JP2007336138A - Video recording apparatus and video recording program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a video recording apparatus which is capable of long-time recording and can be simplified in system configuration. <P>SOLUTION: The video recording apparatus includes a readout 2, storage 3, reencoder 4, selector 5, and output portion 6. Video data read out frame by frame by the readout 2 are separated into intra-coded frame data coded without using interframe prediction, and non-intra-coded frame data coded by using interframe prediction and is stored separately in the storage 3. Then, only the intra-coded frame data are reencoded by the reencoder 4. Frame data selected by the selector 5 using an amount of image characteristic obtained during the reencoding processing are recorded in a recording medium by the output portion 6. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a video recording apparatus and a video recording program for re-encoding and recording video data encoded by a video encoding method.

  Conventionally, in surveillance camera systems aimed at surveillance of crime prevention, accidents, disasters, etc., a surveillance camera is installed in a remote place where the surveillance object is located, and the captured video is transmitted to the surveillance center via an analog dedicated line and displayed for a long time. In many cases, it is reproduced above or recorded as an analog signal with a long-time recording VTR called a time-lapse (intermittent) VTR as necessary. However, in recent years, as a result of the development of IP network infrastructure, the increase in capacity and price of digital recording media such as hard disks and DVDs, and the digitalization of surveillance cameras, digital video recording of surveillance cameras has also become possible. The replacement of is progressing rapidly.

  In general, in order to digitally record a video taken by a surveillance camera, an analog signal of the photographed video is converted into a digital signal by the inside of the surveillance camera or by a device connected to the surveillance camera, encoded, The encoded digital video signal is transmitted to a recording device via a network and recorded on a digital recording medium such as a hard disk or a DVD. As a video signal encoding method, there are a method in which each frame is independently encoded as one still image, and a method in which inter-frame prediction is performed by using the correlation of consecutive frames. Among these, the method of encoding each frame independently is often used in a system for time-lapse recording of video data of a digital signal, and a representative example is the JPEG method (see, for example, Patent Document 1).

  However, the method of encoding each frame independently has a lower video data compression rate than the method of encoding by performing inter-frame prediction, and requires a wider bandwidth for transmission over a network. There is a problem that.

  On the other hand, representative methods for encoding using inter-frame prediction include encoding methods such as MPEG-2 and MPEG-4, and video data encoded by these methods is digital recording media. In many cases, everything is recorded continuously. Since the encoding method using inter-frame prediction has a high compression rate, the frequency band required for transmission using a network line may be relatively narrow, but all of the encoded video data is recorded. As a result, there is a problem that a recording device having a capacity larger than that of the JPEG video recorded by time lapse is required, resulting in a high system cost. For this reason, there has been used a method of detecting whether or not a suspicious person or object or an abnormal phenomenon is included in the video of the surveillance camera and recording only the video when the abnormality is detected. However, this method is effective, for example, when detecting intruders with a security camera on a holiday or at night when there are few people coming and going, but when continuous monitoring such as disasters and accidents such as rivers and roads is required There are difficulties.

  As a method for solving this problem, inter-frame prediction is not used by detecting a picture start code from video data encoded by the MPEG method, specifying the position of the picture header, and acquiring the picture type from the picture header. There is an example in which only an I frame is recorded (see, for example, Patent Document 2).

  However, in the method of extracting and recording only the I frame described in Patent Document 2, the code amount of the recorded video data depends on the type of the video data and the device that encodes the video data. There is a problem. In addition, when encoding MPEG, in general, encoding is often performed at a fixed bit rate, and in the case of a video with a complicated picture and almost no movement, the code amount allocated to the I frame is relatively large, There is a problem that it is impossible to record to the recording device for a long time.

In addition, in the method described in Patent Document 2, when an abnormal video is captured by the surveillance camera, an abnormality has occurred in the monitoring target area in order to stop the time-lapse recording and switch to normal recording. However, a device such as a sensor for detecting whether or not an abnormality has occurred is required, the video recording apparatus must be notified that an abnormality has been detected, and the procedure of the recording process must be switched, resulting in a complicated system configuration.
JP 2001-111959 A JP 2003-101944 A

  The present invention has been made in view of the above-described conventional circumstances, and can perform recording for a long time without depending on the type of video data and the device that encodes the video data, and the abnormality of the monitoring target area. It is an object of the present invention to provide a video recording apparatus and a video recording program that can simplify the system configuration without requiring a device such as a sensor for detecting an image.

  The video recording apparatus of the present invention includes any one of a first encoding method for encoding a frame image within a frame and a second encoding method for encoding based on a correlation between frames. A video recording apparatus that selects, encodes, and sequentially records in units of frames, a reading unit that sequentially reads out the encoded frame images in units of frames, and a frame image encoded by the first encoding method temporarily The first storage unit for storing automatically, and the i-th (i is a natural number) and (i + 1) -th frame images encoded by the first encoding method (L ≧ 1) Of the second storage unit that temporarily stores the frame image encoded by the second encoding method as the i-th set of frame image groups, and the frame image group stored in the second storage unit, The j-th set that satisfies a predetermined condition (j is Number) and each of the preceding and following M sets (M is a natural number) of the frame image group, all frame images stored in the first storage unit, and the first selection unit A first re-encoding unit that re-encodes using the group of frame images, and a recording unit that records the frame image re-encoded by the first re-encoding unit.

  With this configuration, the image frame encoded within the frame and the image frame encoded based on the correlation between the frames are separated, and the image frame encoded within the frame is re-encoded, so that time-lapse recording can be performed. This makes it possible to record video data for a long time. Also, by temporarily storing the frame image encoded based on the correlation between frames, when an abnormal video is shot by the surveillance camera, all frames are continuously switched to the normal frame rate. Can be recorded.

  As one aspect of the present invention, in the video recording apparatus, the first re-encoding unit includes a decoding unit that decodes a frame image stored in the first storage unit, and the decoding unit A second re-encoding unit that re-encodes the frame image decoded by the second encoding method, a frame image re-encoded by the second re-encoding unit, and the first storage A second selection unit that selects one of the frame images stored in the unit, a frame image selected by the second selection unit, and a frame image group selected by the first selection unit in a time order And a portion provided with a section.

  With this configuration, the image frame encoded within the frame and the image frame encoded based on the correlation between the frames are separated, and the image frame encoded within the frame is re-encoded, so that time-lapse recording can be performed. This makes it possible to record video data for a long time. Also, by temporarily storing the frame image encoded based on the correlation between frames, when an abnormal video is shot by the surveillance camera, all frames are continuously switched to the normal frame rate. Can be recorded.

  Furthermore, as one aspect of the present invention, in the video recording apparatus, the predetermined condition includes: a feature amount calculated based on encoding information acquired at the time of re-encoding, and a frame image to be re-encoded. Also included is a case where the difference from the feature amount calculated from the frame image re-encoded immediately before is not less than a predetermined range.

  With this configuration, in addition to the effects described above, it is possible to switch between full frame recording and type lapse recording by using feature amounts based on encoding information at the time of re-encoding. This eliminates the need for a device such as a sensor for detecting whether or not the image has occurred, and simplifies the configuration of the video recording system that can switch the recording mode.

  The video recording program of the present invention includes any one of a first encoding method for encoding a frame image within a frame and a second encoding method for encoding based on a correlation between frames. A video recording program for causing a computer to execute a function of selecting and encoding and sequentially recording in units of frames, wherein a function of sequentially reading out the encoded frame images in units of frames and encoding with the first encoding method For temporarily storing the frame image in the first storage unit and the i th (i is a natural number) and (i + 1) th frame image encoded by the first encoding method. A function of temporarily storing frame images encoded by the second encoding method (L ≧ 1) in the second storage unit as an i-th set of frame image groups, and the second storage unit Memorized frame Among the image groups, a function of selecting the frame image group of the j-th set (j is a natural number) that satisfies a predetermined condition and the M groups before and after each (M is a natural number), and all of the frame images stored in the first storage unit A computer executes a frame image, a function of re-encoding using the selected frame image group, and a function of recording the re-encoded frame image.

  This program separates the image frame encoded within the frame from the image frame encoded based on the correlation between the frames, and re-encodes the image frame encoded within the frame, so that time-lapse recording can be performed. This makes it possible to record video data for a long time. Also, by temporarily storing the frame image encoded based on the correlation between frames, when an abnormal video is shot by the surveillance camera, all frames are continuously switched to the normal frame rate. Can be recorded.

  Further, as one aspect of the present invention, in the video recording program, the re-encoding function includes a function of decoding a frame image stored in the first storage unit, and a function of decoding the decoded frame image. A function of re-encoding by the second encoding method, a function of selecting one of the frame image re-encoded by the second encoding method and the frame image stored in the first storage unit, The selected frame image and a function of rearranging the selected frame image group in time order are included.

  This program separates the image frame encoded within the frame from the image frame encoded based on the correlation between the frames, and re-encodes the image frame encoded within the frame, so that time-lapse recording can be performed. This makes it possible to record video data for a long time. Also, by temporarily storing the frame image encoded based on the correlation between frames, when an abnormal video is shot by the surveillance camera, all frames are continuously switched to the normal frame rate. Can be recorded.

  Furthermore, as one aspect of the present invention, in the video program, the predetermined condition includes a feature amount calculated based on encoding information acquired at the time of re-encoding and a frame image immediately before the frame image to be re-encoded. The difference between the feature amount calculated from the re-encoded frame image is not less than a predetermined range.

  In addition to the above effects, this program can switch between full-frame recording and type-lapse recording by using the feature amount based on the encoding information at the time of re-encoding. This eliminates the need for a device such as a sensor for detecting whether or not this occurs, and simplifies the configuration of a video recording system that can switch the recording mode.

  According to the present invention, it is possible to perform recording for a long time without depending on the type of video data and the device that encodes the video data, and a device such as a sensor for detecting an abnormality in the monitoring target area. It is possible to provide a video recording apparatus and a video recording program that can simplify the system configuration without being required.

  In the video recording apparatus according to the embodiment of the present invention, the MPEG-4 Visual encoding method is used as the encoding method of video captured by the surveillance camera. In the MPEG-4 encoding system, a profile is defined in which various applications are assumed and encoding tools for realizing each application are combined. Here, a simple profile encoding process, which is the most basic profile that stipulates that a decoder conforming to MPEG-4 Visual must always decode, will be described.

  Hereinafter, an outline of processing of the MPEG-4 Visual encoding method will be described. FIG. 1 is a diagram showing a schematic configuration of an MPEG-4 Visual encoding apparatus. In the MPEG-4 Visual encoding system, encoding is performed in units called VOP (Video Object Plane). The simplest VOP corresponds to a frame, and in this embodiment, this case will be described.

  In FIG. 1, an MPEG-4 Visual encoding device 1 includes an input unit 10, a motion detection / encoding unit 11, a prediction unit 12, a conversion unit 13, a DCT encoding unit 14, a reference image generation unit 15, and a reference image storage. A unit 16, a parameter setting unit 17, and an output unit 18 are provided.

  The input unit 10 inputs video data to be encoded in units of VOPs, and divides each VOP into macroblocks having units of 16 × 16 pixels.

  The motion detection / encoding unit 11 includes a motion detection unit 111, a macroblock type setting unit 112, and a motion vector encoding unit 113, as shown in FIG. In FIG. 2, the motion detection unit 111 calculates a motion vector using the macroblock data input from the input unit 10 and the reference image data stored in the reference image storage unit 16, and the macroblock type The setting unit 112 sets a macroblock type based on the calculated motion vector. In addition, the motion vector encoding unit 113 performs variable length encoding only for the set macroblock type non-intra macroblock.

  The prediction unit 12 generates a predicted image using the motion vector calculated by the motion detection / encoding unit 11 and the reference image data stored in the reference image storage unit 16.

  As illustrated in FIG. 3, the conversion unit 13 includes a prediction error generation unit 131, a DCT unit 132, and a quantization unit 133. In FIG. 3, the prediction error generation unit 131 generates error image macroblock data from the macroblock data input from the input unit 10 and the prediction image generated by the prediction unit 12 only for non-intra macroblocks. In the case of an intra macroblock, the macroblock data input from the input unit 10 is output as it is. Further, the DCT unit 132 performs two-dimensional discrete cosine transform (DCT) for each 8 × 8 pixel block on the macroblock data output from the prediction error generation unit 131, and the quantization unit 133 receives the data from the DCT unit 132. The output two-dimensional DCT coefficient is quantized.

  As shown in FIG. 4, the DCT encoding unit 14 includes a DC / AC prediction unit 141, a scanning unit 142, and a variable length encoding unit 143. In FIG. 4, the DC / AC prediction unit 141 performs two-dimensional quantization DCT from a value predicted based on the two-dimensional quantization DCT coefficient output from the conversion unit 13 and the two-dimensional DCT coefficient of an adjacent macroblock. The error component of the coefficient is calculated. The scan unit 142 rearranges the error components of the two-dimensional quantized DCT coefficients calculated by the DC / AC prediction unit 141 in the order of encoding, and the variable length encoding unit 143 performs the rearranged two-dimensional quantization. The DCT coefficient error component is encoded.

  As shown in FIG. 5, the reference image generation unit 15 includes an inverse quantization unit 151, an inverse DCT unit 152, and an image reconstruction unit 153. In FIG. 5, an inverse quantization unit 151 inversely quantizes the two-dimensional quantized DCT coefficient output from the transform unit 13, and an inverse DCT unit 152 performs inverse DCT on the inversely quantized two-dimensional DCT coefficient. Further, the image reconstruction unit 153 reconstructs an image using the pixel values output from the inverse DCT unit 152. At this time, in the case of a non-intra macroblock, the pixel value of the prediction image generated by the prediction unit 12 is added to reconstruct the image.

  The reference image storage unit 16 stores the image reconstructed by the reference image generation unit 15.

  The parameter setting unit 17 sets a quantization parameter used when encoding the next macroblock based on the code amount output from the motion detection / encoding unit 11 and the DCT encoding unit 14.

  The output unit 18 outputs the codes output from the motion detection / encoding unit 11 and the DCT encoding unit 14 as MPEG-4 Visual encoded data.

  Next, the operation of the MPEG-4 Visual encoding apparatus 1 of the present embodiment configured as described above will be described. FIG. 6 is a flowchart for explaining the operation procedure of the MPEG-4 Visual encoding apparatus 1.

  First, in step S10, video data to be encoded is input to the input unit 10.

  In step S20, the MPEG-4 encoding parameter is initialized in the parameter setting unit 17, and in the subsequent step S30, the motion detection / encoding unit 11 divides the moving image data into macroblocks in units of VOPs. A subroutine for motion detection / encoding processing described later is executed.

  Here, the motion detection / encoding processing subroutine of the motion detection / encoding unit 11 in the above description will be described. FIG. 7 is a flowchart for explaining the processing procedure of the motion detection / encoding subroutine.

  First, the motion detection unit 111 shown in FIG. 2 performs motion detection using the reference image stored in the reference image storage unit 16 to calculate a motion vector (step S31).

  Next, the macro block type setting unit 112 sets a macro block type using the result of motion detection (step S32), and determines whether the macro block type is a non-intra macro block (step S33).

  As a result of the determination, if the set macroblock type is a non-intra macroblock, the motion vector encoding unit 113 performs variable length encoding on the motion vector (step S34), and outputs it to the output unit 18 to complete the subroutine processing. (Step S34).

  On the other hand, if it is determined in step S33 that the macroblock type is not a non-intra macroblock, the macroblock data input from the input unit 10 is output to the output unit 18 without being encoded. The subroutine processing is terminated.

  Returning to the flowchart shown in FIG. 6, in step S <b> 40, the prediction unit 12 generates a prediction image using the reference image stored in the reference image storage unit 16 and the motion vector calculated by the motion detection unit 111. .

  In the subsequent step S50, the conversion unit 13 executes a conversion processing subroutine described below. FIG. 8 is a flowchart for explaining the procedure of the conversion processing subroutine in the conversion unit 13.

  First, the prediction error generation unit 131 shown in FIG. 3 generates a prediction error using the macroblock data input from the input unit 10 and the prediction image data generated by the prediction unit 12 (step S51). At this time, if the input macro block data is an intra macro block, the macro block data is used as prediction error data as it is.

  Next, in the DCT unit 132, the prediction error data is subjected to two-dimensional discrete cosine transform (DCT) for each block of 8 × 8 pixels (step S52), and the quantization scale input from the parameter setting unit 17 by the quantization unit 133 is obtained. The two-dimensional DCT coefficient is quantized using (step S53).

  Returning to the flowchart shown in FIG. 6 again, in step S60, the DCT encoding unit 14 executes a subroutine of DCT encoding processing.

  Here, a subroutine of the DCT encoding process in the DCT encoding unit 14 will be described. FIG. 9 is a flowchart for explaining the procedure of the DCT encoding processing subroutine.

  First, the DC / AC prediction unit 141 shown in FIG. 4 performs DC / AC prediction from the two-dimensional quantized DCT coefficient output from the converting unit 13 and the two-dimensional DCT coefficient of the adjacent macroblock, and performs two-dimensional quantization. An error component of the DCT coefficient is calculated (step S61).

  Next, the scanning unit 142 rearranges the error components of the two-dimensional quantized DCT coefficients in the order of encoding (step S62), and the variable length coding unit 143 converts the rearranged two-dimensional quantized DCT coefficient error components. Variable length coding is performed and output to the output unit 18 (step S63).

  Returning to the flowchart shown in FIG. 6, in step S <b> 70, the reference image generation unit 15 executes a subroutine for reference image generation processing.

  Here, a reference image generation processing subroutine in the above-described reference image generation unit 15 will be described. FIG. 10 is a flowchart for explaining the procedure of the reference image generation processing subroutine.

  First, the inverse quantization unit 151 shown in FIG. 5 inversely quantizes the two-dimensional quantized DCT coefficient output from the transform unit 13 (step S71), and the inverse DCT unit 152 inversely quantizes the two-dimensional DCT coefficient. Is subjected to inverse DCT to generate a prediction error of the reference image (step S72).

  Next, the image reconstruction unit 153 adds the predicted image and the prediction error of the reference image to generate a reference image (step S <b> 73) and stores it in the reference image storage unit 16. At this time, when the macroblock data is an intra macroblock, the prediction error of the reference image is stored in the reference image storage unit 16 as it is as a reference image.

  Returning to the flowchart shown in FIG. 6, in step S <b> 80, when the parameter setting unit 17 encodes the next macroblock based on the code amount output from the motion detection / encoding unit 11 and the DCT encoding unit 14. Sets the quantization parameter to be used.

  The series of processing from step S10 to step S80 as described above is repeated for each macroblock, and when the processing is completed for one VOP, the type of the next VOP is set in step S90.

  Then, the processing from step S10 to step S90 is repeated for each VOP, the processing ends for all VOPs, and the processing ends when all input video data is encoded.

  Next, a video recording apparatus according to an embodiment of the present invention that uses the MPEG-4 Visual encoding system described above will be described.

  FIG. 11 is a diagram showing a schematic configuration of the video recording apparatus according to the embodiment of the present invention. In the figure, the video recording apparatus of the present embodiment is configured to include a reading unit 2, a storage unit 3, a re-encoding unit 4, a selection unit 5 and an output unit 6. From the MPEG-4 Visual stream, a part of VOP ( (Hereinafter referred to as a frame) is selected, re-encoded and recorded. In the following description, the MPEG-4 Visual stream profile is described as an example of a simple profile, and I-VOP is described as an intra frame and P-VOP is described as a non-intra frame.

  The reading unit 2 reads video data for each set of frames and sends the video data to the storage unit 3.

  The storage unit 3 includes an encoding type determination unit 31, an intra frame storage unit 32, and a non-intra frame storage unit 33. The encoding type determination unit 31 determines the encoding type of the frame read from the reading unit 2, and if it is an intra frame, sends the frame data to the intra frame storage unit 32, and if it is a non-intra frame, The frame data is sent to the non-intra frame storage unit 33.

  The intra frame storage unit 32 and the non-intra frame storage unit 33 temporarily store the frame data sent from the encoding type determination unit 31. Each of the intra frame storage unit 32 and the non-intra frame storage unit 33 has an area where a predetermined number of frames can be stored, and when frames are stored in all of the areas, new frames are stored. Old data is deleted sequentially to store the frames.

  As shown in FIG. 12, the re-encoding unit 4 includes an intra-frame decoding unit 41, an intermittent video re-encoding unit 42, a feature amount acquisition unit 43, and a video reconstruction unit 44. Further, as illustrated in FIG. 13, the intra frame decoding unit 41 includes a variable length decoding unit 411, an inverse quantization unit 412, an inverse DCT unit 413, and an image reconstruction unit 414.

  In FIG. 13, a variable length decoding unit 411 reads frame data to be recorded from the intra frame storage unit 32 or the non-intra frame storage unit 33 and performs variable length decoding, and an inverse quantization unit 412 performs variable length decoding. Dequantize the data. The inverse DCT unit 413 performs inverse DCT on the inversely quantized data, and the image reconstruction unit 414 reconstructs an image by arranging the inverse DCT data at a correct position in the image data.

  In FIG. 12, the intermittent video re-encoding unit 42 re-encodes the image data decoded by the intra-frame decoding unit 41, sends the encoded information to the feature amount acquisition unit 43, and re-encodes the image data. The image is sent to the video reconstruction unit 44.

  The feature amount acquisition unit 43 calculates an image feature amount from the encoded information sent from the intermittent video re-encoding unit 42.

  The video reconstruction unit 44 selects the video data sent from the intermittent video re-encoding unit 42, the frame data stored in the intra frame storage unit 32, and the frame data stored in the non-intra frame storage unit 33. The encoded data is obtained from the frame data thus reconstructed and sent to the output unit 6.

  The selection unit 5 selects which non-intra frame group to record from the frame data stored in the non-intra frame storage unit 33 based on the image feature amount acquired from the re-encoding unit 4. .

  The output unit 6 outputs the video data re-encoded by the re-encoding unit 4 for recording on a recording medium (not shown).

  Next, the operation of the video recording apparatus of the present embodiment configured as described above will be described. FIG. 14 is a flowchart for explaining an operation procedure of the video recording apparatus according to the present embodiment.

  First, in step S100, the reading unit 2 reads frame data in order of one set of frames from the MPEG-4 Visual stream of video data to be recorded.

  Next, the encoding type information is acquired from the read frame data (step S200), and the encoding type determination unit 31 determines whether the frame is an intra frame (step S300).

  If the result of determination in step S300 is that the read frame is an intra frame, the frame data is stored in the intra frame storage unit 32 (step S400), and if it is a non-intra frame, it is stored in the non-intra frame storage unit 33. Store (step S500).

  In this manner, the processing procedure from step S100 to step S500 is repeated for each frame, and the number of frames necessary for encoding is stored in the respective storage areas of the intra frame storage unit 32 and the non-intra frame storage unit 33. Then, the re-encoding unit 4 acquires the frame to be recorded from the intra-frame storage unit 32 or the non-intra frame storage unit 33 and performs re-encoding (step S600).

  Here, the re-encoding processing subroutine of the re-encoding unit 4 in the above description will be described. FIG. 15 is a flowchart for explaining the procedure of a re-encoding processing subroutine in the re-encoding unit 4.

  First, in step S610, the intra frame stored in the intra frame storage unit 32 is decoded. The procedure of step S610 further constitutes a subroutine as shown in the flowchart of FIG. 16, the variable length decoding unit 411 performs variable length decoding to obtain quantized DCT coefficients (step S611), and the inverse quantization unit 412 To obtain a DCT coefficient by inverse quantization (step S612).

  Next, inverse DCT is performed by the inverse DCT unit 413 (step S613), and a reconstructed image is generated using the pixel data obtained by the image reconstruction unit 414 as pixel data of corresponding coordinates (step S614).

  In step S620 shown in FIG. 15, the image data obtained by decoding according to the above procedure is re-encoded. This re-encoding is performed according to the procedure of the flowchart shown in FIG. 6 using the MPEG-4 Visual encoding apparatus having the configuration shown in FIG. 1, and therefore the description thereof is omitted here. However, the output destination is a memory for temporary storage, and a predetermined number of re-encoded frame data is held.

  In step S630, parameters obtained when the frame data is re-encoded in the procedure of step S620 are acquired, and an image feature amount is calculated. Here, the image feature amount is calculated using a spatial distribution of motion vectors acquired when performing inter-frame prediction, a sum of squares of differences between frames, a code amount, a quantization parameter, and the like. Anything other than those described above may be used as long as it can detect a difference from the immediately preceding frame when abnormal data is captured in the video data captured by the surveillance camera.

  In subsequent step S640, using the acquired image feature amount and the image feature amount acquired up to the time of re-encoding of the immediately preceding frame, a non-intra frame before M sets (M is a natural number) of the re-encoded frame. Determine whether to record the group. If it is determined to be recorded, the intra frame data immediately before the non-intra frame group to be recorded is acquired from the intra frame storage unit 32 (step S650).

  Next, the current non-intra frame data group is acquired from the non-intra frame storage unit 33 (step S660).

  On the other hand, if it is determined in the processing procedure of step S640 that the non-intra frame group is not to be recorded, the re-encoded frame data before M sets is acquired from the intra frame storage unit 32 (step S670).

  The frame data obtained by such a procedure is added to the end of the video data to be recorded (step S680), and the re-encoding processing subroutine is terminated.

  Returning to the flowchart shown in FIG. 14, in step S <b> 700, the selection unit 5 selects which non-intra frame group to record based on the image feature amount acquired at the time of re-encoding.

  In step S800, the video data re-encoded by the re-encoding unit 4 is output by the output unit 6 for recording on a recording medium (not shown), and the process ends.

  As described above, according to such a video recording apparatus according to the embodiment of the present invention, only the intra frame is extracted from the MPEG-4 encoded data and re-encoded, thereby reducing the amount of calculation. Since the amount of data can be greatly reduced, video data can be recorded for a long time at low cost. Further, by determining whether or not the video data includes abnormal data using the re-encoding information, it is possible to easily switch between time-lapse recording and normal recording without complicating the system configuration.

  In the embodiment of the present invention, the simple profile is described as an example of the profile of the MPEG-4 Visual encoded stream, but other profiles can be recorded with the same processing configuration.

  In addition, the present invention relates to MPEG-1, MPEG-2, H.264. Even if an encoding technique other than MPEG-4 such as H.264 is used, the processing configuration is essentially the same as long as it is video data encoded using inter-frame prediction, and can be recorded by similar processing. it can.

  In addition, as the re-encoding process, for example, the motion amount is not detected and the process is performed assuming that the motion vectors are all zero, thereby further reducing the amount of calculation.

  The video recording program of the present invention causes a computer such as a processor (not shown) that controls the video recording apparatus to execute the above-described operation, and is held in a predetermined storage device inside and outside the video recording apparatus. It can also be acquired from another server device or the like via a network.

  Although various embodiments of the present invention have been described, the present invention is not limited to the matters shown in the above-described embodiments, and those skilled in the art may modify or apply the description based on the description of the specification and well-known techniques. The present invention is intended to be included in the scope for which protection is sought.

  The video recording apparatus and the video recording program of the present invention are capable of recording for a long time without depending on the type of video data and the device that encodes the video data, and for detecting an abnormality in the monitoring target area. This is advantageous in that the system configuration can be simplified without requiring devices such as sensors, and is useful for surveillance camera systems and the like.

The figure which shows schematic structure of an MPEG-4Visual encoding apparatus. The figure which shows schematic structure of the motion detection and encoding part of an MPEG-4Visual encoding apparatus. The figure which shows schematic structure of the conversion part of an MPEG-4Visual encoding apparatus The figure which shows schematic structure of the DCT encoding part of an MPEG-4Visual encoding apparatus. The figure which shows schematic structure of the image generation part for a reference of an MPEG-4Visual encoding apparatus. Flow chart for explaining the operation procedure of the MPEG-4 Visual encoding apparatus Flowchart for explaining the processing procedure of the motion detection / coding unit in the MPEG-4 Visual coding apparatus Flowchart for explaining the procedure of the conversion unit in the MPEG-4 Visual encoding apparatus Flowchart for explaining the processing procedure of the DCT encoding unit in the MPEG-4 Visual encoding apparatus Flowchart for explaining the processing procedure of the reference image generation unit in the MPEG-4 Visual encoding device The figure which shows schematic structure of the video recording apparatus of embodiment which concerns on this invention. The figure which shows schematic structure of the re-encoding part of the video recording device which concerns on embodiment The figure which shows schematic structure of the intra-frame decoding part of the video recording device which concerns on embodiment The flowchart for demonstrating the operation | movement procedure of the video recording apparatus of embodiment which concerns on this invention. Flowchart for explaining the processing procedure of the re-encoding unit of the video recording apparatus according to the embodiment The flowchart for demonstrating the process sequence of the intra-frame decoding part in the video recording device which concerns on this embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 MPEG-4Visual encoding apparatus 11 Input part 12 Motion detection and encoding part 13 Prediction part 14 Conversion part 15 DCT encoding part 16 Reference image generation part 17 Reference image memory | storage part 18 Parameter setting part 19 Output part 2 Reading part 3 Storage Unit 31 Encoding Type Determination Unit
32 Intra-frame storage unit 33 Non-intra-frame storage unit 4 Re-encoding unit 41 Intra-frame decoding unit 42 Intermittent video reconstruction unit 43 Feature quantity acquisition unit 44 Video reconstruction unit 5 Selection unit 6 Output unit

Claims (6)

A frame image is encoded by selecting one of a first encoding method for encoding within a frame and a second encoding method for encoding based on the correlation between the frames, in units of frames. In order to record in order,
A reading unit for sequentially reading the encoded frame images in units of frames;
A first storage unit for temporarily storing a frame image encoded by the first encoding method;
i-th (i is a natural number) and (i + 1) -th encoded by the second encoding method of L (L ≧ 1) between the frame images encoded by the first encoding method A second storage unit that temporarily stores frame images as a group i frame image group;
First selection for selecting the frame image group of the j-th set (j is a natural number) that satisfies a predetermined condition and the M groups (M is a natural number) before and after satisfying a predetermined condition from the group of frame images stored in the second storage unit And
A first re-encoding unit that re-encodes all the frame images stored in the first storage unit and a frame image group selected by the first selection unit;
A recording unit for recording the frame image re-encoded by the first re-encoding unit;
A video recording apparatus comprising: The video recording apparatus according to claim 1,
The first re-encoding unit includes:
A decoding unit for decoding the frame image stored in the first storage unit;
A second re-encoding unit that re-encodes the frame image decoded by the decoding unit using the second encoding method;
A second selection unit that selects one of the frame image re-encoded by the second re-encoding unit and the frame image stored in the first storage unit;
A rearrangement unit that rearranges the frame image selected by the second selection unit and the frame image group selected by the first selection unit in time order;
A video recording apparatus comprising: The video recording apparatus according to claim 1 or 2,
The predetermined condition is a feature amount calculated based on encoding information acquired at the time of re-encoding and a feature amount calculated from a frame image re-encoded immediately before the frame image to be re-encoded. A video recording apparatus in which the difference is greater than or equal to a predetermined range. A frame image is encoded by selecting one of a first encoding method for encoding within a frame and a second encoding method for encoding based on the correlation between the frames, in units of frames. A video recording program for causing a computer to execute the function of recording in order,
A function of sequentially reading the encoded frame images frame by frame;
A function of temporarily storing a frame image encoded by the first encoding method in a first storage unit;
i-th (i is a natural number) and (i + 1) -th encoded by the second encoding method of L (L ≧ 1) between the frame images encoded by the first encoding method A function of temporarily storing frame images as a group i frame image group in the second storage unit;
A function of selecting a set of frame images satisfying a predetermined condition among the group of frame images stored in the second storage unit (j is a natural number) and each of the preceding and subsequent M sets of frame images (M is a natural number);
A function of re-encoding all the frame images stored in the first storage unit and the selected frame image group;
A function of recording the re-encoded frame image;
A video recording program that causes a computer to execute. The video recording program according to claim 4,
The re-encoding function is:
A function of decoding the frame image stored in the first storage unit;
A function of re-encoding the decoded frame image by the second encoding method;
A function of selecting one of the frame image re-encoded by the second encoding method and the frame image stored in the first storage unit;
A function of rearranging the selected frame image and the selected frame image group in time order;
Including video recording program. The video recording program according to claim 4 or 5,
The predetermined condition is a feature amount calculated based on encoding information acquired at the time of re-encoding and a feature amount calculated from a frame image re-encoded immediately before the frame image to be re-encoded. A video recording program in which the difference is greater than or equal to a predetermined range.

Images