Classifications

• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
  • G06T1/00—General purpose image data processing
  • G06T1/0021—Image watermarking
  • G06T1/005—Robust watermarking, e.g. average attack or collusion attack resistant
  • G06T1/0057—Compression invariant watermarking
• • G—PHYSICS
  • G06—COMPUTING OR CALCULATING; COUNTING
  • G06F—ELECTRIC DIGITAL DATA PROCESSING
  • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
  • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
  • G06F16/51—Indexing; Data structures therefor; Storage structures

Definitions

• the invention relates to a method for quickly storing image data, in particular to a method for quickly storing image data capable of quickly compressing and storing image data, and a method for identifying valuable documents using the image data fast storage method for storing value document image data and identifying Device. Background technique
• the ordinary deposit and withdrawal machine processes 8 price documents per second, each of which has an average processing time of 125ms, a system overhead of about 20ms, and a serial communication of 15ms.
• the value file identification (including category, denomination, orientation, authenticity, serial number identification) is about 65ms, so the image storage time must be controlled within 15ms.
• the image storage time must be controlled within 15ms.
• To store complete image information at least three sample images need to be stored.
• one is a white light image with a resolution of 480*800; one is an infrared image with a resolution of 240*400; Transmitted image with a resolution of 240*400.
• the white light image only uses the 480*800 resolution for the serial number area.
• the actual resolution is 240*400, so you can choose to store all three images in the same way. You can also convert white images into 240*400 resolution images and store three 240*400 resolution images.
• the time-consuming situation is as follows:
• the traditional image storage method can not meet the requirement of storing three high-resolution images in 15ms, and even cause the system to crash. Therefore, if the image needs to be stored, the device recognition speed can be reduced or the device recognition function can be reduced.
• Another object of the present invention is to provide a value document identification method that can quickly store the collected document image data.
• the present invention provides a method for quickly storing image data, which is used for compressing and storing value document image data collected in a document of value document, comprising the following steps: (A1) obtaining long image data; (A2) extracting the long integer image data by using a bit-and-bit operation using N mutually corresponding data masks, the extraction method is to extract M points in each row of M*N points, and extract one point in each column of L points.
• N is an integer greater than or equal to 2
• L and M are integers greater than or equal to 1
• (A3) Data extracted by integrating N data masks by a bitwise OR operation to obtain encoded image data and stored.
• the image data collected directly by the value document identification device is single-byte image data, so The single-byte image data collected by the image is forced to be converted into long image data before storage.
• the step (A2) further includes: dividing the long image data into N partitions, and extracting the N partitions by a bit-and-operation operation using N mutually corresponding data masks Long integer image data.
• the present invention also provides a value document identification method for identifying a current value document, comprising the steps of: (B1) collecting a sample image of a current value document and obtaining Corresponding image data; (B2) performing value document identification according to image data of the current value document to obtain value document identification data; (B3) forcing the image data collected by the image into long image data; (B4 And compressing and storing the converted image data using the image data fast storage method.
• step (B1) it is detected whether the current value document enters the sample area, "Yes” then collects the image data, and "No” continues the detection.
• the invention enables automatic detection without manual manipulation.
• the sample image includes an infrared sample image, a transmission sample image, and a white light sample image
• the step (B4) specifically includes: directly storing the long image of the infrared sample image and the transmission sample image Data and storing the long image data of the white light image using the image data fast storage method.
• infrared image, transmission image, and white light image are often required.
• the image data is converted into a long image. After the image data, it is possible to store its image data at a faster speed.
• the value document identification method of the present invention further comprises the following steps: (B5) restoring the compressed stored image data to obtain decompressed image data; (B6) determining the authenticity of the current value document according to the image data of the current value file, To obtain further valuable document identification data.
• step 5 restoring the compressed stored image data to obtain decompressed image data
• step 6 determining the authenticity of the current value document according to the image data of the current value file, To obtain further valuable document identification data.
• the value document identification data includes current The type of the price document, the denomination, the orientation, the serial number, the authenticity, etc., wherein the serial number information of the current value document needs to be identified according to the image data of the white light sample image collected, and the authenticity information of the current value document And other current value document identification data can be analyzed using the stored image data, so the present invention also needs to decompress the compressed stored image data.
• the step (B5) is specifically: calculating data corresponding to each pixel on the decompressed image template according to the size data of the decompressed image template and the compressed stored image data to obtain decompressed image data.
• the present invention also provides a corresponding value document identification device for identifying a current value document, which includes a collection module, an identification module, a compression module and a storage module, and the collection module collects a sample image of the current value document and Obtaining corresponding image data; the identification module performs value file identification according to the image data of the current value file to obtain the value file identification data; the compression module includes a conversion unit and a coding unit, and the conversion unit forcibly converts the collected image data into Long integer image data; the coding unit extracts the long image data by a bit-and-bit operation using N mutually corresponding data masks, and then integrates data extracted by N data masks by "bit or" operation to The encoded image data is obtained and stored.
• the extraction method is to extract M points in each row of M*N points, extract one point in each column L point, N is an integer greater than or equal to 2, and L and M are integers greater than or equal to 1. ;
• the storage unit is used to store data.
• the coding unit divides the long image data into N partitions, and extracts long integer images in N partitions by using a bit-and-operation operation using N mutually corresponding data masks. data.
• the value document identification device further comprises a decompression module for restoring the encoded image data to obtain decompressed image data.
• the collection module is a contact image sensor, and the contact image sensor can detect whether the current value document enters the sample area, and collect image data after the current value document enters the sample area, "No" Then continue testing.
• the invention enables automatic detection without manual operation.
• the image data fast storage method of the present invention extracts multi-point data by using a plurality of "bit-and-bit” operations at a time by using a plurality of corresponding data masks, and extracts the data by a "bit or" operation.
• the data is re-encoded and stored to obtain encoded image data, which greatly reduces the number of operations and realizes fast compression storage of image data.
• it also improves the use of the above image data to save quickly.
• the value document identification method and the value document identification device of the present invention also forcibly convert the collected single-byte image data into long integer image data, replacing the traditional single-byte storage method, and reducing the operation. The number of times further increases the storage speed of the image data.
• FIG. 1 is a flow chart of a method for identifying a value document of the present invention
• FIG. 2 is a flow chart of a method for quickly storing image data according to the present invention
• Figure 3 is a block diagram showing the structure of the value document identifying apparatus of the present invention.
• 4a is a schematic view of the white light ray image before compression according to the present invention.
• Figure 4b is a schematic view of the image of the white light ray used in the present invention after compression
• Fig. 4c is a schematic view showing the image of the white light ray used in the present invention after decompression. detailed description
• the value document identification method 100 of the present invention includes the following steps: (1 1) collecting a sample image of a current value document and obtaining corresponding image data; (12) performing image data according to the current value file. Price file identification; (1 3) forcing the image data collected by the image into long image data; (14) compressing and storing the converted image data using the image data fast storage method 200; (15) restoring the compressed storage Image data; (16) determining the authenticity of the current value document based on the decompressed image data.
• it is detected whether the current value document enters the sample area "Yes" collects the image data, and "No” continues the detection.
• the invention enables automatic detection without manual operation.
• the image data fast storage method 200 includes the following steps: (21) obtaining long integer image data, and dividing the long integer image data into N partitions; (22) using N phases The mutually corresponding data mask extracts the long integer image data in N partitions by a bitwise AND operation, and extracts M points by M*N points in each row, and extracts a point of L points in each column.
• N is an integer greater than or equal to 2
• L and M are integers greater than or equal to 1
• Data extracted by N data masks is integrated by a bitwise OR operation to obtain encoded image data and stored.
• the value document identification device 300 includes a collection module 31, an identification module 32, a compression module 33, a storage module 34, and a decompression module 35.
• the collection module 31 collects the image of the current value document and obtains the image.
• the corresponding image data S p the identification module 32 performs the value document identification according to the image data of the current value document, thereby obtaining the value document identification data Si for output (can be output to the display or the alarm device);
• the compression module 33 will gather S P to the image data converted to a long integer mandatory image data S, and the image data S long compression operation (suction operation point and re-encoding) to obtain coded image data S.
• decompression module 35 for restoring the coded image data S c to obtain decompressed image data S d; 34 a storage module for storing data.
• a contact type image sensor may detect whether the current value document has entered Bian-like domain, and enters the application zone preclude preclude the image data S p in the set current value document, if the current If the value document does not enter the sample area, the test will continue.
• the compression module 33 may include a conversion unit 331 and the coding unit 332, the converting unit 331 to preclude the set image data S p to cast long integer image data S i; the encoding unit 332
• the long image data Si is equally divided into N partitions, N is an integer greater than or equal to 2, and the long integer image data in the N partitions is extracted by a bitwise operation using N mutually corresponding data masks. Then, the data extracted by the N data masks is integrated by the "bit or" operation to obtain the encoded image data S c and stored, and the extraction method is to extract M points of M*N points in each row, and extract a point of L points in each column.
• N is an integer greater than or equal to 2
• L and M are integers greater than or equal to 1.
• the sample image collected by the collection module 31 includes an infrared sample image, a transmission sample image, and a white light sample image, and obtains image data S pl of the infrared sample image and image data of the projection sample image.
• the image data S p3 of the S p2 and the white light sample image, and the compression storage process of the image data S pl , S p2 , and S p3 are as follows:
• the compressed image data S p2 of the sample image is projected
• the resolution of the infrared image and the projected image is 240*480
• the image data S pl and S p2 collected by the image are single-byte image data
• the conversion unit 331 in the compression module 33 collects the image.
• the single-byte image data S pl , S p2 are stored as eight pixels as a whole, thereby converting the single-byte image data S pl , S p2 into 8-byte image data S u , S 12 and transported to the storage.
• the converting unit 331 integrates the collected single-byte image data Sp3 into eight pixels as a whole, so that the single-byte image data Sp3 is converted into 8-byte image data S13 ;
• the encoding unit 332 divides the 8-byte image data S 13 into a U zone and a D zone, the U zone represents the upper half of the white light image (one grid represents one pixel), and the D zone represents a white light image. In the lower half, the subscripts of the values in the grid represent the number of rows and columns of the pixel points;
• A1 represents an 8-byte image data in the U area
• B1 represents an 8-byte image data in the D area
• the data mask M1 is used to extract 8-byte image data in the U area:
• the data mask M2 is used to extract 8-byte image data in the D area:
• an 8-byte data in the U area extracts 4 valid bytes
• an 8-byte data in the D area also extracts 4 valid bytes, that is, a total of 16 bytes of data extracts 8 bytes of valid data.
• the pre-compression image original sample image
• the compressed image and the decompressed image of the white light sample image, wherein the lattice with the diagonal line in FIG. 4a is valid in the data to be drawn.
• the encoded image data S stored in the R area is required.
• the recovery process is performed to obtain the decompressed image data S d , and the compressed image is restored to a normal image.
• the specific decompression step is as follows: the upper half of the decompressed image is recorded as a region, and the lower half is recorded as a region, in this embodiment.
• the number of lines of the image after decompression is half of the number of lines of the image before compression
• the number of columns of the image after decompression is half of the number of columns of the image before compression, according to the size data of the image template after decompression and the encoded image data s.
• the present invention is based on the operation based on the 8-byte data type. Using “bitwise and arithmetic" and “bitwise OR operation”, 16 bytes of data processing is completed each time, of which 8 bytes of valid data, the number of moves is one-eighth of the traditional method, and each move requires operation: Press the "bit and" operation twice, press the "bit or” operation once, and perform one shift operation.
• the TMS320 platform based on 600M CPU frequency stores two medium resolution (240*400) infrared sample images and transmission sample images, a high resolution (480*800) white light image, traditional
• the method storage time is greater than 70ms, which is much higher than the required 15ms.
• the above embodiment is based on an 8-byte (64-bit) processor.
• the processor of the present invention supports 16 bytes (128 bits)
• the single-byte image data is converted into 16-byte image data, and the data mask M1 is taken.
• M1 0x00ff00ff00ff00ff00ff00ff00ff00ff00ff00;
• M2 0xff00ff00ff00ff00ff00ff00ff00ff00ff00ff00ff00ff00ff00ff00ff00ff00;
• R (Al&Ml) I (A2&M2)
• 32 bytes of data can be realized once, and 16 bytes of valid data are extracted.
• a 4-byte processor and a 32-byte processor can all use similar processing methods.
• the extraction method is to extract a point every four points, and the data is equally divided into four regions of U1, U2, U3, and U4, and four data masks are taken.
• Al, A2, A3, A4 are an 8-byte data of the U1, U2, U3, U4 areas, respectively
• the contact image sensor 31 can detect whether the current value document enters the sample area, and collect the infrared sample image, the transmission sample image, and the white light sample image after the current value document enters the sample area.
• identification module 32 according to the image data S p3 be valuable document identification, in order to obtain valuable document identification data Si ( The serial number information of the current value document); the compression module 33 forcibly converts the collected image data S pl , S p2 , S p3 into the long image data S u , S 12 and S 13 , and the length The integer image data S 13 is subjected to a compression operation to obtain encoded image data S.
• the storage module 34 stores the long integer image data S u , S 12 and the encoded image data S c ; and the decompression module 35 restores the encoded image data S.
• the identification module 32 performs the value document identification based on the long image data S u , S 12 and the decompressed image data S d and determines the authenticity of the current value document to obtain further value document identification.
• Data Si the type, denomination, orientation, authenticity information of the current value document).

Landscapes

• Engineering & Computer Science (AREA)
• Theoretical Computer Science (AREA)
• Physics & Mathematics (AREA)
• General Physics & Mathematics (AREA)
• Databases & Information Systems (AREA)
• General Engineering & Computer Science (AREA)
• Data Mining & Analysis (AREA)
• Software Systems (AREA)
• Inspection Of Paper Currency And Valuable Securities (AREA)
• Image Processing (AREA)
• Compression Or Coding Systems Of Tv Signals (AREA)
• Compression Of Band Width Or Redundancy In Fax (AREA)
• Image Analysis (AREA)
• Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Priority Applications (5)

Applications Claiming Priority (2)

Publications (1)

Family

ID=46292199

Family Applications (1)

Country Status (8)

Families Citing this family (5)

Citations (5)

Family Cites Families (6)

• 2011
  • 2011-11-24 CN CN201110380229.9A patent/CN102521278B/zh active Active
• 2012
  • 2012-10-26 EP EP12850974.2A patent/EP2784695B1/en active Active
  • 2012-10-26 IN IN855MUN2014 patent/IN2014MN00855A/en unknown
  • 2012-10-26 US US14/348,389 patent/US9224185B2/en not_active Expired - Fee Related
  • 2012-10-26 WO PCT/CN2012/083584 patent/WO2013075571A1/zh not_active Ceased
  • 2012-10-26 AU AU2012343073A patent/AU2012343073B2/en not_active Ceased
• 2014
  • 2014-05-06 CL CL2014001185A patent/CL2014001185A1/es unknown
  • 2014-05-09 ZA ZA2014/03346A patent/ZA201403346B/en unknown

Patent Citations (5)

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events