JP5848993B2 - Image processing device - Google Patents

Description

  The present invention relates to an image processing apparatus, and more particularly to a code amount control algorithm in image compression.

  Patent Document 1 below discloses an image processing apparatus according to the background art. In the image processing apparatus, based on the generated code amount related to Y (for example, 1120) macroblocks processed immediately before and the target code amount per X (for example, 1200) macroblocks less than one frame, An allowable code amount that can be used for N (for example, 80) macroblocks to be processed immediately after is obtained. Also, an expected code amount that is expected to be used for N macroblocks to be processed immediately after is obtained. Then, based on the allowable code amount and the expected code amount, a quantization parameter related to the processing target macroblock is set.

JP 2011-217112 A

  However, according to the image processing apparatus disclosed in Patent Document 1, image quality may be deteriorated immediately after an image to be processed is switched from a still image to a moving image. This will be specifically described below.

  FIG. 9 is a diagram illustrating frames before and after the image to be processed is switched from a still image to a moving image. When the image to be processed is a still image, the generated code amount is zero or close to it, and therefore the generated code amount related to the lower end region R1 of the final frame of the still image is zero or very small. Therefore, the allowable code amount is the maximum value or very large at the time of switching from the still image to the moving image. Therefore, the quantization parameter is set to be excessively small for the upper end region R2 of the first frame of the moving image, and accordingly, the generated code amount in the region R2 increases rapidly. As a result, for the region R3 that follows the region R2, the allowable code amount decreases rapidly, so that the quantization parameter is set excessively large. As described above, the region R2 in which the quantization parameter is set to be excessively small and the region R3 in which the quantization parameter is set to be excessively large are repeated, so that the image quality is deteriorated until the quantization parameter is stabilized at a certain level.

  The present invention has been made to solve such a problem, and an object thereof is to provide an image processing apparatus capable of avoiding deterioration in image quality immediately after an image to be processed is switched from a still image to a moving image. .

  An image processing apparatus according to a first aspect of the present invention includes: an encoder that performs an encoding process including a quantization process for an image signal; and a control unit that controls a quantization parameter in the quantization process. A first processing unit that obtains a generated code amount used for the immediately preceding first predetermined number of macroblocks, and a second predetermined number or less of macroblocks less than the total number of macroblocks included in one frame Based on the target code amount and the generated code amount obtained by the first processing unit, the allowable code that can be used for the third predetermined number of macroblocks immediately after including the current macroblock to be processed A second processing unit for determining the amount; a third processing unit for determining an expected code amount expected to be used for the third predetermined number of macroblocks; and the second processing unit. Therefore, a fourth processing unit that sets a quantization parameter related to the current macroblock to be processed, based on the allowable code amount obtained by the third processing unit and the expected code amount obtained by the third processing unit; And when the generated code amount of each macroblock is less than or equal to a preset lower limit for the first predetermined number of macroblocks, the first processing unit generates the generated code of the macroblock The amount is set as the lower limit value.

  According to the image processing apparatus according to the first aspect, the first processing unit, when the generated code amount of each macroblock is less than or equal to a preset lower limit value with respect to the first predetermined number of macroblocks The generated code amount of the macroblock is set as the lower limit value. For this reason, even if the actual generated code amount is zero or close to it because the image to be processed is a still image, the first processing unit regards that a certain amount of code has occurred. Accordingly, since the allowable code amount is not excessively large when the image to be processed is switched from the still image to the moving image, the quantization parameter is not set excessively small by the fourth processing unit. As a result, immediately after the image to be processed is switched from a still image to a moving image, an area where the quantization parameter is set too small and an area where the quantization parameter is set too large do not occur. Can be avoided.

  The image processing apparatus according to the second aspect of the present invention is the image processing apparatus according to the first aspect, in particular, the lower limit value set when the image to be processed is a still image, and the image to be processed is a moving image. It is the same as the lower limit value set in some cases.

  According to the image processing apparatus according to the second aspect, the same lower limit value is set when the image to be processed is a still image and when it is a moving image. Therefore, since it is not necessary to change the lower limit value according to the image, it is possible to simplify the configuration and processing.

  In the image processing device according to the third aspect of the present invention, in particular, in the image processing device according to the second aspect, the lower limit value is in the range of 20% to 25% of the target code amount per macroblock. It is characterized by being set.

  According to the image processing apparatus according to the third aspect, the lower limit value is set in a range of 20% to 25% of the target code amount per macroblock. By setting the lower limit value within this range, it is possible to effectively avoid degradation of image quality in both the upper end area of the first frame of the moving image and the subsequent area.

  In the image processing device according to the fourth aspect of the present invention, in particular, in the image processing device according to the first aspect, the lower limit value is different depending on whether the image to be processed is a still image or a moving image. The lower limit set when the image to be processed is a still image is larger than the lower limit set when the image to be processed is a moving image.

  According to the image processing device of the fourth aspect, the lower limit value set when the image to be processed is a still image is larger than the lower limit value set when the image to be processed is a moving image. Therefore, when the image to be processed is a still image, setting the lower limit value to a certain large value can effectively avoid the deterioration of image quality immediately after the image to be processed is switched from the still image to the moving image. It becomes possible. Also, when processing a movie after the image to be processed has been switched from a still image to a movie, set the lower limit value to a small value (may be zero) so that the movie can be made normal without being affected by the lower limit value. As a result, it is possible to avoid the deterioration of the image quality of the moving image due to the setting of the lower limit value.

  ADVANTAGE OF THE INVENTION According to this invention, the image processing apparatus which can avoid degradation of the image quality immediately after the image to process switches from a still image to a moving image can be obtained.

It is a figure which shows an example of the encoding process per macroblock. It is a block diagram which shows the structure of the image processing apparatus which concerns on embodiment of this invention. It is a block diagram which shows the structural example of the control part shown in FIG. It is a block diagram which shows the structural example of the 3rd process part shown in FIG. It is a figure which shows the example of a parameter set. It is a flowchart which shows an example of the determination method of an attribute value. It is a figure which shows an example of the selection method of a parameter set. It is a block diagram which shows the structure of the image processing apparatus which concerns on a modification. It is a figure which shows the frame before and after the image to process switches from a still image to a moving image.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In addition, the element which attached | subjected the same code | symbol in different drawing shall show the same or corresponding element.

  FIG. 1 is a diagram illustrating an example of an encoding process in units of macroblocks. FIG. 1 illustrates an image of one frame having 1280 pixels in the row direction (horizontal direction) and 720 pixels in the column direction (vertical direction). One frame is divided into macroblocks every 16 pixels in the row direction and every 16 pixels in the column direction. Therefore, one frame is divided into 80 macroblocks in the row direction and 45 macroblocks in the column direction. A set of 80 macroblocks corresponding to one row is called a macroblock line.

  FIG. 2 is a block diagram showing a configuration of the image processing apparatus 1 according to the embodiment of the present invention. As shown in FIG. 2, the image processing apparatus 1 includes an image generation unit 4, an encoder 2, and a control unit 3. The image generation unit 4 generates an image and inputs the image signal S 1 to the encoder 2. The encoder 2 can be MPEG-2, MPEG-4, or H.264. The image signal S4 after compression encoding is output by performing image processing such as quantization processing and encoding processing on the image signal S1 before compression encoding. To do. The image signal S4 is transmitted from the image processing apparatus 1 to a display device (not shown) via a wireless LAN or the like, and an image (including a still image and a moving image) is displayed on the display device.

  The control unit 3 receives the image signals S1 and S4 and also receives a signal S2A indicating the maximum allowable code amount per macroblock and a signal S2B indicating the target code amount per macroblock. Here, the maximum allowable code amount per macroblock is the maximum bit rate (for example, 9 Mbps) from the image processing apparatus 1 to the display apparatus, the frame rate of the image (for example, 60 fps), and the total number of macroblocks in one frame. (For example, 3600). Similarly, the target code amount per macroblock is based on the target bit rate (for example, 8.1 Mbps) from the image processing apparatus 1 to the display apparatus, the frame rate of the image, and the total number of macroblocks in one frame. Is calculated. Based on these signals S1, S2A, S2B, and S4, the control unit 3 controls the quantization parameter in the quantization process in the encoder 2 with the control signal S3. In the image processing apparatus 1 according to the present embodiment, the generated code amount for X macroblocks corresponding to the allowable transmission delay (for example, 1200 macroblocks included in the 15 macroblock lines) is X macroblocks. The quantization parameter is controlled so as not to exceed the maximum allowable code amount (for example, 9000000/60/3 = 50000 bits).

  FIG. 3 is a block diagram illustrating a configuration example of the control unit 3 illustrated in FIG. 2. As shown by the connection relationship in FIG. 3, the control unit 3 includes a first processing unit 11, a second processing unit 12, a third processing unit 13, and a fourth processing unit 14.

  The first processing unit 11 obtains the generated code amount used for the Y macroblocks processed immediately before based on the image signal S4. Specifically, the generated code amount for each of the Y macroblocks is obtained, and the generated code amounts for the Y macroblocks are calculated by summing them. The calculated generated code amount for Y macroblocks is input from the first processing unit 11 to the second processing unit 12 as a signal S11.

  Here, when the signal S5 is input to the first processing unit 11, a predetermined lower limit M regarding the generated code amount per macroblock is set in advance. When the generated code amount for each of the macroblocks is equal to or lower than the lower limit value M when calculating the generated code amount for the Y macroblocks, the first processing unit 11 sets the generated code amount of the macroblock as the lower limit value M. deal with. The lower limit value M is set to a common value when the current processing target image is a still image and when it is a moving image.

  Referring to FIG. 9, lower limit value M is an optimum value that can effectively avoid deterioration in image quality in both the upper end region R2 of the first frame of the moving image and the subsequent region R3 (and subsequent regions). Set to If the lower limit value M is too small, the generated code amount related to the region R2 becomes large and the generated code amount related to the region R3 becomes small, so that the image quality of the region R3 deteriorates. If the lower limit M is too large, the amount of generated code related to the region R2 becomes small and the amount of generated code related to the region R3 increases, so that the image quality of the region R2 deteriorates. On the other hand, if the lower limit value M is too large, when a moving image is normally processed after the region R3, even if the actual generated code amount of a certain macroblock is less than or equal to the lower limit value M, the generated code amount of the macroblock is Since it is handled as the lower limit value M, the quantization parameter is set larger than the appropriate value, and the image quality deteriorates.

  The optimum value of the lower limit value M may be obtained by experiment or simulation, and in the example of the present embodiment, it is set to an arbitrary value within the range of 20% to 25% of the target code amount per macroblock. The In the example of the present embodiment, the target code amount per macroblock is about 41.7 bits (= 9000000/60/80/45), so the lower limit M is about 8.3 to 10.4 bits. Set within the range. In the case where the lower limit value is set to 10 bits, for example, the first processing unit 11 determines that the macro block when the generated code amount related to a specific macro block in the immediately preceding Y macro blocks is 10 bits or less. Are treated as if a 10-bit code amount has occurred.

  The second processing unit 12 obtains a target code amount per X macroblocks based on the signals S2A and S2B. Further, the second processing unit 12 performs N macroblocks (for example, immediately after processing) based on the calculated target code amount and the generated code amount related to the Y macroblocks input from the first processing unit 11 (for example, The allowable code amount that can be used with respect to the current macro block to be processed (80 macro blocks included in one macro block line) is obtained. The obtained allowable code amount is input from the second processing unit 12 to the fourth processing unit 14 as a signal S12.

  Here, the second processing unit 12 predicts the amount of generated code from the number of macroblocks (X) corresponding to the allowable transmission delay because the number of generated macroblocks (Y) is already determined. If the number is equal to or greater than the value obtained by subtracting the number of possible macroblocks (N) (that is, the relationship Y ≧ XN), the immediately preceding XN is calculated from the target code amount for X macroblocks. A value obtained by subtracting the generated code amount used for the macroblocks is set as an allowable code amount usable for the N macroblocks to be processed immediately after. On the other hand, the number of macroblocks whose generated code amount has already been determined (Y) is the number of macroblocks (N) from which the generated code amount can be predicted from the number of macroblocks (X) corresponding to the allowable transmission delay. Is less than the value obtained by subtracting (number of), that is, when the relationship of Y <XN is satisfied, the generated code amount used for the immediately preceding Y macroblocks from the target code amount for Y + N macroblocks. A value obtained by subtracting is used as an allowable code amount usable for N macroblocks to be processed immediately after. Therefore, even when the relationship of Y ≧ X−N is satisfied or when the relationship of Y <X−N is satisfied, the allowable code amount can be appropriately obtained.

  The third processing unit 13 obtains an expected code amount that is expected to be used for the immediately subsequent N macroblocks based on the image signal S1 (details will be described later). The obtained expected code amount is input from the third processing unit 13 to the fourth processing unit 14 as a signal S13.

  Based on the allowable code amount input from the second processing unit 12 and the expected code amount input from the third processing unit 13, the fourth processing unit 14 determines the quantization parameter related to the current macroblock to be processed. Set (details will be described later). The set quantization parameter is input from the fourth processing unit 14 to the encoder 2 shown in FIG. 2 as the signal S3.

  FIG. 4 is a block diagram illustrating a configuration example of the third processing unit 13 illustrated in FIG. 3. As shown in the connection relationship of FIG. 4, the third processing unit 13 includes an overall evaluation value calculation unit 21, a partial evaluation value calculation unit 22, a Sobel filter processing unit 23, an expected code amount calculation unit 24, and a storage unit 25. Configured.

  Based on the image signal S1, the overall evaluation calculation unit 21 calculates an activity evaluation value act1 relating to the entire area of each macroblock (vertical 16 pixels × horizontal 16 pixels) for each of the immediately following N macroblocks. Here, the activity evaluation value is an index indicating the degree of variation of the pixel value in the macroblock. For example, the sum of absolute differences between the luminance average value of the macroblock and the luminance value of each pixel is calculated as the macroblock luminance value. Obtained as a value divided by the number of pixels in the block. By using the activity evaluation value act1 in this way, for a macroblock having an attribute that is smooth as a whole (that is, a macroblock in which deterioration in image quality is conspicuous), by estimating a small quantization parameter (that is, a large code amount), An appropriate expected code amount can be obtained.

  The partial evaluation value calculation unit 22 calculates the minimum value act2 of the activity evaluation values for the plurality of partial areas of each macroblock for each of the N macroblocks immediately after, based on the image signal S1. That is, a vertical 4 pixel × horizontal 16 pixel region including the upper side of each macroblock, a vertical 4 pixel × horizontal 16 pixel region including the lower side, a vertical 16 pixel × horizontal 4 pixel region including the left side, and a vertical side including the right side. The activity evaluation values for the area of 16 pixels × 4 horizontal pixels are calculated, and the minimum value act2 is obtained. By using the minimum value act2 in this way, for example, a macroblock having an attribute that is complex as a whole, does not include a character area, and does not include a smooth portion (that is, deterioration in image quality is not noticeable). For a macro block, an appropriate expected code amount can be obtained by estimating a large quantization parameter (that is, a small code amount). In addition, for example, with respect to a macroblock having an attribute that is complex as a whole and does not include a character area but includes a smooth portion in part (that is, a macroblock in which deterioration in image quality is conspicuous in a smooth portion) By estimating a small quantization parameter (that is, a slightly large code amount), an appropriate expected code amount can be obtained.

  Based on the image signal S1, the Sobel filter processing unit 23 performs edge extraction processing using a Sobel filter for each of the plurality of partial areas of each macroblock for each of the N macroblocks immediately after the image signal S1. The maximum value Sobel is obtained. That is, a vertical 4 pixel × horizontal 16 pixel region including the upper side of each macroblock, a vertical 4 pixel × horizontal 16 pixel region including the lower side, a vertical 16 pixel × horizontal 4 pixel region including the left side, and a vertical side including the right side. Edge extraction processing is executed for an area of 16 pixels × 4 horizontal pixels, and the maximum value Sobel of the filter result is obtained. By using the maximum value Sobel in this way, for example, with respect to a macroblock having an attribute that is complicated as a whole and includes a character area in part (that is, a macroblock in which deterioration in image quality is conspicuous in the character area), By estimating a certain quantization parameter (that is, a medium code amount), an appropriate expected code amount can be obtained.

  FIG. 5 is a diagram illustrating an example of a parameter set stored in the storage unit 25 illustrated in FIG. In each of the plurality of parameter sets QPS0 to QPS7, quantization parameters are respectively set corresponding to the macroblock attribute values ACT0 to ACT4. For example, for the parameter set QPS4, a quantization parameter of value “38” is set corresponding to the attribute value ACT0, a quantization parameter of “20” is set corresponding to the attribute value ACT1, and the attribute value A quantization parameter of “38” is set corresponding to ACT2, a quantization parameter of “44” is set corresponding to attribute value ACT3, and a quantization parameter of “26” corresponding to attribute value ACT4. Parameter is set. For example, for the parameter set QPS7, a quantization parameter of “51” is set corresponding to the attribute values ACT0 to ACT4.

  In addition, a code amount (reference code amount) that is assumed to be generated when the quantization parameter is set to a reference value (for example, “51” which is the maximum value in the standard of H.264) is the attribute value of the macroblock. They are set corresponding to ACT0 to ACT4. In this example, 32 bits corresponding to attribute value ACT0, 4 bits corresponding to attribute value ACT1, 32 bits corresponding to attribute value ACT2, 32 bits corresponding to attribute value ACT3, and attribute value ACT4 Thus, an 8-bit reference code amount is set for each. By appropriately setting the reference code amount, an appropriate expected code amount can be obtained for each of the plurality of parameter sets QPS0 to QPS7.

  With reference to FIG. 4, the expected code amount calculation unit 24 calculates the activity evaluation value act1 obtained by the overall evaluation value calculation unit 21 and the partial evaluation value calculation unit 22 for each of the N macroblocks immediately after. Based on the obtained minimum value act2, the maximum value Sobel obtained by the Sobel filter processing unit 23, and the parameter set read from the storage unit 25, the expected code amount expected to be used for the immediately following N macroblocks is calculated. Ask. Specifically, it is as follows.

  The expected code amount calculation unit 24 first determines an attribute value for each of the N macroblocks immediately after. FIG. 6 is a flowchart illustrating an example of an attribute value determination method. As shown in the flowchart of FIG. 6, the attribute value of each macroblock is determined to be the attribute value ACT1 when the first condition that the activity evaluation value act1 is “2” or less is satisfied, and the first condition If the second condition that the activity evaluation value act1 is “6” or less is not satisfied, the attribute value ACT4 is determined, and the second value is not satisfied and the maximum value Sobel is “120” or more. When the third condition is satisfied, the attribute value ACT0 is determined. When the fourth condition that the third condition is not satisfied and the minimum value act2 is “5” or less is determined, the attribute value ACT2 is determined. When the fourth condition is not satisfied, the attribute value ACT3 is determined. The expected code amount calculation unit 24 counts how many macroblocks determined by the attribute values ACT0 to ACT4 are included in the immediately following N macroblocks.

  Next, the predicted code amount calculation unit 24 calculates the total predicted code amount of the immediately following N macroblocks when each of the parameter sets QPS0 to QPS6 is applied. As shown in FIG. 5, the quantization parameter of the parameter set QPS0 to QPS6 is set to any one of “20”, “26”, “32”, “38”, “44”, and “51”. Yes. H. According to the H.264 standard, if the quantization parameter is reduced by “6”, the quantization accuracy is doubled. Therefore, when the quantization parameter is set to “20”, “26”, “32”, “38”, “44”, and “51”, the code amount assumed to be generated is the reference code amount, respectively. It becomes 32 times, 16 times, 8 times, 4 times, 2 times, and 1 time. Therefore, immediately after the N macroblocks, the macroblocks determined to have the attribute values ACT0, ACT1, ACT2, ACT3, and ACT4 are included in K0, K1, K2, K3, and K4, respectively. In this case, for example, when the parameter set QPS4 is applied, the total expected code amount is 32 (bits) × 4 (times) × K0 (pieces) + 4 × 32 × K1 + 32 × 4 × K2 + 32 × 2 × K3 + 8 × 16 × K4 It is obtained by the following operation. The expected code amount calculation unit 24 also calculates the total expected code amount when the other parameter sets QPS0 to QPS3, QPS5, and QPS6 are applied in the same manner as described above. Then, the total expected code amount when each of the parameter sets QPS0 to QPS6 is applied is input to the fourth processing unit 14 shown in FIG. 3 as a signal S13.

  FIG. 7 is a diagram illustrating an example of a parameter set selection method performed by the fourth processing unit 14. The allowable code amount for the N macroblocks immediately after is obtained as a value obtained by subtracting the generated code amount for the XN macroblocks processed immediately before from the target code amount for the X macroblocks.

  The fourth processing unit 14 selects a parameter set that has the maximum expected code amount below the allowable code amount among the plurality of parameter sets QPS0 to QPS6. In the example shown in FIG. 7, the parameter set QPS4 is selected. And the 4th process part 14 sets the quantization parameter regarding the macroblock of the now process target based on selected parameter set QPS4. For example, when the attribute value of the current macroblock to be processed is the attribute value ACT0, the quantization parameter for the macroblock is set to “38” defined at the intersection of QPS4 and ACT0 in FIG. To do. According to this example, it is possible to improve the image quality as much as possible while suppressing the generated code amount below the target code amount.

  As another example, the fourth processing unit 14 selects a parameter set whose predicted code amount is closest to the allowable code amount from among a plurality of parameter sets QPS0 to QPS6. In the example shown in FIG. 7, the parameter set QPS3 is selected. And the 4th process part 14 sets the quantization parameter regarding the macroblock of the now process target based on selected parameter set QPS3. For example, if the attribute value of the current macroblock to be processed is the attribute value ACT0, the quantization parameter for the macroblock is set to “32” defined at the intersection of QPS3 and ACT0 in FIG. To do. According to this example, since the generated code amount can be brought close to the target code amount, stable code amount control can be performed. In this example, when the expected code amount exceeds the allowable code amount, and the sum of the generated code amount and the expected code amount is, for example, 105% or more of the target code amount, the expected code exceeding the allowable code amount. Selection of parameter sets that are quantities may be prohibited.

  In addition, when the predicted code amount exceeds the allowable code amount even when the parameter set QPS6 having the smallest predicted code amount is selected from among the plurality of parameter sets QPS0 to QPS6, the fourth processing unit 14 performs processing illustrated in FIG. The indicated parameter set QPS7 is selected. When the parameter set QPS7 is selected, the quantization parameter related to the current macroblock to be processed is set to “51” which is the maximum value in the standard.

<Modification>
FIG. 8 is a block diagram illustrating a configuration of the image processing apparatus 1 according to the modification. A signal S6 indicating whether the currently generated image is a still image or a moving image is input from the image generation unit 4 to the control unit 3. Thereby, the control unit 3 can determine whether the currently processed image is a still image or a moving image (that is, whether each frame is a still image frame or a moving image frame).

  In the above embodiment, the lower limit value M is set to a common value between the case where the image to be processed is a still image and the case where the image is a moving image. On the other hand, in this modification, the lower limit M is set to a different value depending on whether the image to be processed is a still image or a moving image. For example, the lower limit M is set to about 50% of the target code amount per macroblock when the image to be processed is a still image, and is set to zero when the image to be processed is a moving image. . Referring to FIG. 9, when determining the amount of generated code related to region R1, first processing unit 11 sets lower limit M to 50% because region R1 is a frame of a still image. Further, when the amount of generated code related to the regions R2 and R3 is obtained, the lower limit value M is set to zero because the regions R2 and R3 are frames of moving images.

<Summary>
As described above, according to the image processing apparatus 1 according to the present embodiment, the first processing unit 11 sets the generated code amount of each macroblock for the Y macroblocks processed immediately before the preset lower limit value. If it is equal to or less than M, the generated code amount of the macroblock is set to the lower limit value M. Therefore, even if the actual generated code amount is zero or close to it because the image to be processed is a still image, the first processing unit 11 regards that a certain amount of code has occurred. Accordingly, since the allowable code amount is not excessively large when the image to be processed is switched from the still image to the moving image, the quantization parameter is not set excessively small by the fourth processing unit 14. As a result, immediately after the image to be processed is switched from a still image to a moving image, an area where the quantization parameter is set too small and an area where the quantization parameter is set too large do not occur. Can be avoided.

  Further, according to the image processing apparatus 1 according to the present embodiment, the same lower limit value M is set for the case where the image to be processed is a still image and the case where the image is a moving image. Therefore, it is not necessary to change the lower limit value M according to the image, so that the configuration and processing can be simplified.

  Further, according to the image processing apparatus 1 according to the present embodiment, the lower limit value M is set within the range of 20% to 25% of the target code amount per macroblock. By setting the lower limit value M within this range, it is possible to effectively avoid deterioration in image quality in both the upper end region R2 of the first frame of the moving image and the subsequent region R3.

  Further, according to the image processing device 1 according to the modification, the lower limit value M set when the image to be processed is a still image is larger than the lower limit value M set when the image to be processed is a moving image. . Therefore, when the image to be processed is a still image, the lower limit value M is set to a somewhat large value to effectively avoid deterioration in image quality immediately after the image to be processed is switched from a still image to a moving image. Is possible. In addition, when a moving image is processed after the image to be processed is switched from a still image to a moving image, by setting the lower limit value M to a small value (may be zero), the moving image is not affected by the lower limit value M. Processing can be performed normally, and as a result, it is possible to avoid the deterioration of the image quality of the moving image associated with setting the lower limit value M.

DESCRIPTION OF SYMBOLS 1 Image processing apparatus 2 Encoder 3 Control part 11 1st process part 12 2nd process part 13 3rd process part 14 4th process part

Claims (4)

An encoder that performs encoding processing including quantization processing on an image signal;
A control unit for controlling a quantization parameter in the quantization process;
With
The controller is
A first processing unit for determining a generated code amount used for the immediately preceding first predetermined number of macroblocks;
Based on a target code amount per second macroblock that is less than or equal to a second predetermined number less than the total number of macroblocks included in one frame and the generated code amount obtained by the first processing unit, A second processing unit that obtains an allowable code amount that can be used for a third predetermined number of macroblocks immediately after including the target macroblock;
A third processing unit for obtaining an expected code amount expected to be used for the third predetermined number of macroblocks;
Based on the allowable code amount obtained by the second processing unit and the expected code amount obtained by the third processing unit, a quantization parameter relating to a current macroblock to be processed is set. 4 processing units;
Have
When the generated code amount of each macroblock is less than or equal to a preset lower limit for the first predetermined number of macroblocks, the first processing unit determines the generated code amount of the macroblock as the lower limit. An image processing apparatus as a value.   The image processing apparatus according to claim 1, wherein the lower limit value set when the image to be processed is a still image is the same as the lower limit value set when the image to be processed is a moving image.   The image processing apparatus according to claim 2, wherein the lower limit value is set within a range of 20% to 25% of a target code amount per macroblock. The lower limit value is set to a different value depending on whether the image to be processed is a still image or a moving image,
The image processing apparatus according to claim 1, wherein the lower limit value set when the image to be processed is a still image is larger than the lower limit value set when the image to be processed is a moving image.

Images