JP2012109846A - Image encoder and camera system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To eliminate the need of a repetitive compression processing, to suppress an increase of memory capacity and to efficiently suppress an excessive increase of a generated code amount when input image data suddenly changes while occurrence of block noise and mosquito noise due to excessive compression is suppressed.SOLUTION: An image encoder includes: a DCT section 8; a quantization section 9; an encoding section 10; a quantization coefficient calculation necessity determination section 2 determining that calculation of a quantization coefficient is required when previous encoding information shows possibility of the increase of the generated code amount; a DC component acquiring section 3 acquiring a DC component generated from a luminance/color difference signal when the calculation of the quantization coefficient is required; a DC component preserving section 5 preserving the DC components of previous and this input image data; an image change operation section 6 determining a change of input image data by comparison of the DC components of input image data and the previous image data and generating an image change amount; and a quantization coefficient calculating section 7 calculating the quantization coefficient for encoding based on the obtained image change amount and transferring the coefficient to the quantization section.

Description

  The present invention relates to an image encoding apparatus including an image input unit, a DCT unit (discrete cosine transform unit), a quantization unit, and an encoding unit, and more specifically, when input image data such as a scene change suddenly changes The present invention relates to a technique for suppressing an excessive increase in code amount. Applicable products include digital still cameras, digital video cameras, in-vehicle cameras, and network cameras.

  In devices equipped with an imaging function, such as digital video cameras, digital still cameras, mobile phone terminals with cameras, and portable electronic notebooks with cameras, the amount of video and still image data obtained by imaging is large, so it is recorded as is. Then, a recording medium with a limited capacity will soon fill up the recording capacity. For this reason, recording is generally performed after reducing the amount of data by compressing by various methods. For example, as a compression method, a moving picture experts group (MPEG) method or the like is used for moving image data, and a joint photographic experts group (JPEG) method or the like is used for still image data. In a conventional image compression processing apparatus, an image is compressed step by step to a predetermined data size.

JP 2007-116633 A

  However, in order to obtain an appropriate compression rate, if a plurality of compression processes including retry are performed, the required time becomes long and the imaging interval cannot be set short. In addition, repeated compression processing requires a large-capacity memory for holding all data of the original image.

  If the compression rate is set high to suppress the data size to a predetermined value by a single compression process, block noise and mosquito noise are generated, and image quality is degraded.

  Also, if the input image data such as a scene change changes suddenly and there is no correlation between the previously compressed image data and the currently compressed image data, the amount of code after compression increases excessively. If the code amount exceeds the upper limit of the standard, the frame must be discarded and the encoding process must be performed again.

  Patent Document 1 discloses a technique for suppressing a code amount at the time of image encoding based on a camera shake detection amount obtained by a camera shake detection unit, but this is a different approach from the present invention. It is.

  The present invention has been created in view of such circumstances, eliminating the need for repeated compression processing, suppressing an increase in memory capacity, and suppressing occurrence of block noise and mosquito noise due to excessive compression. An object of the present invention is to efficiently suppress an excessive increase in the amount of generated codes when image data changes rapidly.

  The present invention solves the above problems by taking the following measures.

  << 1 >> An image encoding apparatus according to the present invention is premised on an image input unit for inputting image data from the outside, and a DCT unit for generating a DC component (DC component) by inputting a luminance / color difference signal from the image input unit. And a quantization unit that inputs and quantizes a DC component by the DCT unit, and an encoding unit that encodes information quantized by the quantization unit.

  In addition to the above premise configuration, the image coding apparatus according to the present invention includes a quantization coefficient calculation necessity determination unit, a DC component acquisition unit, a DC component storage unit, an image change calculation unit, and a quantum, each configured as follows. A conversion coefficient calculation unit is provided.

  The quantization coefficient calculation necessity determination unit is configured to determine whether or not to perform quantization coefficient calculation by the quantization coefficient calculation unit based on the previous encoded information stored. The quantization coefficient calculation necessity determination unit determines that the quantization coefficient calculation is necessary when the previous coding information indicates that the generated code amount may increase, and the DC component acquisition unit and the DC component storage , The image change calculation unit, and the quantization coefficient calculation unit are activated. Conversely, if the previous coding information does not indicate the possibility of increase in the amount of generated code, it is determined that the quantization coefficient calculation is unnecessary, and at this time, the DC component acquisition unit and the like are left in an inactive state. The

  When activated, the DC component acquisition unit is configured to acquire a DC component generated from the luminance / color difference signal from the image input unit and store the acquired DC component in the DC component storage unit. Yes.

  The DC component storage unit is configured to store the DC component acquired by the DC component acquisition unit for each input image data. Here, the DC component of the previous input image data already encoded and the DC component of the current input image data are stored.

  The image change calculation unit obtains the DC component of the input image data from the DC component storage unit and the DC component of the previous input image data that has already been encoded, and generates an image change amount by comparing both DC components. The image change amount is passed to the quantization coefficient calculation unit.

  The quantization coefficient calculation unit calculates a quantization coefficient for encoding based on the image change amount obtained by the image change calculation unit, and passes the calculated quantization coefficient to the quantization unit.

  The quantization unit receives the quantization coefficient from the quantization coefficient calculation unit, and quantizes the DC component input from the DCT unit with the received quantization coefficient.

  The effect | action by this structure is as follows. The quantization coefficient calculation necessity determination unit determines whether or not to perform the quantization coefficient calculation by the quantization coefficient calculation unit based on the previous encoded information stored. That is, when the previous coding information indicates that the generated code amount may increase, it is determined that the quantization coefficient calculation is necessary, and the DC component acquisition unit, the DC component storage unit, the image change calculation unit, the quantization Activate the coefficient calculator. Conversely, if the previous coding information does not indicate the possibility of increase in the amount of generated code, it is determined that the quantization coefficient calculation is unnecessary, and at this time, the DC component acquisition unit and the like are left in an inactive state. The The activated DC component acquisition unit acquires the DC component generated from the luminance / color difference signal from the image input unit, and stores the acquired DC component in the DC component storage unit. The DC component storage unit stores the previously encoded DC component of the previous input image data and the DC component of the current input image data. The image change calculation unit obtains the DC component of the input image data from the DC component storage unit and the DC component of the previous input image data already encoded, generates an image change amount by comparing both DC components, To the conversion factor calculation unit. The quantization coefficient calculation unit calculates a quantization coefficient for encoding based on the image change amount obtained by the image change calculation unit, and passes the calculated quantization coefficient to the quantization unit. The DCT unit receives the luminance / color difference signal from the image input unit, generates a DC component, and outputs the DC component to the quantization unit. The quantization unit receives the DC component from the DCT unit, receives the quantization coefficient from the quantization coefficient calculation unit, and quantizes the DC component with the received quantization coefficient. The encoding unit encodes the information quantized by the quantization unit. Then, the encoded information is passed to the quantization coefficient calculation necessity determination unit.

  One of the points here is determination of necessity of quantization coefficient calculation by the quantization coefficient calculation necessity determination unit. When the previous encoding information received from the encoding unit indicates that the generated code amount may increase, the quantization coefficient calculation necessity determination unit determines that the quantization coefficient calculation is necessary. Otherwise, it is determined that the quantization coefficient calculation is unnecessary.

  The fact that the previous encoded information indicates the possibility of increase in the amount of generated code means that, for example, the previously encoded quantization coefficient is small below a certain level. When the quantization coefficient is small, the generated code amount increases. When the quantization coefficient is small and the generated code amount may increase, the quantization coefficient calculation necessity determination unit includes a DC component acquisition unit, a DC component storage unit, an image change calculation to suppress the generated code amount. And the quantization coefficient calculation unit are activated. As a result, a quantization coefficient is generated according to an image change amount that is a comparison result between the DC component of the input image data and the DC component of the previous input image data that has already been encoded. This is another point.

  The larger the image change amount, the larger the quantization coefficient, and as a result, the generated code amount is sufficiently suppressed. When the image change amount is small, the quantization coefficient may be reduced, and as a result, the generated code amount may be suppressed on a small scale.

  As described above, when the input image data changes abruptly, the quantization coefficient is increased according to the scale of the change, so that it is possible to efficiently suppress an excessive increase in the generated code amount.

  <2> The means for generating the DC component acquired by the DC component acquisition unit in the configuration of <1> is configured separately from the DCT unit, and is based on the luminance / color difference signal from the image input unit. There is an aspect in which a DC component generation unit that generates components is provided. The DC component generation unit is configured to generate a DC component from the luminance / color difference signal from the image input unit and pass it to the DC component acquisition unit when the DC component acquisition unit is activated.

  << 3 >> Similarly, as a means for generating the DC component acquired by the DC component acquisition section in the configuration of <1>, there is an aspect in which the DC component generation means in the DCT section is used.

  When the dedicated DC component generation unit different from the general DCT unit as in << 2 >> is provided, the DC component generation unit is dedicated, so the DC component acquisition unit, DC Each processing in the component storage unit, the image change calculation unit, and the quantization coefficient calculation unit is speeded up. As a result, it is possible to accurately cope with a sudden change in the input image data.

  On the other hand, in the case of the configuration in which the DC component obtained by the general DCT unit is passed to the DC component acquisition unit as in <3> above, a dedicated DC component generation unit is unnecessary because of the dual configuration, and the cost increase is suppressed. This is advantageous.

  According to the present invention, when input image data such as a scene change suddenly changes, the quantization coefficient is increased in real time according to the scale of the change, so that it is not necessary to perform repeated compression processing and increase the memory capacity. It is possible to efficiently suppress an excessive increase in the amount of generated codes while suppressing generation of block noise and mosquito noise due to excessive compression.

1 is a block diagram showing a configuration of a camera system in which an image encoding device according to Embodiment 1 of the present invention is incorporated. The flowchart which shows operation | movement of the image coding apparatus of Example 1 of this invention. Explanatory drawing of calculation of DC component difference amount of input image data and comparison image data in the Example of this invention The graph which calculates a quantization coefficient from the DC component difference amount in the Example of this invention FIG. 5 is a block diagram illustrating a configuration of a camera system in which an image encoding device according to a second embodiment of the present invention is incorporated. The flowchart which shows operation | movement of the image coding apparatus of Example 2 of this invention. The flowchart which shows operation | movement of the image coding apparatus of Example 3 of this invention.

  The image encoding apparatus of the present invention having the above configuration <1> can be further advantageously developed in the following embodiment.

  << 4 >> For the quantization coefficient calculation unit in the configurations of the above << 1 >> to << 3 >>, the DC component difference amount that is the difference between the two DC components is used as the image change amount obtained by the image change calculation unit. There is a mode in which a quantization coefficient for encoding is calculated and generated based on the DC component difference amount. This is one of the simplest methods for calculating the quantization coefficient.

  << 5 >> About the quantization coefficient calculation necessity determination part in the configuration of the above << 1 >> to << 4 >>, using the quantization coefficient as the previous encoding information, the quantization coefficient is more than a specific threshold value. When it is large, there is an aspect in which it is configured to determine that calculation of the quantization coefficient is unnecessary. The fact that the quantization coefficient is larger than a certain value means that there is a margin in terms of the amount of generated code. That is, there is no need to suppress the amount of generated codes, and in this case, the quantization coefficient calculation unit disables the quantization coefficient calculation. That is, even when the input image data changes rapidly, the quantization coefficient need not be updated if the quantization coefficient is large and the generated code amount has a margin.

  <6> The DC component acquisition unit in the configurations of <1> to <5> is configured to acquire the DC component of the input image data in units of macroblocks.

  <7> Similarly, the DC component acquisition unit in the configurations of the above << 1 >> to << 5 >> may be configured to acquire the DC component of the input image data in units of pictures.

  <8> Similarly, the DC component acquisition unit in the configurations of << 1 >> to << 5 >> may be configured to acquire the DC component of the input image data in units of DCT execution blocks.

  As in the above << 6 >> to << 8 >>, the mode of the DC component acquisition unit with respect to what unit the DC component of the input image data is acquired, in units of macroblocks, pictures, or DCT execution blocks But either is fine. The input image data changes abruptly by increasing the quantization coefficient as the DC component difference amount between the input image data acquired in each unit and the DC component of the previous input image data already encoded increases. Even in this case, it is possible to accurately achieve the effect of sufficiently suppressing the generated code amount.

  <9> With respect to the DC component storage unit in the configurations of the above <1> to <8>, the DC component of the previous input image data that has already been encoded as the DC component to be compared with the DC component of the input image data There is an aspect of being configured to set.

  << 10 >> For the DC component storage unit in the same configurations of << 1 >> to << 8 >>, an average value of DC components for an arbitrary frame is set as the DC component to be compared with the DC component of the input image data. There is also an aspect of being configured to.

  As in the above << 9 >> and << 10 >>, the DC component storage unit regarding the type of DC component to be compared with the DC component to be compared with the input image data is previously encoded. The input image data may be a DC component, or an average value of DC components for an arbitrary frame. The larger the DC component difference amount, which is the difference between the DC component of the input image data and the previously encoded DC component, the larger the quantization coefficient, so that the generated code amount is sufficient even when the input image data changes suddenly. Therefore, it is possible to accurately achieve the effect of suppressing the movement. When the average value is used, the calculation of the difference between the DC component of the input image data and the DC component encoded last time becomes stable.

  <11> In the configuration of <6> in which the DC component acquisition unit acquires the DC component of the input image data in units of macroblocks, the image change calculation unit should be compared with the DC component of the input image data. There is a mode in which the unit of the DC component is made into a macroblock unit in accordance with the DC component acquisition unit.

  << 12 >> In addition, in the configuration of << 7 >> in which the DC component acquisition unit acquires the DC component of the input image data in units of pictures, the image change calculation unit is compared with the DC component of the input image data. There is a mode in which the unit of the power DC component is set to the picture unit in accordance with the DC component acquisition unit.

  << 13 >> Also, in the configuration of <8> above, in which the DC component acquisition unit acquires the DC component of the input image data in units of DCT execution blocks, the image change calculation unit is compared with the DC component of the input image data. There is a mode in which the unit of the DC component to be performed is made into a DCT execution block unit in accordance with the DC component acquisition unit.

  As in the above << 11 >> to << 13 >>, the DC component acquisition is performed for each aspect of the image change calculation unit regarding the unit of the DC component to be compared with the DC component of the input image data. In accordance with the section, the macro block unit, the picture unit, or the DCT execution block unit can accurately achieve the effect of sufficiently suppressing the generated code amount even when the input image data changes suddenly. Become.

  << 14 >> The invention relating to the image coding apparatus according to the above << 1 >> to << 13 >> can be developed as follows in a camera system.

  That is, a camera system according to the present invention includes an imaging device that converts an imaging signal into a luminance / color difference signal, and the above << 1 >> to << 13 >> in which the luminance / color difference signal is input from the imaging device and image coding processing is performed. One of the image encoding devices is provided.

  In the camera system, when the input image data changes abruptly, the quantization coefficient is increased according to the scale of the change, so that it is possible to suppress an excessive increase in the generated code amount.

  Embodiments of an image encoding apparatus according to the present invention will be described below with reference to the drawings. In the following embodiments, components having functions similar to those of the other embodiments are denoted by the same reference numerals, and description thereof is omitted.

Example 1
In the first embodiment, in order to suppress the generated code amount in a sudden scene change, the DC component of the input image data is acquired before encoding, and is compared with the DC component of the previous input image data that has already been encoded. An example in which the amount of code is suppressed by determining the amount of change and calculating the quantization coefficient for encoding based on the amount of change of the DC component will be described. In this embodiment, a dedicated DC component generation unit is provided, and the DC component of the input image data can be acquired before encoding.

  FIG. 1 is a block diagram showing a configuration of a camera system in which an image encoding device according to Embodiment 1 of the present invention is incorporated. In FIG. 1, 20 is an image encoding device, 1 is an image input unit that inputs luminance / color difference signals from an imaging device 11 outside the image encoding device 20, and 2 is obtained from the encoding unit 10 in the previous encoding. The quantization coefficient calculation necessity determination unit 3 for determining whether or not to perform the quantization coefficient calculation according to the coding information, and 3 to the DC component generation unit 4 for luminance / color difference of the input image data A DC component acquisition unit that outputs a signal and acquires a DC component An of input image data, 4 is a DC component generation unit that generates a DC component based on a luminance / color difference signal from the image input unit 1, and 5 is a DC component acquisition A DC component storage unit 6 that stores the DC component An of the input image data acquired from the unit 3 and outputs the DC component An-1 of the previous input image data that has already been encoded and the DC component An of the input image data; Input image from DC component storage 5 The DC component An of the data and the DC component An-1 of the previous input image data already encoded as the comparison target are acquired, and the DC component difference amount ΔAn (= An−An−) of the two image data is obtained as the image change amount. 1) an image change calculation unit that calculates 1), 7 obtains a DC component difference amount ΔAn from the image change calculation unit 6, and calculates a quantization coefficient qn from the DC component difference amount ΔAn, 8 is a quantization A DCT unit that receives a DCT request from the coefficient calculation necessity determination unit 2, acquires a luminance / color difference signal from the image input unit 1, and generates a DC component An of the input image data, and 9 is a DC component An generated by the DCT unit 8. Is quantized based on the quantized coefficient qn calculated by the quantized coefficient calculating unit 7, 10 obtains quantization information from the quantizing unit 9, performs encoding, and converts the encoded data into an image processing apparatus 12 and the mark to be output to the recording medium 13 Unit, 11 image pickup device 12 is an image processing apparatus, 13 is a recording medium. The image encoding device 20 includes a group of components indicated by reference numerals 1 to 10. The camera system 30 includes an imaging device 11 and an image encoding device 20.

  Next, the operation of the image coding apparatus of the present embodiment configured as described above will be described with reference to the flowchart of FIG.

  In step S <b> 1, the image input unit 1 inputs luminance / color difference information in the input image data from the imaging device 11.

  Next, in step S2, the quantization coefficient calculation necessity determination unit 2 determines whether the previously encoded quantization coefficient qn-1 is equal to or less than a predetermined threshold value. It is determined that the amount may increase, and the process proceeds to step S3. Otherwise, it is determined that the generated code amount does not increase, and the process proceeds to step S14. When the process proceeds to step S14, the DCT unit 8 performs DCT processing of the input image data as usual, then performs quantization processing in step S15, further performs encoding processing in step S16, and ends the processing.

  In step S3, when it is determined that there is a possibility that the generated code amount may increase when the quantization coefficient qn-1 is equal to or smaller than the predetermined threshold value in step S2, the DC component acquisition unit 3 goes to the DC component generation unit 4. A luminance / color difference signal of the input image data is output, and a DC component obtained as a result of DCT is acquired.

  Next, in step S4, the DC component acquisition unit 3 determines a DC component acquisition unit. If the DC component acquisition unit is a DCT execution block unit, the luminance / color difference signal is output to the DC component generation unit 4 in step S5. The DC component An of the input image data is acquired in units of DCT execution blocks. If the DC component acquisition unit is a macroblock unit, a luminance / color difference signal is output to the DC component generation unit 4 in step S6, and the DC component An of the input image data is acquired in macroblock units. If the DC component acquisition unit is a picture unit, a luminance / color difference signal is output to the DC component generation unit 4 in step S7, and the DC component An of the input image data is acquired for each picture.

  In subsequent step S8, the DC component storage unit 5 sets the DC component An-1 of the previous input image data that has already been encoded as the DC component to be compared with the DC component An of the input image data.

  Next, in step S9, the image change calculation unit 6 determines a “unit for comparing DC components” (DC component comparison unit). If the unit is a DCT execution block unit, the process proceeds to step S10.

  In step S10, the difference between the DC component An of the input image data and the DC component An-1 of the previous input image data that has already been encoded is taken in units of DCT execution blocks, and the accumulated difference is already encoded with the input image data. The DC component difference amount ΔAn, which is the difference between the DC components of the previous input image data, is set, and the process proceeds to step S13.

  If the DC component comparison unit is a macroblock unit, in step S11, the difference between the DC component An of the input image data and the DC component An-1 of the previous input image data that has already been encoded is taken in macroblock units, and the difference is calculated. Is the DC component difference amount ΔAn, and the process proceeds to step S13.

  If the DC component comparison unit is a picture unit, in step S12, the difference between the DC component An of the input image data and the DC component An-1 of the previous input image data that has already been encoded is taken in units of pictures, and the difference is accumulated. Is the DC component difference amount ΔAn, and the process proceeds to step S13.

  Next, in step S13, the quantization coefficient calculation unit 7 determines the DC component difference amount ΔAn, determines that the scene change has occurred when the DC component difference amount ΔAn is large, and increases the quantization coefficient qn to be encoded. When the DC component difference amount ΔAn is small, it is determined that the scene has not changed, and the quantization coefficient qn is calculated from the DC component difference amount ΔAn under the condition that the quantization coefficient qn to be encoded is not increased. Proceed to step S15.

  Next, in step S15, the quantization unit 9 quantizes the input image data based on the calculated quantization coefficient qn.

  Next, in step S <b> 16, the encoding unit 10 performs encoding based on the quantization information input from the quantization unit 9.

  Next, a method for calculating the DC component difference amount ΔAn between the input image data and the comparison image data will be described with reference to FIG.

  When the DC component comparison unit of the input image data and the comparison image data is (a) to (r) of the DCT execution block unit, the input image data (a) and the comparison image data (b) that are paired in the DCT execution block unit. Similarly, the difference absolute value of the DC component between the input image data (c) and the comparison image data (d), which are paired with each other, is compared with the input image data (e). DC component absolute difference between image data (f), DC component absolute difference between input image data (g) and comparison image data (h), input image data (i) and comparison image data (j) DC component difference absolute value, input image data (k) and comparison image data (l) DC component difference absolute value, input image data (m) and comparison image data (n) DC component absolute difference Value, input image data (o) The difference absolute value of the DC component between the comparison image data (p) and the difference absolute value of the DC component between the input image data (q) and the comparison image data (r) are calculated, and the total of these difference absolute values is calculated as the input image. The DC component difference amount ΔAn between the data and the comparison image data is used.

  Next, a method for calculating the quantization coefficient qn will be described with reference to FIG.

  A threshold value A0 to Ak is given to the DC component difference amount ΔAn between the input image data and the comparison image data, and if the DC component difference amount ΔAn is within the threshold value, the quantization is compared with the prepared threshold value. Coefficients B0 to Bn of the coefficient qn are set. As the DC component difference amount ΔAn is larger, it can be considered that the scene change amount is larger. Therefore, the generated code amount can be suppressed by increasing the value of the quantization coefficient qn according to the magnitude of the DC component difference amount ΔAn. It becomes.

  As described above, according to the present embodiment, when the input image data changes suddenly as in a scene change, the quantization coefficient is increased according to the scale of the change. It can be suppressed efficiently. In addition, since the dedicated DC component generation unit 4 independent from the general DCT unit 8 is provided, the processing of each unit is speeded up, and it is possible to accurately cope with a sudden change in input image data.

(Example 2)
FIG. 5 is a block diagram showing a configuration of a camera system in which an image encoding device according to the second embodiment of the present invention is incorporated. In the second embodiment, the DC component generation unit 4 is deleted from the configuration of FIG. 1 in order to reduce the system cost. Instead, the DC component supplied to the DC component acquisition unit 5 is sent from the DCT unit 8.

  The operation of the image coding apparatus according to the present embodiment will be described with reference to the flowchart of FIG.

  In step S <b> 21, the image input unit 1 inputs luminance / color difference information in the input image data from the imaging device 11.

  Next, in step S22, the quantization coefficient calculation necessity determination unit 2 determines whether the previously encoded quantization coefficient qn-1 is equal to or less than a predetermined threshold value. It is determined that the amount may increase, and the process proceeds to step S23. Otherwise, it is determined that the generated code amount does not increase, and the process proceeds to step S34. When the process proceeds to step S34, the DCT unit 8 performs DCT processing of the input image data as usual, then performs quantization processing in step S35, further performs encoding processing in step S36, and ends the processing.

  In step S23, the DC component acquisition unit 3 requests the DCT unit 8 to send a DCT request in step S23, which proceeds after determining that there is a possibility that the generated code amount may increase when the quantization coefficient qn-1 is equal to or smaller than the predetermined threshold value in step S22. To obtain the DC component obtained as a result of DCT. This step S23 is different from step S3 in FIG. This is because the DC component generation unit 4 in FIG. 1 is not in the present embodiment, and instead uses the DCT unit 8.

  Next, in step S24, the DC component acquisition unit 3 determines a DC component acquisition unit. If the DC component acquisition unit is a macroblock unit, the DC component An of the input image data is macro-converted from the DC component acquisition unit 3 in step S26. Get in blocks. If the DC component acquisition unit is a DCT execution block unit, the DC component An of the input image data is acquired from the DC component acquisition unit 3 in DCT execution block units in step S25. If the DC component acquisition unit is a picture unit, in step S27, the DC component An of the input image data is acquired from the DC component acquisition unit 3 in picture units.

  In subsequent step S28, the DC component storage unit 5 sets the previously encoded DC component An-1 of the previous input image data as the DC component to be compared with the DC component An of the input image data.

  Next, in step S29, the image change calculation unit 6 determines a DC component comparison unit. If the unit is a macroblock unit, the process proceeds to step S31.

  In step S31, the difference between the DC component An of the input image data and the DC component An-1 of the previous input image data that has already been encoded is taken in macroblock units, and the accumulated difference is already encoded with the input image data. The DC component difference amount ΔAn that is the difference between the DC components of the previous input image data is set, and the process proceeds to step S33.

  If the DC component comparison unit is a DCT execution block unit, in step S30, the difference between the DC component An of the input image data and the DC component An-1 of the previous input image data that has already been encoded is calculated in units of the DCT execution block. The accumulated difference is set as the DC component difference amount ΔAn, and the process proceeds to step S33.

  If the DC component comparison unit is a picture unit, in step S32, the difference between the DC component An of the input image data and the DC component An-1 of the previous input image data that has already been encoded is taken in units of pictures, and the difference is accumulated. Is the DC component difference amount ΔAn, and the process proceeds to step S33.

  Next, in step S33, the quantization coefficient calculation unit 7 determines the DC component difference amount ΔAn, determines that the scene has changed when the DC component difference amount ΔAn is large, and increases the quantization coefficient qn to be encoded. When the DC component difference amount ΔAn is small, it is determined that the scene has not changed, and the quantization coefficient qn is calculated from the DC component difference amount ΔAn under the condition that the quantization coefficient qn to be encoded is not increased. Proceed to step S35.

  Next, in step S35, the quantization unit 9 quantizes the input image data based on the calculated quantization coefficient qn.

  Next, in step S <b> 36, the encoding unit 10 performs encoding based on the quantization information input from the quantization unit 9.

  As described above, according to the present embodiment, when the input image data changes suddenly as in a scene change, the quantization coefficient is increased according to the scale of the change as in the first embodiment. An excessive increase in the amount can be efficiently suppressed. In addition, since the DC component obtained by the general DCT unit 8 is configured to be passed to the DC component acquisition unit 3, the dedicated DC component generation unit 4 in the case of the first embodiment is not necessary. Cost increase can be suppressed.

(Example 3)
FIG. 7 is a flowchart showing the operation of the image coding apparatus according to the third embodiment of the present invention. In this embodiment, an average value of DC components for an arbitrary frame is set as a DC component to be compared. This is different from the first and second embodiments.

  In step S <b> 41, the image input unit 1 inputs luminance / color difference information in the input image data from the imaging device 11.

  Next, in step S42, the quantization coefficient calculation necessity determination unit 2 determines whether the previously encoded quantization coefficient qn-1 is equal to or less than a predetermined threshold value. It is determined that the amount may increase, and the process proceeds to step S43. Otherwise, it is determined that the generated code amount does not increase, and the process proceeds to step S54. When the process proceeds to step S54, the DCT unit 8 performs DCT processing of the input image data as usual, then performs quantization processing in step S55, further performs encoding processing in step S56, and ends the processing.

  In step S43, the DC component acquisition unit 3 sends a DCT request to the DCT unit 8 in step S43, which proceeds after determining that there is a possibility that the generated code amount may increase when the quantization coefficient qn-1 is equal to or smaller than a predetermined threshold value in step S42. To obtain the DC component obtained as a result of DCT.

  Next, in step S44, the DC component acquisition unit 3 determines the DC component acquisition unit. If the DC component acquisition unit is a picture unit, the DC component An of the input image data is converted from the DC component acquisition unit 3 to the picture unit in step S47. Get in. If the DC component acquisition unit is a DCT execution block unit, the DC component An of the input image data is acquired from the DC component acquisition unit 3 in the DCT execution block unit in step S45. If the DC component acquisition unit is a macroblock unit, the DC component An of the input image data is acquired from the DC component acquisition unit 3 in macroblock units in step S46.

  In subsequent step S48, the DC component storage unit 5 sets an average value of DC components for an arbitrary frame as a DC component to be compared with the DC component An of the input image data.

  Next, in step S49, the image change calculation unit 6 determines a DC component comparison unit. If the unit is a picture unit, the process proceeds to step S52.

  In step S52, the difference between the DC component An of the input image data and the DC component An-1 of the previous input image data that has already been encoded is taken in picture units, and the accumulated difference is already encoded with the input image data. The DC component difference amount ΔAn that is the difference between the DC components of the previous input image data is set, and the process proceeds to step S53.

  If the DC component comparison unit is a DCT execution block unit, in step S50, the difference between the DC component An of the input image data and the DC component An-1 of the previous input image data that has been encoded is calculated in units of the DCT execution block. The accumulated difference is set as the DC component difference amount ΔAn, and the process proceeds to step S53.

  If the DC component comparison unit is a macroblock unit, in step S51, the difference between the DC component An of the input image data and the DC component An-1 of the previous input image data that has already been encoded is calculated in macroblock units. Is the DC component difference amount ΔAn, and the process proceeds to step S53.

  Next, in step S53, the quantization coefficient calculation unit 7 determines the DC component difference amount ΔAn, determines that the scene has changed when the DC component difference amount ΔAn is large, and increases the quantization coefficient qn to be encoded. When the DC component difference amount ΔAn is small, it is determined that the scene has not changed, and the quantization coefficient qn is calculated from the DC component difference amount ΔAn under the condition that the quantization coefficient qn to be encoded is not increased. Proceed to step S55.

  Next, in step S55, the quantization unit 9 quantizes the input image data based on the calculated quantization coefficient qn.

  In step S56, the encoding unit 10 performs encoding based on the quantization information input from the quantization unit 9.

  As described above, according to the present embodiment, since the average value is adopted, even if the value of the DC component An of the input image data becomes abnormal, the DC component An of the input image data and the DC component previously encoded are used. Can be calculated stably.

  The technology of the image coding apparatus of the present invention increases the quantization coefficient in real time according to the scale of the change even when the input image data such as a scene change suddenly changes, thereby eliminating the need for repeated compression processing, It suppresses an increase in memory capacity, suppresses the generation of block noise and mosquito noise due to excessive compression, and efficiently suppresses an excessive increase in the amount of generated codes. For example, digital video cameras, digital still cameras, cameras It is useful in devices having an imaging function such as a mobile phone terminal with a camera and a portable electronic notebook with a camera.

DESCRIPTION OF SYMBOLS 1 Image input part 2 Quantization coefficient calculation necessity judgment part 3 DC component acquisition part 4 DC component generation part 5 DC component preservation | save part 6 Image change calculation part 7 Quantization coefficient calculation part 8 DCT part 9 Quantization part 10 Encoding part DESCRIPTION OF SYMBOLS 11 Imaging device 12 Image processing device 13 Recording medium 20 Image encoding device 30 Camera system

Claims (14)

An image input unit for inputting image data from the outside;
A DCT unit for inputting a luminance / color difference signal from the image input unit to generate a DC component;
A quantization unit that quantizes a DC component input from the DCT unit using an input quantization coefficient;
An encoding unit for encoding the information quantized by the quantization unit;
When the previous encoded information received from the encoding unit indicates that the generated code amount may increase, it is determined that quantization coefficient calculation is necessary, and the previous encoded information increases the generated code amount. A quantization coefficient calculation necessity determination unit that determines that the calculation of the quantization coefficient is unnecessary when the possibility is not indicated;
When the necessity determination of the quantization coefficient calculation is required,
A DC component acquisition unit for acquiring a DC component generated from a luminance / color difference signal from the image input unit;
A DC component storage unit that stores the DC component of the previous input image data encoded by the DC component acquisition unit and the DC component of the current input image data;
An image change calculation unit that obtains the DC component of the input image data and the DC component of the previous input image data that has already been encoded from the DC component storage unit, and generates an image change amount by comparing both DC components;
An image code comprising: a quantization coefficient for encoding based on an image change amount obtained by the image change calculation unit; and a quantization coefficient calculation unit that passes the calculated quantization coefficient to the quantization unit Device.   As a means for generating the DC component acquired by the DC component acquisition unit, a DC component generation unit that is configured separately from the DCT unit and generates the DC component based on a luminance / color difference signal from the image input unit. The image encoding device according to claim 1, comprising:   The image coding apparatus according to claim 1, wherein a DC component generation unit in the DCT unit is used as the unit that generates the DC component acquired by the DC component acquisition unit.   The quantization coefficient calculation unit generates a DC component difference amount that is a difference between the two DC components as the image change amount obtained by the image change calculation unit, and performs encoding for encoding based on the DC component difference amount. The image coding device according to any one of claims 1 to 3, wherein the image coding device is configured to calculate a quantization coefficient.   5. The quantization coefficient calculation necessity determination unit is configured to determine that calculation of a quantization coefficient is unnecessary when a previously encoded quantization coefficient is greater than a specific threshold value. The image encoding device according to any one of the above.   The image coding apparatus according to any one of claims 1 to 5, wherein the DC component acquisition unit is configured to acquire a DC component of input image data in units of macroblocks.   The image encoding device according to any one of claims 1 to 5, wherein the DC component acquisition unit is configured to acquire a DC component of input image data in units of pictures.   The image coding apparatus according to any one of claims 1 to 5, wherein the DC component acquisition unit is configured to acquire a DC component of input image data in units of DCT execution blocks.   The DC component storage unit is configured to set the DC component of the previous input image data that has already been encoded as the DC component to be compared with the DC component of the input image data. The image encoding device according to any one of the above.   The DC component storage unit is configured to set an average value of DC components for an arbitrary frame as a DC component to be compared with a DC component of input image data. The image encoding device according to claim 1.   The image coding apparatus according to claim 6, wherein the image change calculation unit sets a unit of a DC component to be compared with a DC component of input image data as a macroblock unit.   The image coding apparatus according to claim 7, wherein the image change calculation unit sets a unit of a DC component to be compared with a DC component of input image data as a picture unit.   The image coding apparatus according to claim 8, wherein the image change calculation unit uses a DC component unit to be compared with a DC component of input image data as a DCT execution block unit.   The image coding apparatus according to any one of claims 1 to 13, wherein the image coding apparatus converts an image pickup signal into a luminance / color difference signal, and the image coding process is performed by inputting the luminance / color difference signal from the image pickup apparatus. And a camera system.

Images