JP2013206489A - Image acquisition device, image acquisition method, and image acquisition program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image acquisition device having high operational efficiency.SOLUTION: An image acquisition device 1 according to one aspect of the present invention includes: first recording means 2 which cannot be randomly accessed, and is adapted to store an image file P; and catalog data generating means 4 for collecting a header part and a frame index out of an image file P stored in the first recording means 2 so as to generate a catalog file C, and acquiring a desired image file out of image files P stored in the first recording means 2 on the basis of the catalog file C.

Description

  The present invention relates to an image acquisition apparatus, an image acquisition method, and an image acquisition program, and more particularly to an apparatus, a method, and a program for acquiring a desired image file from an image file stored in a recording unit that cannot be randomly accessed.

  Image files used for broadcasting and the like use image file wrapping standards such as MXF (Material eXchange Format) (specified by SMPTE377M and related SMPTE standards), image data, audio data, auxiliary data, etc. Are arranged in time series.

  Image data refers to data that has been compressed with high efficiency using DCT (Discrete Cosine Transform) coding technology or the like. However, since variable code amount control (VBR) or inter-frame compression technology is used, the minimum unit of image data ( The amount of data in a frame or field is not constant.

  When a part of time is cut out from one image file and copied as another image file, the cut-out start point (IN point) and the cut-out section (DURATION) are generally designated by the frame position. Since the image data has a different data length in each frame, a simple calculation using a known multiplier or the like cannot determine the start position of the frame data corresponding to the IN point by the number of bytes.

  In an image file wrapping standard such as MXF, a frame index data area is provided in a file, and the data position of image data, audio data, and auxiliary data at a specific frame position stored in the file by referring to this frame index data Can be converted into byte position information in the file, and data according to the designated IN point and DURALION can be cut out.

  The storage location in the file of the frame index data area can be freely defined, and there may be all data near the beginning (header part) or near the end (footer part), but in a distributed manner in the middle part (body part) Sometimes placed.

  Therefore, in order to realize the clipping process with all types of image files, it is necessary to randomly access the image file and collect all stored frame indexes, and then obtain the byte position of the clipped portion.

  Generally, in an apparatus that requires image cropping, an image file is often stored in a random access medium such as a hard disk drive (HDD) or a memory card. A technique for cutting out a part of a stored image file and copying it as another file can be realized relatively easily by using a general technique by referring to the frame index.

  By the way, in recent years, an archive system for storing a large number of image files in a recording device such as an LTO (Linear Tape-Open) for a long period of time has become common.

  As shown in FIG. 7, the magnetic recording drive 101 used in LTO or the like stores data while winding the magnetic tape 102b wound around the cartridge reel 102a of the LTO magnetic recording tape 102 around the winding reel 103. In this configuration, data is read out by the magnetic head 104, signal processing is performed by the signal processing circuit 105, and output as read data. Incidentally, the upper part of FIG. 7 shows a schematic diagram of the magnetic recording drive 101, and the lower part shows an image file stored on the magnetic tape 102b.

  For this reason, recording devices such as LTO can record a large amount of data at a high density and at a low cost, and the sequential access is very fast, but because of the take-up type tape media, the high-speed random accessibility is extremely poor. ing.

  On the other hand, a recording device 201 such as an HDD records image file data on an infinite number of tracks carved in the rotational direction of a magnetic recording disk 202 that rotates in a fixed manner, as shown in FIG. By simply swinging the magnetic head 204 provided at the tip based on the operation of the actuator 205, it is possible to instantaneously access data far from the address, and after the read data is processed by the signal processing circuit 206, the read data is It is the structure which outputs. Therefore, the recording device 201 has extremely good random accessibility. Incidentally, a schematic diagram of the recording apparatus 201 is shown in the upper part of FIG. 8, and an image file stored in the magnetic recording disk 202 is shown in the lower part.

  However, since the recording device 201 such as an HDD costs about tens of times as much as the capacity of the recording device such as an LTO, it is not suitable for storing a large amount of image files such as an archive device.

  Therefore, there is a case in which a recording device such as an LTO is used as a main recording device and a recording device such as an HDD is used as a cache device for primary reading of data.

JP 2007-265467 A

  In the above-mentioned method, it is necessary to copy all of the original image file to be cut out to a recording device such as an HDD before cutting out the image file. Therefore, when the cut-out source file is very large and the cut-out part is relatively small However, the operation efficiency was extremely poor because the processing took a long time.

  The objective of this invention is providing the image acquisition apparatus, the image acquisition method, and image acquisition program which solve the subject mentioned above.

  An image acquisition apparatus according to an aspect of the present invention includes a first recording unit that is not randomly accessible and stores an image file, and a header section and a frame from the image file stored in the first recording unit. Catalog data generating means for collecting an index to generate a catalog file and acquiring a desired image file from the image file stored in the first recording means based on the catalog file.

  An image acquisition method according to an aspect of the present invention includes a step of generating a catalog file by collecting a header part and a frame index from an image file recorded in a first recording means that is not randomly accessible, and based on the catalog file And obtaining a desired image file from the image file stored in the first recording means.

  An image acquisition program according to an aspect of the present invention includes a process for generating a catalog file by collecting a header part and a frame index from an image file recorded in a first recording unit that is not randomly accessible to a computer, and the catalog Obtaining a desired image file from the image file stored in the first recording means based on the file.

  According to the present invention, it is possible to provide an image acquisition device, an image acquisition method, and an image acquisition program with high operational efficiency.

1 is a configuration block diagram illustrating an image acquisition device according to a first embodiment of the present invention. It is a block diagram showing the configuration of a more preferable image acquisition apparatus according to the first embodiment of the present invention. It is a flowchart figure which shows the production | generation process of a catalog file in the image acquisition apparatus of the 1st Embodiment of this invention. It is a figure which shows the structure of a general image file. The upper part shows the structure of the image file, and the lower part shows the structure of the catalog file generated from the image file. It is a flowchart figure which shows the acquisition process of the cutout image file in the image acquisition apparatus of the 1st Embodiment of this invention. It is a schematic diagram which shows the structure of recording apparatuses, such as general LTO. It is a schematic diagram which shows the structure of recording apparatuses, such as a common HDD.

  An image acquisition apparatus, an image acquisition method, and an image acquisition program according to an embodiment of the present invention will be described. However, the present invention is not limited to the following embodiment. In addition, for clarity of explanation, the following description and drawings are simplified as appropriate.

<First Embodiment>
As shown in FIG. 1, in the image acquisition device 1 of the present embodiment, the catalog file generation unit 4 previously collects a header part and a frame index from an image file P to generate a catalog file, and based on the catalog file A desired image file is acquired from the image file P stored in the recording unit 2 that cannot be randomly accessed.

  As a result, the recording unit 2 that takes a long time for random access processing can acquire a cut-out image file without frequently performing a cueing operation, so that the processing time for a cut-out request for an image file can be minimized, It can contribute to the improvement of operational efficiency.

  Below, the preferable structure of the image acquisition apparatus 1 of the said structure is demonstrated. As shown in FIG. 2, the image acquisition apparatus 1 includes a first recording unit 2, a second recording unit 3, and a catalog file generation unit 4.

  The first recording unit 2 is not randomly accessible and stores the image file P. The first recording unit 2 of the present embodiment is composed of an LTO recording apparatus. That is, the first recording unit 2 includes an LTO magnetic recording drive 5, an LTO magnetic recording tape 6, and a cart robot 7, as shown in FIG.

  Since the LTO magnetic recording drive 5 has already been described in the background art section, detailed description thereof is omitted, but the image file P is written to or read from the LTO magnetic recording tape 6. A plurality of LTO magnetic recording tapes 6 are prepared and stored in a rack 8.

  The cart robot 7 takes out the desired LTO magnetic recording tape 6 from the rack 8 and loads it into the LTO magnetic recording drive 5. Here, the cart robot 7 operates based on an image cut-out request input from the outside when acquiring a desired image file.

  The second recording unit 3 stores the catalog file C generated by the catalog file generation unit 4. The second recording unit 3 is preferably composed of a randomly accessible recording device such as an HDD.

  The catalog file generation unit 4 generates a catalog file C by collecting the header part and the frame index from the image file P stored in the first recording unit 2, and creates the catalog file C stored in the second recording unit 3. Based on this, a desired image file is acquired from the image file P stored in the first recording unit 2. Here, the function of the catalog file generation unit 4 of the present embodiment is realized by using an arithmetic device such as a CPU (Central Processing Unit) mounted on the image acquisition device 1. However, the function of the catalog file generation unit 4 may be realized using hardware resources.

  Next, a catalog file generation process in the image acquisition apparatus 1 will be described. As shown in FIGS. 2 and 3, an image file P is input to the catalog file generation unit 4 from the outside (S1). Here, the image file P of the present embodiment is an image file represented by MXF.

  As shown in FIG. 4, such an image file P is composed of a header part, a body part, and a footer part, and can be determined by inserting a unique data packet serving as a partition called Partition Pack at the head of each part. It has become.

  In the header part, metadata (Header Metadata) is mainly stored, and all or part of the frame index (Index) may be stored.

  The footer part may not store anything, but the same metadata as the header part may be inserted. The footer unit may also store a frame index. In general, when metadata is stored in the footer part, the metadata in the header part is often incomplete (undefined), and some metadata that was undefined when the header part was generated is described again in the footer part. May be completed.

  In the body portion, image data, audio data, and auxiliary data are usually divided into units called “Edit Units” and are continued after the Body Partition Pack. The length of the edit unit can be defined in any way, but generally corresponds to a video frame unit or a field unit.

  The catalog file generation unit 4 outputs the input image file P to the first recording unit 2 (S2). The first recording unit 2 stores the input image file P in the LTO magnetic recording tape 6.

  At this time, the catalog file generation unit 4 refers to the contents of the sequentially input image file P, and performs header part analysis, footer part analysis, and collection of frame indexes included in the body part (S3). Here, the metadata included in the header part may be in an unfinished state. In this case, the final completed state data may be included in the metadata included in the footer part. The completed metadata is generated by combining the metadata.

  In addition, the frame index may be diffused and arranged in the header part, the footer part, and the body part, and all the frame indexes are generated as completion indexes.

  Then, the catalog file generation unit 4 generates a catalog file by combining the completed metadata and the completed index data (S4).

  Here, a specific example of processing for collecting a metadata and a frame index from an MXF image file and generating a catalog file will be described.

  The image file P shown in the upper part of FIG. 5 has a structure in which the body part is divided into a plurality of parts. In this case, each of the second and subsequent body parts has a frame index, and frame data (edit unit) indicated by the frame index. ) Is only included in the previous body part. Therefore, when collecting the frame indexes from the image file P in the example of FIG. 5, the catalog file generating unit 4 collects all the frame indexes included in all the body portions.

  Then, as shown in the lower part of FIG. 5, the catalog file generation unit 4 generates completed metadata from the metadata included in the header portion and the metadata of the footer portion including the metadata that has not been completed in the header portion. The data obtained by concatenating the frame indexes collected from the body part is used as a catalog file.

  The catalog file generator 4 outputs the catalog file C generated as described above to the second recording unit 3 (S5). The second recording unit 3 stores the input catalog file C. Since the catalog file C has a very small capacity compared to the image file P, the catalog file C has a small capacity, and even if many catalog files C are stored in an expensive recording device (second recording unit 3) such as an HDD. There is no particular adverse effect. For example, for an image file compressed at a data rate of about 50 Mbps, the data amount per second is 50 Mbits, that is, 6.25 Mbytes, but the index data is about 240 bytes, and the header part and footer part are the maximum. It is about several megabytes. For this reason, a capacity of 220 Tbytes or more is required to store an image file for 10,000 hours, and storage to a recording device such as an LTO with a low unit cost is reasonable, but a catalog file for 10,000 hours is about 9 Gbytes. Therefore, there is no problem even if it is stored in a recording device such as an HDD.

  Next, a cutout image acquisition process in the image acquisition apparatus 1 will be described. As shown in FIG. 2 and FIG. 6, when an image cut-out request for reading out a temporal portion of the image file P stored in the LTO magnetic recording tape 6 is input to the image acquisition apparatus 1, the image cut-out is performed. Based on the request, the catalog file generation unit 4 reads the catalog file C corresponding to the requested image file (cut-out image file) (S11). Then, the catalog file generating unit 4 analyzes the frame index of the catalog file C and calculates a necessary body part address (S12). Incidentally, the image cutout request includes information on the name of the target image file, the start point (IN point) at which the cutout is started, and the cutout section (DURATION). The IN point and DURATION are specified in units of video frames. Therefore, the byte position of the image file P to be cut out is acquired from the index information of the corresponding catalog file C based on the IN point information of the image cut-out request, and the frame section defined by the DURATION information is based on the IN point. Is calculated from the index information of the catalog file C.

  Next, the catalog file generation unit 4 reads the completed metadata from the catalog file C in order to configure the header part of the cut-out image file after cut-out. Based on the input image cut-out request, the data size after cut-out, etc. After the metadata corresponding to is changed, it is stored in, for example, the second recording unit 3 as data of the header portion of the image cutout file after cutout (S13). However, the storage location of the data in the header portion of the image cutout file after cutout may be stored in a recording unit provided in the image acquisition device.

  Next, the catalog file generation unit 4 instructs the cart robot 7 to take out the LTO magnetic recording tape 6 including the address from the rack 8 and load it into the LTO magnetic recording drive 5 based on the input image cutout request. (S14). Based on this instruction, the cart robot 7 takes out the corresponding LTO magnetic recording tape 6 from the rack 8 and loads it into the LTO magnetic recording drive 5.

  Next, the catalog file generation unit 4 gives a first instruction to send the LTO magnetic recording tape 6 from the start address of the corresponding cut-out image file to a position where the byte address corresponding to the IN point calculated in the step S12 is further added. This is performed for the LTO magnetic recording drive 5 of the recording unit 2 (S15). Based on the instruction, the LTO magnetic recording drive 5 sends the LTO magnetic recording tape 6 to the head of the body part in the cut-out image file.

  Next, the catalog file generation unit 4 instructs the LTO magnetic recording drive of the first recording unit 2 to read data from the LTO magnetic recording tape 6 by the data size of the required number of bytes corresponding to DURATION calculated in the step S12. 5 (S16). Based on the instruction, the LTO magnetic recording drive 5 reads data from the LTO magnetic recording tape 6 and outputs the data to the catalog file generation unit 4.

  Next, the catalog file generation unit 4 connects the body part data read in step S16 to the header part data changed in step S13, connects the footer part (without metadata) as fixed data, and cuts it out. Obtained as an image file (S17).

  In this way, it is possible to obtain a cutout image file without frequently cuing the LTO magnetic recording tape 6 that takes a long random access processing time, and to minimize the processing time for an image cutout request.

  Usually, a recording device (first recording unit 2) such as an LTO performs data read at a high speed of 100 Mbytes / second or more when performing sequential read in which sequential data recorded on the LTO magnetic recording tape 6 is read in order. It is a device that can read out. However, the random read that reads out the dispersed data while feeding the tape to a specific byte position in the file depends on the feed width of the tape, but the reading performance is remarkably deteriorated.

  To perform partial segmentation of an image file, as a pre-process for identifying the segment to be segmented, the front (header portion), the rear (footer portion), and the intermediate positions (body portions) that are sporadically arranged at the address in the image file Frame index data stored in the header, and the metadata distributed in the header and footer are combined to generate header data after cutting. The above processing is applied to the LTO magnetic recording tape. To execute the program while randomly accessing the stored data, the tape feeding and returning time is much longer than the data reading time, which is not practical.

  For this reason, a method of reading out all image files to the HDD cache device and performing extraction processing in the HDD with high random access performance can be considered. However, with the recent improvement in quality of image files, the image files become very large data. Therefore, for example, even if it is desired to read a part of several megabytes from an image file of several hundred gigabytes, data of several hundred gigabytes must be read, and both processing time and HDD cache device capacity are wasted. It was.

  In the present embodiment, the tape feeding to the LTO magnetic recording tape 6 may be performed once, and the data to be read out may be only a portion necessary for cutting out, which is efficient. Further, the required capacity of the HDD cache device is very small because the catalog file is extremely small data, so it is not particularly expensive and economical.

<Other embodiments>
In the first embodiment, an example in which an HDD (Hard Disk Drive) is used as the second recording unit 3 is given. However, in recent years, an SSD (Random Access Performance using a generalized semiconductor memory is extremely high. A solid state disk) may be used.

  The present invention is not limited to the above-described embodiment, and can be changed as appropriate without departing from the spirit of the present invention.

DESCRIPTION OF SYMBOLS 1 Image acquisition apparatus 2 1st recording part 3 2nd recording part 4 Catalog file generation part 5 LTO magnetic recording drive 6 LTO magnetic recording tape 7 Cart robot 8 Rack 101 LTO magnetic recording drive 102 LTO magnetic recording tape 102a Cartridge reel 102b Magnetic tape 103 Take-up reel 104 Magnetic head 105 Signal processing circuit 201 Recording device 202 such as HDD Magnetic recording disk 203 Arm 204 Magnetic head 206 Signal processing circuit P Image file C Catalog file

Claims (6)

A first recording means that is randomly inaccessible and stores an image file;
A catalog file is generated by collecting a header part and a frame index from the image file stored in the first recording unit, and based on the catalog file, from the image file stored in the first recording unit Catalog data generating means for acquiring a desired image file;
An image acquisition apparatus comprising:   The image acquisition apparatus according to claim 1, wherein the catalog file is stored in a second recording unit that is randomly accessible. A step of collecting a header part and a frame index from an image file to be recorded in a first recording means that cannot be randomly accessed to generate a catalog file;
Obtaining a desired image file from the image file stored in the first recording means based on the catalog file;
An image acquisition method comprising:   The image acquisition method according to claim 3, wherein the catalog file is stored in a second recording unit that is randomly accessible. On the computer,
A process of collecting a header part and a frame index from an image file recorded in a first recording means that cannot be randomly accessed to generate a catalog file;
Obtaining a desired image file from the image file stored in the first recording means based on the catalog file;
An image acquisition program that executes   The image acquisition method according to claim 5, wherein the catalog file is stored in a second recording unit that is randomly accessible.

Images