JP2016136398A - Image processing apparatus and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To achieve adequate reading even when a plurality of objects to be read are subjected to batch reading processing.SOLUTION: Fig. 16 shows display content of the whole image displayed together with a final read state (a completion mark etc.) in an overlay state. Namely, the whole image including objects to be read (for example, a bar code) is applied to pattern-analyze, a specific pattern is separately extracted for each of the objects to be read, and then each extracted specific pattern is analyzed to separately recognize information for each of the objects to be read. On the basis of a processing result of above processing, a processing state (the completion mark etc.) up to a current point is separately added to each of the objects to be read which are included in the whole image.SELECTED DRAWING: Figure 16

Description

  The present invention relates to an image processing apparatus and program having an image recognition function.

  In general, for example, when managing entry / exit and inventory of a large number of products stored in a library, etc., a barcode reader is used, and the barcode attached to the product is read to manage the inventory and inventory. Management and so on. At that time, it takes a lot of time and labor to periodically carry out inventory counting work and check one by one while collating the contents of the list with the stored goods. Therefore, in order to perform such confirmation work efficiently, conventionally, bar code is continuously read from a large number of articles stored in a library or the like, and management such as loading and unloading and inventory is performed. An article management system is disclosed (see Patent Document 1).

JP 2000-289810 A

  The above-described prior art compares the article number registered in advance with the barcode (article number) read from the article actually stored in the storage, and registers the newly read article number as movement data. However, when reading a barcode as an object to be read continuously from a large number of stored items, the barcode may be read twice or a reading error may occur. There were problems such as.

  An object of the present invention is to enable appropriate processing to be performed on a target image that is a recognition target included in a captured image.

The invention according to claim 1 is an image processing device having an image recognition function, wherein a target image to be recognized is extracted from a captured image captured by an imaging unit, and image analysis of the target image is performed. A specifying unit that specifies a type of the target image; and a recognition process based on the image recognition function is performed on the target image specified by the specifying unit, and the target is determined according to a result of the recognition process. A determination unit that determines whether or not the image has been normally recognized by the recognition process for the image, and a target image that has been normally recognized by the determination unit, the contents of the result of the recognition process are stored and the specifying unit Storage control means for controlling to store type information indicating the specified type regardless of whether or not the target image specified in step B can be normally recognized in the recognition process. It is characterized by that.
The invention according to claim 2 is an image processing apparatus having an image recognition function, wherein a target image to be recognized is extracted from a captured image captured by an imaging unit, and image analysis of the target image is performed. A specifying unit that specifies a type of the target image; and a recognition process based on the image recognition function is performed on the target image specified by the specifying unit, and the target is determined according to a result of the recognition process. A discriminating unit that discriminates whether or not the image can be normally recognized by the recognition processing for the image, and controls to store type information indicating the type of the target image specified by the specifying unit, and the type of the target specified If the image is a target image that can be normally recognized by the discriminating means, storage control means for controlling to further store the contents of the recognition processing result is provided.
The invention according to claim 6 is an image processing apparatus having an image recognition function, wherein a target image to be recognized is extracted from a captured image captured by an imaging unit, and image analysis of the target image is performed. A specifying unit that specifies a type of the target image; and a recognition process based on the image recognition function is performed on the target image specified by the specifying unit, and the target is determined according to a result of the recognition process. Discriminating means for discriminating whether or not the image has been normally recognized by the recognition processing for the image, and the discriminating means, if there is a target image that could not be normally recognized by the initial discrimination by the discriminating means, The type of the target image is identified from the captured image captured by changing the imaging conditions by the means, and the recognition process for the identified target image of the type is performed again. Again, it is judged whether or not recognized, characterized in that.

  ADVANTAGE OF THE INVENTION According to this invention, an appropriate process can be implement | achieved with respect to the target image used as the recognition target contained in a captured image, and it becomes rich in practicality.

The block diagram which showed the basic component of the information reader. The figure which illustrated the state which displayed as a whole picture the picked-up image which carried out the high resolution imaging | photography of the whole several load piled up by the imaging part. The figure which illustrated each specific pattern extracted as reading object (barcode, a logo, etc.) by carrying out pattern analysis of the whole picture of Drawing 2. The figure which showed the content of the management table memory | storage part M2 after each performing a reading process (recognition process) with respect to each of the reading object (specific pattern) which exists in the whole image shown in FIG. The figure for demonstrating the plane coordinate system which shows the position of each specific pattern in a whole image. The flowchart which showed the recognition process (read process) which reads and recognizes all the reading objects which exist in the whole image collectively. 7 is a flowchart following the operation of FIG. 8 is a flowchart following the operation of FIG. 9 is a flowchart following the operation of FIG. The figure which showed the state which divided | segmented the non-extraction area | region which could not extract reading object (specific pattern) in the whole image into several blocks. The figure which showed the content of the management table memory | storage part M2 after dividing the non-extraction area | region in a whole image into a some block. The figure which illustrated the specific pattern extracted by carrying out pattern analysis of the n times expansion photography picture. The figure which showed the content of the management table memory | storage part M2 after extracting a specific pattern from a n times enlarged picked-up image. The figure which showed the content of the management table memory | storage part M2 after extracting a specific pattern from a nx2 magnification expansion picked-up image. The figure which showed the final process content of the management table memory | storage part M2. The figure which showed the content of the whole image by which the final reading result was overlay-displayed. The figure which showed each whole image which image | photographed sequentially the whole motor vehicle currently drive | working on a highway for every predetermined timing in 2nd Embodiment. The flowchart which showed the reading process (highway monitoring process) in the case of reading a registration number from a license plate in order to monitor the motor vehicle which drive | works a highway as the reading process in 2nd Embodiment. 19 is a flowchart detailing a photographing & reading process (step B4 in FIG. 18).

(Embodiment 1)
The first embodiment of the present invention will be described below with reference to FIGS.
FIG. 1 is a block diagram showing basic components of the information reading apparatus.
The information reading device has an imaging function capable of capturing a high-definition image. For example, the information reading device captures the entire plurality of loads (products) stacked in a warehouse or the like with high resolution, and acquires the captured image as an entire image. Extract a specific pattern (image part such as a barcode) as all the reading objects (for example, one-dimensional barcode, two-dimensional barcode, logo, OCR characters, etc.) existing in the whole image by pattern analysis, By analyzing each of the specific patterns (reading objects), all the reading objects existing in the entire image are read collectively. The information reading device is, for example, a stationary device fixedly installed at a predetermined location in a warehouse or the like so as to photograph the luggage from the front while facing a collection location (storage location such as a warehouse). Type information reading device (package monitoring device).

  The control unit 1 operates by supplying power from a power supply unit (for example, a commercial power supply, a secondary battery, etc.) 2 and controls the overall operation of the stationary information reader according to various programs in the storage unit 3. The control unit 1 is provided with a CPU (Central Processing Unit) and a memory (not shown). The storage unit 3 has, for example, a configuration including a ROM, a flash memory, and the like, and a program storage in which programs for realizing the present embodiment and various applications are stored in accordance with the operation procedures illustrated in FIGS. A management table storage unit M2 that stores a reading result of the unit M1, a barcode, and the like, an image storage unit M3 that stores a captured image, and a dictionary storage unit M4 for information recognition.

  The RAM 4 is a work area for temporarily storing various information necessary for the operation of the stationary information reader such as flag information and screen information. The display unit 5 uses, for example, any one of a high-definition liquid crystal display, an organic EL (Electro Luminescence) display, and an electrophoretic display (electronic paper). The display device is a connected external monitor, but may be provided in the main body of the information reading device. The display unit 5 displays the reading result and the like with high definition. A transparent touch panel that detects the contact of a finger is laminated on the surface of the display unit 5, for example, a capacitive touch screen. A screen (touch screen) is configured.

  The operation unit 6 is an external keyboard that is separated from the main body of the information reading apparatus and connected by code connection or communication, but may be provided in the main body of the information reading apparatus. Although not shown, the operation unit 6 includes, for example, a power key, a numeric key, a character key, and various function keys as various buttons in the form of push buttons. The control unit 1 includes the operation unit 6. For example, various business processes such as inventory management, incoming / outgoing inspection, and incoming / outgoing management are performed as processing in accordance with the input operation signal from.

  The communication unit 7 transmits and receives data via a wide area communication network such as a wireless LAN (local area network) or the Internet, and communicates with an external storage device (not shown) connected via the wide area communication network. Upload and download data at. The image pickup unit 8 constitutes a digital camera capable of high-resolution photography equipped with a high-power zoom of 10 times optical, and is used when reading a barcode or the like attached to each product. The image pickup unit 8 is provided on the main body side of the information reading apparatus and is not shown. In addition to an area image sensor such as a C-MOS or CCD image pickup device, a distance measuring sensor, a light amount sensor, an analog processing circuit, A signal processing circuit, a compression / decompression circuit, and the like are provided, and the optical zoom is adjusted and controlled, and auto-focus drive control, shutter drive control, exposure, white balance, and the like are controlled. In addition, the imaging unit 8 includes a bifocal lens and a zoom lens that can switch between telephoto and wide-angle, and performs telephoto / wide-angle imaging. In addition, the imaging unit 8 is provided with a shooting direction changing function that can freely and manually change the shooting direction (the shooting direction) vertically and horizontally.

FIG. 2 is a diagram exemplifying a state in which a captured image obtained by photographing the whole of a plurality of piled loads with the imaging unit 8 at a high resolution is displayed as an entire image.
The entire image includes an image portion to be read such as a barcode printed or attached on the surface of each package (rectangular area in the figure). A pattern that identifies a set of data existing in the entire image as a region to be read by pattern analysis and extracts the specific patterns (image portions such as barcodes, logos, and OCR characters). The extraction process is executed. The data set portion is an area specified by comprehensively determining the data set density, set area, set shape, etc., and this area is extracted as a specific pattern indicating the image portion to be read. ing.

FIG. 3 is a diagram illustrating each specific pattern extracted as a reading target by performing pattern analysis on the entire image of FIG.
In the figure, the number “100” is an identification number for identifying an entire image including a reading target such as a barcode. The numbers “101” to “116” are specific pattern identification information. That is, when the specific pattern is extracted for each reading target existing in the entire image, identification numbers (series No.) sequentially assigned to identify each of the extracted specific patterns are illustrated. In this case, a state in which a total of 16 specific patterns “101” to “116” are extracted from the entire image is illustrated.

Here, the outline of the reading process in this embodiment will be briefly described.
First, in the present embodiment, an entire image including a reading target such as a barcode is captured and acquired by the imaging unit 8, and then each of the specific patterns extracted as described above is sequentially analyzed. A recognition process (reading process) is performed in which all reading objects existing in the entire image are collectively read and recognized. In this recognition processing, information such as a barcode is read by specifying the type of the reading target and collating with the contents of the recognition dictionary storage unit M4.

  In this case, the processing status up to the present time is added for each reading target based on the processing result for extracting the specific pattern and the processing result for recognizing information. Here, as the processing status, for example, a specific pattern is extracted from the entire image and information is normally recognized (reading completed state: read completed state), or a specific pattern is extracted from the entire image. The processing state for each reading target, such as a state that could not be performed (non-extracted state) or a state in which a specific pattern could be extracted from the entire image but information could not be recognized normally (reading error state) It is.

  And in this embodiment, based on the above-mentioned recognition result, various processes are performed in the following procedures (a) to (f). First, if (a) the specific pattern was extracted but could not be recognized normally, the shooting direction was changed and the aim was adjusted to the specific pattern. A portion where the image is generated (in this case, a reading target corresponding to the specific pattern) is photographed with an n-times (for example, 2 times) zoom. (B) A recognition process is performed on the enlarged image captured with the n-times zoom in this way. (C) As a result, when it cannot be normally recognized, a process of extracting a specific pattern by performing pattern analysis on the enlarged image is performed, and then a recognition process is further performed on the extracted specific pattern. .

  (D) If the enlarged image is not normally recognized even if the enlarged image is analyzed in this way, the zoom magnification is further increased, and the location where the problem occurs is photographed with an n × 2 zoom. (E) And after performing the process which extracts a specific pattern by performing pattern analysis with respect to this enlarged image of this nx2 zoom, the recognition process is further performed with respect to the extracted specific pattern. (F) In this way, if the image cannot be normally recognized even if the image is taken with the n × 2 times zoom, it is determined that the image cannot be read and is left to the user's judgment. A process of saving each photographed enlarged image and each magnified image photographed with n × 2 times zoom is performed.

  Note that the processing procedures (a) to (f) described above are procedures when a specific pattern can be extracted, but recognition processing is not performed normally as a result of performing recognition processing on the specific pattern. However, the present invention is not limited to this. For example, when a specific pattern cannot be extracted in order to cope with a case where printing of a barcode or the like is thin and unclear, the above (b) is excluded. The other processes (a) and (c) to (f) are performed in the same manner. The processing status in such a processing procedure is added (stored) in the management table storage unit M2 in association with each reading target, and the processing status for each reading target is displayed on the displayed whole image. For example, “finished mark or the like is additionally displayed (overlapped display) (see FIG. 16 described later).

FIG. 4 shows the contents of the management table storage unit M2 after executing the reading process (recognition process) for each of the reading targets (specific patterns) existing in the entire image shown in FIG. FIG.
The management table storage unit M2 stores and manages the read information for each read target (specific pattern). “No.”, “Status”, “Upper left coordinate”, “Lower right coordinate”, “Type”, “ Each item includes “reading recognition contents” and “image identification information”. “No.” is an identification number (for example, “101” to “116”) for identifying each of the extracted specific patterns, as shown in FIG. “Status” indicates the processing status of the reading target (specific pattern) up to the present time, and “Done” shown in FIG. 4 indicates that the specific pattern is extracted from the entire image and information can be normally recognized (reading completed). “Status: read completed status”, and “error” indicates a status where a specific pattern can be extracted from the entire image but information cannot be recognized normally (read error status).

“Upper left coordinates” and “Lower right coordinates” are rectangular area designation information for specifying the position and size of the specific pattern (rectangular area) extracted from the entire image, and are two-point coordinates (the upper left coordinates of the rectangular area and The position and size of the area are represented by (lower right coordinates). In this case, if the horizontal direction of the entire image is the X-axis direction and the vertical direction is the Y-axis direction in the planar coordinate system shown in FIG. 5, for example, the pattern region indicated by the identification number “101” has the “upper left coordinate”. It is expressed as (27, 1), and the “lower right coordinate” is expressed as (31, 2). In the pattern region indicated by the identification number “102”, the “upper left coordinate” is represented as (31, 4), and the “lower right coordinate” is represented as (34, 7). Since the actual coordinate value is in units of pixels, the value is, for example, 10 ⁿ times. Here, 10 ⁿ is the number of pixels per 1st grid (one coordinate in the plane coordinate system) shown in FIG.

  “Type” indicates the type of the reading target (specific pattern). In the example shown in the figure, “specific pattern such as logo”, “two-dimensional barcode”, and “one-dimensional code” are stored. The “read recognition content” is information read by the recognition process for each reading target. As described above, the management table storage unit M2 is configured to store the recognition result (reading result) and the processing status up to the present time in association with each reading target existing in the entire image. The “image identification information” is information for identifying the image stored in the image storage unit M3, that is, information for identifying the entire image, the above-mentioned n-fold magnified image, and the n × 2-fold magnified image. , The shooting date and time, the shooting location, the image number, and the like, and the contents of the management table storage unit M2 and the content of the image storage unit M3 are associated (linked) with this “image identification information”. I have to.

  Next, the operation concept of the stationary information reading apparatus in the first embodiment will be described with reference to the flowcharts shown in FIGS. Here, each function described in these flowcharts is stored in the form of a readable program code, and operations according to the program code are sequentially executed. In addition, operations according to the above-described program code transmitted via a transmission medium such as a network can be sequentially executed. The same applies to other embodiments described later. In addition to the recording medium, an operation unique to the present embodiment can be executed using a program / data supplied externally via a transmission medium. .

6 to 9 are flowcharts showing a recognition process (reading process) for collectively reading and recognizing all reading objects existing in the entire image.
First, the control unit 1 activates the imaging unit 8, and when a plurality of whole packages stacked in a warehouse or the like are photographed at high resolution (step A1 in FIG. 6), the entire captured image is captured from the imaging unit 8. The image is acquired as an image, and the “image identification information” is generated and stored in the image storage unit M3 together with the entire image (step A2), and the entire image is displayed on the entire display unit 5 as shown in FIG. Step A3).

  In this state, the control unit 1 performs a pattern extraction process to identify all the reading targets existing in the entire image by pattern analysis of the entire image, and extract the specific patterns, respectively ( Step A4). Then, “No.”, “upper left coordinate”, and “lower right coordinate” are generated for each extracted specific pattern, and are stored in the management table storage unit M2 together with the above-described “image identification information” of the entire image (Step S2). A5). In this case, as shown in FIG. 3, specific patterns No. “101” to “116” are extracted, and “No.”, “upper left coordinate”, and “lower right coordinate” are included as information about the pattern. In addition to being stored in the management table storage unit M2, “image identification information” for identifying the entire image is stored.

  Then, the management table storage unit M2 is referred to specify the “No.” in ascending order (step A6), and the “upper left coordinate” and the “lower right coordinate” corresponding to the specified No. are read out. In addition to specifying the type (one-dimensional barcode, two-dimensional barcode, logo, etc.) by analyzing the image part (specific pattern) specified by the coordinates, the type is collated with the contents of the recognition dictionary storage unit M4. Thus, a recognition process for reading and recognizing information is performed (step A7). As a result, it is checked whether the information has been normally recognized (step A8). When the information has been normally recognized (YES in step A8), the management table storage unit M2 corresponding to the designated number indicates the type of the reading target and the recognition result. Are stored in “type” and “read recognition contents” (step A9). Next, an “done” mark is displayed in an overlay (overlapped display) on the image portion (specific pattern portion) on the entire image corresponding to the designated number (step A10), and the management table storage unit corresponding to the designated number is displayed. By storing “Done” in “Status” in M2 (step A11), it is clearly indicated that the state is normally recognized (read completed state: read completed state).

  If the recognition process is not performed normally as a result of executing the recognition process for the specific pattern corresponding to the designated number (NO in step A8), the image portion (specific pattern portion) on the entire image corresponding to the designated number. "Error" mark is displayed in an overlay (overlapping display) (step A12), and "error" is stored in the "status" in the management table storage unit M2 corresponding to the designated number (step A13). Explicitly indicate that the status was not recognized (read error status). If the type of the reading target can be determined even if the information cannot be normally recognized, the type of the reading target is set to “type” in the management table storage unit M2 corresponding to the designated number. You may make it memorize | store.

  When processing for one specific pattern is completed, it is checked whether all the specific patterns ("No.") have been specified (step A14), and the process returns to step A6 until all "No." are specified. After the next “No.” is designated, the above operation is repeated. As a result, the contents of the management table storage unit M2 are as shown in FIG. 4 at the stage where all “No.” have been specified. Here, when all the “No.” designations have been completed (YES in step A14), the process proceeds to step A15 in FIG. 7, and an area that has not been extracted as a specific pattern (non-extracted area) in the entire image is predetermined. After dividing into multiple blocks with the unit size, the "No.", "Upper left coordinates", and "Lower right coordinates" are generated for each divided block, and the "Status" is set as "Non-extracted" management table The storage unit M2 stores and manages it. In this case, when the non-extracted area is divided into a plurality of blocks with a predetermined unit size, the non-extracted area is divided into units corresponding to other extracted areas, that is, the same as each area (block) already extracted as a specific pattern. The non-extraction area is divided into a plurality of blocks so as to be the size.

FIG. 10 shows a state where the non-extraction area is divided into a plurality of blocks. In this case, as described above, the non-extraction region is divided into a plurality of blocks according to the size and arrangement state so that the size and arrangement state of each specific pattern extracted by the pattern analysis of the entire image are the same. In the figure, identification numbers “120” to “151” are identification numbers for identifying individual blocks newly allocated by dividing the non-extraction area.
FIG. 11 shows the contents of the management table storage unit M2 after the non-extraction area in the entire image is divided into a plurality of blocks. Each block newly assigned this time has its “No.” as an identification number. “120” to “151” are stored, “non-extracted” is stored as the “status”, and coordinate data indicating the position and size is stored as the “upper left coordinate” and the “lower right coordinate”.

  Then, referring to the management table storage unit M2, “No.” is designated in ascending order (step A16), the “status” is read, and “done”, “not extracted”, or “error” is discriminated. (Step A17). Now, No. “101” is designated first, but since “Status” of this designation No. is “Done”, “Type” and “Read recognition contents” corresponding to this designation No. are stored in the management table. The information is read from the storage unit M2 as read information and delivered to a business application (for example, an inventory management application) (step A18). Then, it is checked whether or not all “No.” have been designated (step A19). The process returns to the above-described step A16 until all “No.” is designated, and after specifying the next “No.” The content of “status” is determined, and if it is “completed”, the above operation is repeated.

  If the “status” of the designated number is “error” (step A16), the imaging unit 8 is activated and the orientation of the imaging unit 8 is changed to change the actual reading target corresponding to the designated number. After aiming at the position, n-times (for example, optical double) zoom photographing is performed (step A20). In this case, the “upper left coordinate” and “lower right coordinate” corresponding to the designated number are the positions of the specific pattern on the entire image, and the distance to the baggage (subject) when the entire image is captured and the above-described values. Based on the position of the specific pattern, a change amount of the shooting direction is obtained and the direction is adjusted in the shooting direction. Then, by analyzing the image (enlarged image) taken with the n-fold zoom, the type of the reading target is specified, and recognition processing for reading and recognizing information is performed (step A21).

  As a result, when the recognition is successful (YES in step A22), the recognition result is stored in the “type” and “read recognition content” in the corresponding management table storage unit M2 (step A23). Then, the “completed” mark is displayed in an overlay on the image portion (specific pattern portion) on the entire image corresponding to the designated number, and the “status” in the management table storage unit M2 corresponding to the designated number is set to “error”. To “done” and stored (step A24). Thereafter, the process proceeds to step A18 described above, and the “type” and “read recognition content” corresponding to the designated number are read from the management table storage unit M2 as read information and delivered to the business application.

  If it is assumed that “error” No. “110” is designated as “status” (step A 17), aiming at the actual reading target corresponding to that and shooting again with n-times zoom ( In step A20), a recognition process is performed on the captured image (step A21). In this case, since the two patterns are included in the pattern corresponding to the designated number, the recognition process recognizes normally. Is determined to be impossible (NO in step A22). As a result, the process proceeds to step A27 in FIG. 8, where pattern analysis is performed on the captured image (enlarged image) of the n-fold zoom, and all the reading targets existing in the enlarged image are extracted as specific patterns, respectively. Perform the extraction process.

  FIG. 12 is a diagram exemplifying a specific pattern extracted by pattern analysis of an n-fold magnified captured image. FIG. 13 is a diagram illustrating the contents of the management table storage unit M2 after the specific pattern is extracted from the n times magnified captured image. If the specific pattern cannot be extracted as a result of pattern analysis of the n times magnified photographed image (NO in step A28 in FIG. 8), the flow proceeds to the flow of FIG. 9 described later, but the specific pattern can be extracted. When it is completed (YES in step A28), the "No.", "upper left coordinate", and "lower right coordinate" are generated for each of the extracted specific patterns and stored in the management table storage unit M2 (step A29). .

  Here, in the example of FIG. 12, as a result of performing pattern analysis on the specific pattern indicated by the designation No. “110”, it is found that two one-dimensional barcodes are included, A specific pattern is extracted corresponding to each of the two barcodes. As shown in FIG. 13, the management table storage unit M2 newly assigns identification numbers “163” and “164” corresponding to the two specific patterns and stores them in “No.”. These “upper left coordinates” and “lower right coordinates” are stored. Then, in order to designate one of the specific patterns extracted this time, after specifying No. “163”, recognition processing for the specific pattern is performed (step A30).

  As a result, when the recognition is successful (YES in step A31), the recognition result is stored in the management table storage unit M2 (step A32) or zoomed by n times as in steps A23 and A24 of FIG. The “n times: done” mark indicating that the image has been recognized by shooting is displayed in an overlay, or the “status” is rewritten and stored from “error” to “finished” (step A33). Next, it is checked whether an unspecified pattern exists in the specific pattern extracted this time (step A34). However, since the pattern of No. “164” is not specified (YES in step A34), Returning to step A30 described above, one unspecified No. pattern is designated and recognized, and then the above-described operation is performed.

  Further, when the recognition process is not successful (NO in step A31), an “error” mark is displayed in an overlay on the image portion (specific pattern portion) on the entire image corresponding to the designated number. “Error” is stored in “Status” in the management table storage unit M2 corresponding to No. (step A35). Then, the process proceeds to step A34 for checking the presence / absence of an undesignated number. When all the designations are completed (NO in step A34), the current process (steps A29 to A35) is executed. The presence or absence of “error” is checked (step A36). If there is at least one error (YES in step A36), the flow proceeds to the flow of FIG. 9 described later, but if there is no “error” (step A36). (NO in A36), the process proceeds to step A18 in FIG. 7, and the current read information is delivered to the business application.

  As described above, when the “status” of the designation number “110” is “error”, the same operation as described above is performed. Further, even when “error” No. “114” is designated as “status” (step A17 in FIG. 7), the same operation as described above is performed. In other words, the image is re-photographed with n-times zoom while aiming at the reading object (step A20), and then the recognition process is performed (step A21). NO at step A22), the process proceeds to step A27 in FIG. 8, and a specific pattern is extracted by performing pattern analysis on the captured image (enlarged image) of n times zoom. When a specific pattern can be extracted in this way (YES in step A28), the "No.", "upper left coordinate", and "lower right coordinate" are generated for each extracted pattern, and the management table storage unit M2 Is stored and managed (step A29).

  Here, in the example of FIG. 12, as a result of performing pattern analysis on the specific pattern of the designated number “114”, it is found that three one-dimensional barcodes are included. A specific pattern is extracted corresponding to each barcode. Then, as shown in FIG. 13, the identification numbers “165”, “166”, “167” are newly assigned and stored in “No.” corresponding to these three specific patterns. “Coordinates” and “lower right coordinates” are stored. As a result of the recognition process (step A30), when the specific pattern can be recognized normally (YES in step A31), the “n times: done” mark is displayed in an overlay, or the “status” is changed from “error” to “error”. The read information is stored at the same time as rewriting to “done” (steps A32 and A33). FIG. 13 shows a case where all of the numbers “165”, “166”, and “167” can be recognized normally.

  As described above, in the example of FIGS. 12 and 13, when two more specific patterns are extracted from the specific pattern with the specified number “110”, three more from the specific pattern with the specified number “114”. In the case where a specific pattern is extracted, a state in which recognition for all the extracted patterns is normally performed is shown, but one of them (for example, No. “164”, No. “165”) is shown. ) However, if it cannot be recognized normally, the flow proceeds to the flow of FIG. The flow in FIG. 9 is executed when the specific pattern cannot be extracted from the enlarged image when the “status” of the designation number is “error” as described above (NO in step A28). Even if the specific pattern can be extracted, it is executed when the recognition result is “error” (YES in step A36).

  First, in the flow of FIG. 9, after aiming at the reading object corresponding to the designated No. and taking an image with a zoom of n × 2 times (step A37), the following basically describes the above-described FIG. The same operations as in steps A27 to A36 are performed (steps A38 to A47). Here, the difference from Steps A27 to A36 in FIG. 8 is that when the recognition process is successful and the recognition is normal (YES in Step A42), it indicates that the recognition was possible with n × 2 × zoom photography. n × 2: The “complete” mark is displayed in an overlay (step A44). If the mark is not normally recognized (NO in step A42), the “NG” mark indicating that processing is impossible is displayed in an overlay, and the “ The “status” is rewritten from “error” to “NG” indicating that processing is impossible (step A 46), and instead of determining the presence or absence of “error”, the presence or absence of “NG” is determined (step A 47).

  In addition, if the specific pattern cannot be extracted even if the pattern analysis is performed on the enlarged image photographed with the n × 2 zoom in this way (NO in step A39), the process proceeds to step A19 in FIG. It is determined whether or not the “No.” has been specified. If “NG” is not present (NO in step A47), the process proceeds to step A18 in FIG. 7, and the read information is delivered to the business application. If “NG” is present (YES in step A47), the process proceeds to the next step A48 to generate the “image identification information” corresponding to the above-mentioned n-fold magnified image and n × 2-fold magnified image, Each enlarged image is stored in the image storage unit M3 together with the “image identification information”, and the generated “image identification information” is stored in the management table in association with the “No.” of the reading target that is “NG”. The reading target that has become “NG” by being stored in the part M2 and each enlarged image are associated (linked). Thereafter, the process proceeds to step A18 in FIG. 7, and the read information corresponding to the normal reading is transferred to the business application except for NG.

  On the other hand, if the “status” of the designated number is “non-extracted” (step A16 in FIG. 7), the imaging unit 8 is activated and the orientation of the imaging unit 8 is changed to actually correspond to the designated number. After aiming at the position of the reading object, n times (for example, optical 2 times) zoom photographing is performed (step A25). Thereafter, the flow proceeds to the flow of FIG. 8, and the above-described operation is repeated thereafter. In this case, the non-extracted block numbers “120”, “121”, and “123” shown in FIG. 10 are cases where printing is thin and the pattern cannot be extracted by the first pattern analysis. As shown in FIG. 12, by performing pattern analysis on the captured image at the n times zoom of the non-extracted block No. “120”, the specific pattern (No. “160”) and the non-extracted block No. “121” are obtained. The specific pattern (No. “161”) from “” and the specific pattern (No. “162”) from the non-extracted block No. “123” can be extracted.

  In this case, as a result of analyzing the specific patterns of No. “160” and No. “162”, information can be normally recognized from the specific patterns. That is, the case where the “logo” can be recognized from the specific pattern of No. “160” and the case where the “OCR character” can be recognized from the specific pattern of No. “162” Since the specific pattern of No. “161” includes three one-dimensional barcodes, even if the specific pattern was analyzed, information could not be normally recognized from the specific pattern. Is the case.

  In this way, even if the image is taken with the n-times zoom, if it cannot be recognized normally (NO in step A31 in FIG. 8), the “status” becomes “error” (step A35). As shown in FIG. 12, the flow of step 9 is followed by pattern analysis on the enlarged image taken with the n × 2 zoom (step A39). As shown in FIG. 12, three one-dimensional barcodes are included. If it is found, specific patterns are extracted corresponding to these three barcodes. FIG. 14 is a diagram showing the contents of the management table storage unit M2 after extracting a specific pattern by pattern analysis of an n × 2 magnified photographed image. In the illustrated example, these three specific patterns are shown. In addition to newly assigning the identification numbers “168”, “169”, “170” and storing them in “No.”, their “upper left coordinates” and “lower right coordinates” are stored (steps). A30). Then, the recognition process is executed corresponding to No. “168”, “169”, “170”. In the example of FIG. 14, No. “168” and “170” were normally recognized, but No. “169” was not normally recognized.

  Thereafter, when all the “No.” designations are completed by repeating the above operation, the final processing content of the management table storage unit M2 is as shown in FIG. 15, and only the “status” of No. “169” is displayed. Becomes “NG” and others are “completed”. FIG. 16 shows the display contents of the entire image in which the final reading result is displayed in an overlay manner. Thus, when all “No.” designations have been completed, this is detected at step A19 in FIG. 7, and the process proceeds to step A26, where the entire image with the completed mark or NG mark displayed as an overlay is displayed. Then, after the entire image in the final state is stored in the image storage unit M3, the flow in FIGS. 6 to 9 ends.

  As described above, in the embodiment described above, the control unit 1 performs a pattern analysis on the entire image including a reading target (for example, a barcode) and extracts a specific pattern for each reading target, and extracts each extracted pattern. A process of recognizing information (for example, barcode information) for each reading object by analyzing a specific pattern is performed, and up to the current time for each reading object included in the entire image based on one of the processing results Since each processing status is added, even if multiple reading objects are read at once, double reading and reading omission can be prevented, and appropriate reading can be realized, which is highly practical. It will be a thing.

  Since the processing status up to the present time is identified and displayed on the image portion corresponding to each reading target on the entire image, the user can grasp the current processing status. In this case, since the whole image is displayed even during the reading operation, the current processing status can be grasped in real time.

  Since the entire image in a state where the processing status is identified and displayed is stored, the user can freely grasp the processing status at any time.

  Since the “status” indicating the current processing status is stored in the management table storage unit M2, for example, the results of reading processing can be aggregated or output as a report for each processing status.

  Since the entire image including a plurality of reading objects is captured and acquired, the entire image can be easily obtained on the spot.

  When the specific pattern cannot be extracted or when the information cannot be recognized normally, the portion is magnified at a predetermined magnification (n times), and then the magnified magnified image is captured. Since specific pattern extraction processing and recognition processing are performed, for example, even when printing such as bar coating is thin and unclear, or when multiple reading objects are included, normal processing is performed by reprocessing after enlarged shooting It is possible to increase the possibility that it can be recognized.

  When the specific pattern cannot be extracted or when the information cannot be recognized normally, the portion is enlarged at a predetermined magnification (n times), and then recognition processing is performed on the enlarged image. Since it did in this way, even when printing, such as a bar coat, is too small, it becomes possible to raise the possibility that it can recognize normally by reprocessing after expansion photography.

  The non-extracted area where the specific pattern could not be extracted is divided for each block of a specific size, and the portion corresponding to the block is magnified at a predetermined magnification for each block, and then this magnified magnified image is captured. Since specific pattern extraction processing and recognition processing are performed on the image, for example, after magnified shooting even in areas where the specific pattern could not be extracted due to thin and unclear printing such as bar coating It is possible to increase the possibility of normal recognition by reprocessing.

  When a non-extracted area where a specific pattern could not be extracted from the entire image was divided into blocks of a predetermined size, it was divided based on the size of the extracted specific pattern. Since there is a possibility that a specific pattern with the same form as the specific pattern in the non-extracted area may exist, the possibility of extraction is increased by dividing the block based on the size of the extracted specific pattern. Can do.

  If re-processing after magnified shooting at a predetermined magnification (n times) was not recognized normally, the portion was magnified again at a magnification (n × 2 times) higher than the predetermined magnification (n times). Thereafter, since the extraction process and the recognition process of the specific pattern are performed, the possibility of normal recognition can be further increased.

  Since an enlarged image that has been magnified and photographed at a predetermined magnification (n times) and an enlarged image that has been magnified and photographed at a magnification (n × 2 times) higher than the predetermined magnification are stored, the enlarged image should be stored for the user. You can search for the cause of the failure to recognize normally.

  In the above-described embodiment, if the recognition process can be normally recognized, the completed mark is displayed as an overlay. However, before the reading process is started, the entire area is displayed in the unprocessed area. In order to show that there is, for example, a light gray shading is displayed over the entire area of the entire image, and if it can be recognized normally, the overlay display displayed at the recognition position is deleted. By doing so, the completion of normal recognition may be identified and displayed on the entire image. Even in this case, the same effect as the above-described embodiment can be obtained, and the processing status can be clarified. Further, the display of the processing status is arbitrary, for example, a figure with “x” in place of the completed mark is displayed in an overlapping manner.

  In the above-described embodiment, the one-dimensional barcode, the two-dimensional barcode, the logo, and the OCR character are exemplified as the reading target. However, as the reading target, a printed character, a handwritten character, a mark sheet, an image (for example, a packing box, Book, face, etc.).

  The information reading apparatus according to the above-described embodiment has an imaging function capable of capturing a high-definition image, so that the entire plurality of loads stacked in a warehouse or the like are captured with high resolution and the captured image is acquired as an entire image. However, the entire image may be acquired in advance from the outside via a communication means or an external recording medium.

  The information reading device of the above-described embodiment is a stationary information reading device that is fixedly installed at a predetermined location to photograph a baggage from the front while facing the baggage accumulation location. May be a handy terminal, OCR (optical character reader), or the like.

(Second Embodiment)
A second embodiment of the present invention will be described below with reference to FIGS.
In the above-described first embodiment, barcodes, logos, and the like as reading targets included in a captured image (entire image) obtained by capturing the entire plurality of loads stacked in a warehouse or the like are collectively displayed. Although read, in this second embodiment, the vehicle traveling on the highway is monitored, and a captured image obtained by sequentially capturing the entire traveling vehicle at every predetermined timing is acquired as an entire image. In addition, the registration numbers are collectively read from the license plate of the automobile as the reading target included in each entire image at each predetermined timing. Here, in both the embodiments, the same or the same names are denoted by the same reference numerals, the description thereof will be omitted, and the following description will focus on the features of the second embodiment. .

  The information reading device of the second embodiment is a stationary information reading device that is fixedly installed so as to be able to take a picture of the entire automobile coming from the upper side of one lane on the highway. And this information reading apparatus carries out high-resolution imaging of one side all lanes, acquires this picked-up image as a whole image, and sets it as all the reading objects (number plate of a car) which exist in this whole image. A specific pattern is extracted by pattern analysis, and each of the specific patterns (reading targets) is analyzed to read the registration numbers from all the reading targets existing in the entire image.

FIG. 17 is a diagram showing the entire images obtained by sequentially capturing the entire automobile traveling on the highway at predetermined timings.
17 (1) shows the entire image taken at 09: 37: 46.85, FIG. 17 (2) shows the entire image taken 0.5 seconds later, and FIG. Further, the whole image taken 0.5 seconds later is shown. In this case, FIG. 17A shows a state in which the registration number is read from the license plates of three automobiles, and the entire captured image is stored in the image storage unit M3. The read registration number is the management table storage unit M2. Is remembered.

  FIG. 17 (2) shows a state where the registration number is read from two newly appearing cars, and the registration number is read from one previously read car, and the entire image is displayed in the image storage unit M3. The two newly registered registration numbers are stored in the management table storage unit M2. However, in order to avoid the redundant storage of the registration numbers that are read twice during the previous and current readings, This time, the registration number stored in the management table storage unit M2 is deleted. FIG. 17 (3) is a state in which the registration number is read from two newly appearing cars, and the registration number is read from one previously read car. In this case as well, The entire image is stored in the image storage unit M3, and the two newly registered registration numbers are stored in the management table storage unit M2. In this case, in order to avoid redundant storage of the registration numbers that are read at the time of the previous and current readings, the registration numbers stored in the management table storage unit M2 this time are deleted.

FIG. 18 is a flowchart showing a reading process (highway monitoring process) when reading a registration number from a license plate to monitor a car traveling on a highway as a reading process in the second embodiment. Execution starts accordingly.
First, in response to power-on, the control unit 1 starts this reading process (highway monitoring process), and a captured image (through image) in which all the lanes on one side of the highway are photographed at high resolution by the imaging unit 8. ) Is acquired for monitoring (step B1). And it will be in a stand-by state until a fixed time (for example, 0.5 second) passes (step B2), and when a fixed time passes (it is YES at step B2), the picked-up image will be analyzed and something will be shown in a picked-up image. Whether a moving object exists (is captured) is checked (step B3). When a moving object is captured in the captured image (YES in step B3), a photographing and reading process is performed (step B4). ).

FIG. 19 is a flowchart detailing the photographing & reading process (step B4 in FIG. 18).
First, the control unit 1 acquires, as an entire image, a captured image obtained by photographing the entire lane on one side of the highway with high resolution from the upper side by the imaging unit 8 (step C1), generates the “image identification information”, and generates the entire image. In addition to storing the entire image in the image storage unit M3, the entire image is displayed on the monitor 5 (step C2), and a specific pattern as all reading objects (number plates) existing in the entire image is obtained by pattern analysis. Extract (step C3). Then, the extracted objects to be read (number plates) are sequentially aimed, and each of the number plates is magnified by n times (for example, 10 times) zoom (step C4). For example, in the case of FIG. 17 (1), the registration numbers “A 12-34”, “I 56-78”, and “U 90-12” are respectively enlarged.

  Next, recognition processing (reading processing) for reading and recognizing the registration number from the specific pattern by sequentially designating and analyzing individual specific patterns (number plates) extracted from the entire image (step C5). As a result of designating and reading the license plate with the registration number “A12-34” and when it is normally recognized (YES in step C6), the corresponding image on the entire image corresponding to the designated plate is displayed. “No.”, “Status”, “Type”, “Reading recognition contents”, “Image identification information” are displayed as the reading information for the designated plate while displaying the “finished” mark on the image portion of the number (step C7). Is generated and stored in the management table storage unit M2 (step C8).

  Here, the “image identification number” stores the “image identification information” of the entire image described above, and associates the entire image in the image storage unit M3 with the read information in the management table storage unit M2. (To be tied). In “Status”, “completed” indicating that the state has been recognized normally (reading completed state: read completed state) is stored, and in “type”, “location recognition content” such as place name and vehicle type is stored. "Stores a registration number.

  When the recognition process for one reading object (number plate) is completed in this way, it is checked whether the recognition process for all the reading objects is completed (step C9), and the process proceeds to the above-described step C5 until all the processes are completed. No. plate, for example, “56-78” is designated and the recognition process is performed. As a result of reading the designated number “56-78” as a result of the reading process, if it cannot be recognized normally (YES in step C6), the magnified image is taken with the n-times zoom for each license plate as described above. (Step C10), a recognition process for reading and recognizing the registration number by analyzing the enlarged image is performed (step C11), and it is checked whether the recognition has been normally performed (step C12).

  Here, when the result of analyzing the magnified shooting is normal recognition (YES in step C12), the process proceeds to step C7 described above, and the “finished” mark is displayed in an overlay on the image portion of the number on the entire image. The read information for the designated plate is stored in the management table storage unit M2 (step C8). Further, when the magnified image is taken with the n-fold zoom and cannot be recognized normally (NO in step C12), the magnified image that has been magnified with the n-fold zoom is stored in the image storage unit M3 together with the “image identification information”. In addition, an “NG” mark indicating that reading is impossible is displayed in an overlay on the image portion of the number on the entire image (step C14).

  Then, “No.”, “status”, and “image identification number” are generated for the designated plate and stored in the management table storage unit M2. In this case, “status” indicates “unreadable” NG "is stored (step C15). Thereafter, the process proceeds to the above-described step C9, where it is checked whether or not the recognition process for all the license plates has been completed, and the process proceeds to the above-described step C5 until all the processes are completed. As a result, in the case of the entire image shown in FIG. 17A, the registration numbers of the three automobiles are normally read and stored in the management table storage unit M2.

  When the reading process for one whole image is finished in this way (step B4 in FIG. 18), the contents of the management table storage unit M2 read this time and the management table read within the past predetermined time (for example, within the past one minute). The contents of the storage unit M2 are compared, and it is checked whether the same registration number is stored (step B5). Now, in the example of FIG. 17 (1), it is the first reading start after power-on, and since it is determined that the same number is not stored (NO in step B5), the end of monitoring is not instructed. Under these conditions (NO in step B7), the process proceeds to the above-described step B2, and the above-described operation is repeated thereafter. Note that the end of monitoring is instructed by a user operation or after a predetermined time has elapsed.

  Here, when the entire image shown in FIG. 17 (2) is taken and read, the newly registered two registration numbers “E34-56” and “O79-90” Although stored in the management table storage unit M2, since one registration number “U90-12” is the same as the registration number stored last time, the duplicate storage is excluded by deleting the registration number stored this time (Steps B5 and B6). Similarly, when the entire image shown in FIG. 17 (3) is captured and read at the next timing, two newly appearing registration numbers “KA 9-87”, “KI 65- 43 ”is stored in the management table storage unit M2, but the duplicate storage of the same registration number“ E34-56 ”as the previous registration stored at the timing of FIG. 17 (2) is excluded (step B5). , B6).

  As described above, in the second embodiment, when the entire image captured at every predetermined timing is sequentially acquired and the same information is included in the information for which the reading target is recognized for each entire image. Since duplicate storage of the same information is suppressed, even when all reading targets are read collectively from each whole image taken at a predetermined timing, it is possible to effectively prevent duplicate storage of the same reading target. Can be read appropriately.

  In the second embodiment described above, the reading process in the case of reading the registration number from the license plate in order to monitor the automobile traveling on the highway is shown. However, the work process in which the product is completed by the flow work is monitored. Therefore, a process of reading a product number, a logo printing state, or the like may be used. In the second embodiment described above, 0.5 seconds is shown as the predetermined timing. However, this value is arbitrary, such as 0.5 seconds, 1 second, 0.5 seconds, 1 second,. There may be.

  Also, in the second embodiment described above, the stationary information reading device fixedly installed is shown, but a portable information reading device or the like may be used. In this case, even if an operator takes images at predetermined timing while moving in a luggage storage area and the like, the same portion to be read can be prevented from being duplicated even if the same portion is taken. For this reason, when the operator sequentially photographs while moving in the luggage storage area or the like, it is not necessary to strictly determine the photographing location, and the entire work can be efficiently performed.

  In addition, the information reading apparatus shown in the above-described embodiment may be separated into a plurality of cases according to function, and is not limited to a single case. In addition, each step described in the above-described flowchart is not limited to time-series processing, and a plurality of steps may be processed in parallel or separately.

DESCRIPTION OF SYMBOLS 1 Control part 3 Memory | storage part 5 Display part 8 Imaging part M1 Program memory | storage part M2 Management table memory | storage part M3 Image memory | storage part M4 Dictionary memory | storage part for information recognition

The invention according to claim 1 is an image processing apparatus having an image recognition function , wherein the extraction unit extracts a specific pattern from an area to be recognized in a captured image captured by the imaging unit; wherein in the area obtained by extracting a specific pattern, the type of recognition of the specific pattern, and recognition of information of the specific pattern, and recognition means for performing, for areas obtained by extracting the specific pattern by the extraction means, the recognition characterized by comprising a control unit that controls to store the recognition result of the type of the specific pattern or information of the specific pattern, specified by the means.
The invention according to claim 7 is an image processing apparatus having an image recognition function , wherein an extraction unit that extracts a specific pattern from an area to be recognized in a captured image captured by the imaging unit, and the extraction unit In the area where the specific pattern is extracted , a recognition means for recognizing the type of the specific pattern and recognition of the information of the specific pattern, and when the recognition means can recognize the recognition of the information, Control is performed to end recognition processing for the recognized area. On the other hand, when the information cannot be recognized by the recognition means and the type of the specific pattern can be specified, the specified type for the area is set. and control means for line Migihitsuji controls a recognition process again according, characterized by comprising a.

Claims (13)

An image processing apparatus having an image recognition function,
A specifying unit that extracts a target image to be recognized from a captured image captured by the imaging unit and performs image analysis of the target image to specify the type of the target image;
Whether or not recognition processing based on the image recognition function is performed on the target image whose type is specified by the specifying unit, and whether or not the target image is normally recognized according to a result of the recognition processing Discriminating means for discriminating;
For the target image that can be normally recognized by the discriminating means, the contents of the recognition processing result are stored, and the type of the target image specified by the specifying means is normally recognized by the recognition processing. Storage control means for controlling to store type information indicating the specified type regardless of whether or not it is possible;
An image processing apparatus comprising: An image processing apparatus having an image recognition function,
A specifying unit that extracts a target image to be recognized from a captured image captured by the imaging unit and performs image analysis of the target image to specify the type of the target image;
Whether or not recognition processing based on the image recognition function is performed on the target image whose type is specified by the specifying unit, and whether or not the target image is normally recognized according to a result of the recognition processing Discriminating means for discriminating;
Control is performed to store type information indicating the type of the target image specified by the specifying unit, and if the target image specified by the type is a target image that can be normally recognized by the determination unit, the recognition is performed. Storage control means for controlling to further store the contents of the processing results;
An image processing apparatus comprising: The storage control means controls to store status information indicating whether or not the discrimination means has been normally recognized in association with the target image specified by the specifying means;
The image processing apparatus according to claim 1, wherein the image processing apparatus is an image processing apparatus. Output control means for controlling to output the status information for the target image stored in the storage control means;
The image processing apparatus according to claim 1, further comprising: For the target image that could not be recognized normally by the determination unit, the recognition process for the target image is performed again from the captured image captured by changing the imaging condition of the imaging unit, and the second recognition process Re-determining whether it was recognized normally,
The image processing apparatus according to claim 1, wherein the image processing apparatus is an image processing apparatus. An image processing apparatus having an image recognition function,
A specifying unit that extracts a target image to be recognized from a captured image captured by the imaging unit and performs image analysis of the target image to specify the type of the target image;
Whether or not recognition processing based on the image recognition function is performed on the target image whose type is specified by the specifying unit, and whether or not the target image is normally recognized according to a result of the recognition processing Discriminating means for discriminating;
Comprising
If there is a target image that could not be normally recognized in the first determination by the determination unit, the determination unit changes the image pickup condition by the image pickup unit, specifies the type of the target image, and determines the type of the target image. Re-recognize the target image of the type specified, and again determine whether or not the re-recognition process was successful.
An image processing apparatus. For the target image that can be normally recognized by the determination means, an output means for outputting the result of the recognition processing;
The image processing apparatus according to claim 6, further comprising: For the target image that can be normally recognized by the discriminating means, the contents of the recognition processing result are stored, and the type of the target image specified by the specifying means is normally recognized by the recognition processing. Storage control means for controlling to store type information indicating the specified type regardless of whether or not it is possible;
The image processing apparatus according to claim 6, further comprising: The specifying means analyzes an image analysis based on a comparison with a plurality of types of pattern images prepared in advance to analyze which type of pattern image corresponds to the type of the target image. Identify,
The image processing apparatus according to claim 1, wherein the image processing apparatus is an image processing apparatus. When there are a plurality of target images extracted from the captured image, the specifying unit specifies a type for each of the plurality of target images,
The determining means determines whether or not the recognition process for each of the plurality of target images has been normally recognized;
The image processing apparatus according to claim 1, wherein the image processing apparatus is an image processing apparatus. A program for controlling a computer of an image processing apparatus having an image recognition function,
The computer,
A specifying unit that extracts a target image to be recognized from a captured image captured by the imaging unit and identifies the type of the target image by performing image analysis of the target image;
Whether or not recognition processing based on the image recognition function is performed on the target image whose type is specified by the specifying unit, and whether or not the target image is normally recognized according to a result of the recognition processing Discriminating means for discriminating,
For the target image that can be normally recognized by the discriminating means, the contents of the recognition processing result are stored, and the type of the target image specified by the specifying means is normally recognized by the recognition processing. Storage control means for controlling to store type information indicating the specified type regardless of whether or not it is possible;
A program designed to function as A program for controlling a computer of an image processing apparatus having an image recognition function,
The computer,
A specifying unit that extracts a target image to be recognized from a captured image captured by the imaging unit and identifies the type of the target image by performing image analysis of the target image;
Whether or not recognition processing based on the image recognition function is performed on the target image whose type is specified by the specifying unit, and whether or not the target image is normally recognized according to a result of the recognition processing Discriminating means for discriminating,
Control is performed to store type information indicating the type of the target image specified by the specifying unit, and if the target image specified by the type is a target image that can be normally recognized by the determination unit, the recognition is performed. Storage control means for controlling to further store the contents of processing results;
A program designed to function as A program for controlling a computer of an image processing apparatus having an image recognition function,
The computer,
A specifying unit that extracts a target image to be recognized from a captured image captured by the imaging unit and identifies the type of the target image by performing image analysis of the target image;
Whether or not recognition processing based on the image recognition function is performed on the target image whose type is specified by the specifying unit, and whether or not the target image is normally recognized according to a result of the recognition processing Discriminating means for discriminating,
Function as
If there is a target image that could not be normally recognized in the first determination by the determination unit, the determination unit changes the image pickup condition by the image pickup unit, specifies the type of the target image, and determines the type of the target image. Re-recognize the target image of the type specified, and again determine whether or not the re-recognition process was successful.
Program.

Images