, the location can be specified using a tablet or mouse.

FIG. 27 shows an example of the memory map of the memory card 26, which consists of a search information area, an image data area, an audio on/off and audio data area. Also,
This search information area contains the start address (Vsta, ASta) and end address (Vsta, ASta) for each of the above areas.
Vend.

Ae nd > is also written. Recording in each area is performed in the order of image data, audio data, and search information according to instructions from the CPU 50.

Next, a case where each of the above-mentioned searches is performed using an inference function (fuzzy search) will be explained with reference to FIGS. 28 to 31.

With such reasoning, searches are performed based on membership functions. This membership function is stored in advance in the suitability storage means in a form corresponding to each search condition. Then, when a search condition is input, a membership function corresponding to the input search condition is selected, and a search is performed in descending order of suitability based on the selected membership function.

Now, Figure 28 shows the length and size) search, for example r50.
This shows a membership function when it is desired to reproduce something on the order of cmJ. “50±5cmJ is a conformity level of “1”
”, and r25cmJ and r75cmJ have a fitness of “0.5
”. Therefore, if you enter the search condition r50cmJ, based on this membership function, the items with suitability “1” will be searched in order of “0.9”, “0.8”, etc.
・Priorities are assigned as follows: first, those with ``50±5cmJ are extracted and reproduced.Next, r44cmJ, r
56cmJ, followed by “4○cmJ, r60cmJ
Nearby ones are extracted in descending order of suitability, such as those near 35cmJ, r65cmJ, etc.
It will be played. In addition, if you enter the search condition ``100cm, J'', r100±5CmJ is set to have a suitability of "1", and r75cmJ is set to have a suitability of "0.5".
”, and as above, those with the highest degree of suitability are sequentially extracted and reproduced.

FIG. 29 shows membership functions when searching for a location, for example, when it is desired to play back a photograph taken in the "Kinki region." Therefore, if you enter "Kinki region",
Based on this membership function, first, the fitness is “1”.
Films shot in Osaka and Kyoto J are extracted and played back.

Next, those taken in "Hyogo" and "Nara" are matched, followed by "Wakayama" and "Shiga", followed by "Mie", "!l! Island", and then "Okayama" and "Fukui". They will be extracted and reproduced in descending order of strength.

Figure 30 shows seasonal searches, such as "spring,""summer," and "autumn."
, which shows each membership function when you want to play back images taken during each season of "winter." for example,"
If you want to play back images taken in ``Spring,'' first select the suitability II.
1 IT photos taken in April and May are extracted and played back. Next, images taken in "June" will be extracted and played back. Also, if you want to play back images taken in "summer", select "July" with a suitability of "1".
, those taken in “August” and “September” are extracted,
will be played. Next, images taken in "June" will be extracted and played back. If you want to play back images taken in "autumn", select "October" and "November" with a suitability of "1".
The images taken are extracted and played back. Next, “9
Images taken on the moon will be extracted and played back. If you want to play back images taken in “winter,” the suitability “
1'' taken in December, January, and February are extracted and played back. Next, images taken in "March" will be extracted and played back.

Note that the seasonal search is not limited to monthly units as described above, but may also be performed by date units. For example, the suitability for "summer" may be set as "0.5" for "mid-September", and the suitability for "autumn" may be set as "0.5".Also, You may also consider other search conditions such as temperature and humidity to perform a search from a broader perspective.For example, if the search condition is "summer + hot", add m to the above "summer". Find the logical product between that and the temperature of 30 degrees Celsius or higher. Similarly, when the search condition is "winter + hot", a logical AND is performed between those that satisfy the above-mentioned "winter J" and those that have a temperature of 25° C. or higher.

Figure 31 shows "morning", "noon", and "evening" in the date and time search.
This shows the membership function when you want to play back images taken in various time periods such as , ``night'', and so on.

If you enter "morning", first "6 o'clock" with a fitness of "1"
The images taken between "9 o'clock" and "9 o'clock" are extracted and played back. Subsequently, images taken in the order of "5 o'clock", "10 o'clock" and [4 o'clock] are extracted and played back. If you enter "noon", first, "12 o'clock" to "
Images taken between 14:00 and 14:00 are extracted and played back. Subsequently, images taken at "11 o'clock" and "15 o'clock" are extracted and played back. When "Evening J" is input, first, images taken at "17:00" and "18:00" with a suitability of "1" are extracted and played back. Subsequently, the image taken at "16:00" is extracted and played back. If you enter "night", first match r! 1″1′.

Images taken between "20:00" and "3:00" are extracted and played back. Subsequently, the images taken at "19:00" and "4FR" are extracted and played back. Note that, as in the case of the seasonal search, the search may be performed from a broader perspective by taking into account search conditions such as temperature and humidity.

In addition to the above search examples, the barometric pressure sensor 19 and GP
It is also possible to search for images taken during mountain climbing from the altitude information provided by the S receiver 23. Furthermore, by adding the focal path 11f and photographing distance to the atmospheric pressure information and altitude information, it is possible to automatically input photographing information as an aerial photograph.

Further, the following can be considered as other determination methods at the time of search. That is, (1) The difference in sound quality between men and women is determined by using the Voice II technology.

(2> Person identification is performed based on the color temperature and pattern recognition of the photographed image of the WB sensor 14.

(3) Adults and children are distinguished from image magnification and pattern recognition.

(4) Determine whether or not glasses are worn by pattern recognition.

(5) The person in the image is memorized in advance, and the person in other images is discriminated using pattern recognition and a learning function.

In this embodiment, a digital memory is used as the recording medium, but an analog memory such as a floppy disk may also be used. Further, although the camera and the playback device are separate bodies, they may be integrated. Furthermore, the data obtained from each sensor may be displayed on the operation and display section 25 or on a separately provided display section.

Furthermore, although this embodiment has been explained using an electronic still camera, each sensor is provided in a silver halide camera, and the photographing information of each of the above sensors is imprinted at an appropriate location within the photographic screen using a well-known date imprinting technique. Good too. In this case, the playback machine includes an optical reading section, a film drive section, a memory, a search processing section according to this embodiment, etc., and the film container is set in the optical reading section and the frames are sequentially or instructed. The photographed image and the imprinted photographic information may be read by COD or the like, extracted as an electrical signal, and recorded in the memory. Then, the same search process as described above is executed using the captured image and the captured information.

〔Effect of the invention〕

As explained above, the camera system according to the present invention includes a camera that records photographed images and audio in association with a recording medium, and a playback device that reproduces the photographed images recorded on the recording medium, a voice registration means for registering voice in the playback device; a collation means for collating the voice registered in the voice registration means with the voice recorded on the recording medium; Since the present invention includes an output control means for outputting for reproduction, it is possible to easily search for an image related to a specific speaker from among the speaker's voice recorded together with the photographed image.

Furthermore, there is no need to manually record audio for each image prior to searching, making it possible to provide an easy-to-use camera system.

[Brief explanation of drawings]

Figure 1 is a block diagram of a still camera, Figure 2 is an explanation of a location code transmitting/receiving device, Figure (A) is a block diagram of a transmitter that transmits a location code, and Figure (B) is a block diagram of a transmitter that transmits a location code. A block diagram of the receiver installed in the camera body that receives the code, Figure 3 is a detailed block diagram of the GPS receiver, Figure 4 is a flowchart explaining the operation of the camera, and Figure 5 is a camera with a search function. A block diagram showing an example of an image playback device, FIG. 6 is a flowchart showing the procedure for registering the image on the memory card in the filing device, and FIGS. 7 and 8 show the procedure for registering the audio on the memory card in the filing device. 9 to 21 are flowcharts showing various search procedures, FIG. 22 is a flowchart showing execution processing such as playback and printing, and FIG.
The figure is a principle diagram in an imaging state to explain how to find the scale size, Figure 24 shows an example of how the scale size is displayed, and (A) is a diagram when the image magnification is displayed as is. Figure (B) is a diagram showing that the unit scale is equivalent to 100 m, Figure 25 is a diagram showing an example of displaying the table of contents, and Figure 2
Figure 6 shows an example of a screen for inputting search conditions during a search, Figure 27 shows an example of a memory map of a memory card, and Figures 28 to 31 show an inference function (fuzzy search) for each search. It is a figure which shows each membership function in the case of performing using. DESCRIPTION OF SYMBOLS 1... Lens, 2... Image pickup element, 5... WB correction circuit, 6... γ correction circuit, 7... Matrix circuit,
8...Encoder, 9...Memory card I/F, 1
0...Microphone, 13...Memory, 14...WB sensor, 15...AF sensor, 16...AFCPU,
17...Temperature sensor, 18...Humidity sensor, 19.
...Atmospheric pressure sensor, 20...CPU, 21...Reception modulation section, 22...Demodulation section, 23...GPS receiver, 2
4...Timekeeping section, 25...Operation and display section, 26...
・Memory card, 201...ROM, 202...D
PSK modulation section, 203...FM modulation section, 206.21
0... Timing generation circuit, 208... FM demodulation section, 209... DPSK demodulation section, 50... CPU, 5
DESCRIPTION OF SYMBOLS 1... Memory card I/F, 52... Special reproduction processing section, 53... Image frame memory, 54... Mixing section, 57... TV monitor, 58... Memory for superimposing , 59... read clock generation circuit, 60.
. . . Character generator, 61 . . . Filing device, 62 .
ROM, 66...Audio data memory, 69...Speaker, 70...Speaker recognition section, SC...Unit scale. Patent applicant Minolta Camera Co., Ltd. Agent
Patent Attorney Etsushi Kotani Patent Attorney
Long 1) Semukai Patent Attorney Takao Ito (A) (B) (A) (B) Figure LL, l

Claims (1)

[Claims] 1. Consisting of a camera having a recording medium that records photographed images and audio in association with each other, and a playback device that plays back the photographed images recorded on the recording medium, and the playback device includes audio registration means that registers the audio. a collation means for collating the voice registered in the voice registration means with the voice recorded on the recording medium; and an output control means for outputting for reproduction an image corresponding to the recorded voice matched by the collation means. A camera system characterized by: