JP4118213B2 - Image compression method, image compression device, imaging device, processing location identification method, processing location identification device, and computer program - Google Patents

Description

  The present invention relates to an image compression method, an image compression apparatus, an imaging apparatus, a processing location specifying method, a processing location specifying device, and a computer program for suppressing distortion associated with image compression processing.

  Conventionally, image compression for still images and moving images has been widely performed in order to reduce the signal amount. For example, image compression in JPEG format using two-dimensional discrete cosine transform is well known for still images. Such JPEG compression is often applied to an imaging apparatus such as a digital camera, and JPEG compression may be performed in software using a computer.

  FIG. 12 is a block diagram illustrating an internal configuration of an imaging apparatus 1 that performs conventional JPEG compression. The inside of the imaging apparatus 1 is roughly divided into an imaging processing unit 2, an image compression unit 3, and a monitor unit 4. The imaging processing unit 2 includes an imaging device 2a that photoelectrically converts an image guided from an optical system (not shown) to generate an image signal, and an A / D conversion unit 2b that performs noise elimination, analog / digital conversion, and the like.

  The image compression unit 3 is configured by sequentially connecting a processing unit 3a, a filter processing unit 3b, a JPEG processing unit 3c, a bus management unit 3d, a memory 3e, an MPU 3f, and an external memory interface unit 3g through an internal bus 3h. The sequential processing unit 3a and the filter processing unit 3b perform necessary preprocessing for image compression, and the JPEG processing unit 3c performs a series of processing related to image compression such as level shift processing and two-dimensional discrete cosine transform processing. The compressed image is output via the external memory interface unit 3g and stored in the external memory 5. Note that the compression rate for image compression can be adjusted as appropriate. JPEG compression is also disclosed in Patent Document 1 below.

FIGS. 13A to 13C show a situation in which the size of the captured image is reduced through each process in the image compression unit 3.
The outer frame of FIG. 13A shows an image range 6a related to the captured image. In this example, the size of the image range 6a is 1344 (the number of horizontal pixels, the same applies hereinafter) × 971 (the number of vertical pixels, the same applies hereinafter). The inner frame in FIG. 13A shows the first processing range 6b (1292 × 966) in which the so-called optical black portion 7 is deleted in the processing of the processing portion 3a. Further, the inner frame of FIG. 13B shows a second processing range 6c (1288 × 962) whose size is reduced from the first processing range 6b by the filter processing of the filter processing unit 3b.

Further, the inner frame shown in FIG. 13C indicates a third processing range 6d (1280 × 960) cut out from the second processing range 6c by the MPU 3f in order to perform compression processing by the JPEG processing unit 3c. The size of the third processing range 6 is such that the number of horizontal and vertical pixels is a multiple of “8”, which is the number of pixel units related to JPEG compression processing. In addition, since the image signal component is YCrCb-based, the signal amount of the color difference signals Cr and Cb is half of the luminance signal Y in the horizontal direction or half in the horizontal and vertical direction, so a multiple of “16” is set to the image size. Often. As described above, the conventional image compression unit 3 performs processing that is determined so as to be cut out from a certain location with a certain number of pixels, in relation to the processing for reducing the image size at each processing stage.
JP 2001-157226 A

  Image compression such as the JPEG format is suitable for natural images such as landscapes in which the image to be compressed is a general landscape, but a binary having a large contrast difference between adjacent pixels in terms of luminance, color difference, etc. There is a problem that distortion such as block noise and mosquito noise occurs at a high compression rate for images and animation images.

  FIG. 14A shows an image before JPEG compression. This image has a size of 40 × 40, and a square patch (size: 8 × 8) is placed at a position where the upper left vertex is a coordinate pixel (15, 15) and the lower right vertex is a coordinate pixel (23, 23). Have. The brightness level of the patch is “64” in numerical value, the brightness level of the background portion is “192”, and the difference in contrast between the two is large. Note that the image in FIG. 14D has the number of pixel units (8 × 8) in the horizontal and vertical directions related to the JPEG compression process (two-dimensional discrete cosine transform: DCT) with respect to the image in FIG. The broken lines are shown in a grid pattern, and the broken lines do not exist in the actual image. Further, when this grid-like broken line is compared with the patch, it can be seen that the patch is located beyond the boundary (broken line) of the pixel unit related to image compression.

  FIG. 14B shows a low-compression image obtained by performing JPEG compression on the image of FIG. 14A at the minimum compression rate, and FIG. 14C shows the image of FIG. A highly compressed image obtained by performing JPEG compression at a compression rate of

  The low-compressed image in FIG. 14B is at a level that does not hinder the image quality, and at first glance it appears that no distortion has occurred. However, when this low-compressed image is scrutinized, there is a slight distortion in the rectangular area from the coordinate pixel (8, 8) to the coordinate pixel (23, 23), and the brightness level is the portion where the rectangular area and the patch overlap. There are pixels that vary from “62” to “65”. On the other hand, the high-compressed image in FIG. 14C has a high compression ratio, so that significant distortion occurs from the coordinate pixel (8, 8) to the rectangular area of the coordinate pixel (23, 23).

  This is because the distortion occurs differently within the number of pixel units (8 × 8 lattice blocks) related to the two-dimensional discrete cosine transform, which is one of the processes in the compressed image in JPEG format. This is because it is observed as block noise. It should be noted that no distortion occurs outside the rectangular area described above.

  In general, JPEG compression removes high-frequency components by converting the component signals related to the contrast of each pixel into frequency components by two-dimensional discrete cosine transform, and then performing quantization, so the degree depends on the compression rate. Although it fluctuates, it is inevitable that distortion occurs in the compressed image. In particular, in a case where an image to be compressed has a large contrast difference between pixels such as a binary image and an animation image, a lot of high-frequency components are included in the boundary portion of contrast, and such high-frequency components are deleted by quantization. Therefore, distortion is likely to occur.

  Note that the above-described problems are not limited to JPEG compression, but are MPEG, H.264, which are compression standards for moving images. The same occurs in image compression processing using two-dimensional discrete cosine transform such as H.262.

The present invention has been made in view of such a problem, and occurs in a compressed image by calculating a difference in contrast between pixels and adjusting a position where image processing such as compression is performed according to the calculated result. An object of the present invention is to provide an image compression method, an image compression device, an imaging device, a processing location specifying method, a processing location specifying device, and a computer program that avoid distortion.
In addition, the present invention calculates the contrast difference between pixels to identify the pixel position where the maximum contrast difference occurs, and the boundary of the pixel unit related to image processing such as compression coincides with the identified pixel position. An object of the present invention is to provide an image compression method, an image compression device, an imaging device, a processing location specifying method, a processing location specifying device, and a computer program that improve the quality of a compressed image by adjusting the location where image processing is performed And

Furthermore, an object of the present invention is to provide an image compression method, an image compression apparatus, and an imaging apparatus that facilitate comparison and detection related to the contrast difference by appropriately increasing or decreasing the calculated contrast difference.
Furthermore, another object of the present invention is to provide an imaging apparatus that performs compression processing so as to suppress the occurrence of distortion on a captured image.

The image compression method of the present invention is an image compression method for compressing an image by preparing a plurality of difference storage means corresponding to each pixel position in a pixel block, which is a unit of image compression, and forming an image. fractionating the absolute value of the difference of a plurality of pixels each having signal and a step of calculating between adjacent pixels, the absolute value of the calculated difference for each pixel position in the pixel block based on the pixel position according to said difference of the Comparing the absolute value of the accumulated difference stored in the difference storage means with the step, integrating the absolute value of the sorted difference in the difference storage means corresponding to the pixel position related to the difference, and storing the difference step a, step a, the pixel block difference memory means for storing the absolute value of the largest accumulated the difference detected corresponding to detecting the absolute value of the largest accumulated the difference by the comparison performed Based on the pixel position, and wherein the step of identifying where to begin the compression process, further comprising the step of performing image compression in pixel block units from specified locations against the image.

Image compression method of the present invention further comprises a step of increasing or decreasing in accordance with the absolute value of the calculated difference to the absolute value of said difference, the step of fractionation to separate the absolute value of the difference step of the increase or decrease is increased or decreased It is characterized by that.

Image compression apparatus of the present invention includes: a calculating means for a plurality of pixels forming an image is computed between pixels adjacent the absolute value of the difference signal, each of which has the absolute value of the difference the calculating means is calculated, said difference Based on the pixel position according to the above, the classification means for classifying into each pixel position in the pixel block that is a processing unit of image compression, and for storing the accumulated absolute value of the difference classified by the classification means, A plurality of difference storage means corresponding to each pixel position in the pixel block, a comparison means for comparing the absolute values of the accumulated differences respectively stored in the difference storage means, and a comparison by the comparison means Based on the pixel position in the pixel block corresponding to the detection means for detecting the absolute value of the accumulated difference and the difference storage means for storing the absolute value of the maximum accumulated difference detected by the detection means, the image in pairs to A position specifying means for specifying where to start the compression process, characterized in that it comprises an image compressing means for performing image compression in pixel block from where it was identified relevant section specifying means.

Image compression apparatus of the present invention includes: a calculating means for a plurality of pixels forming an image is computed between pixels adjacent the absolute value of the difference signal, each of which has the absolute value of the difference which the calculation means has calculated the said difference a adjusting unit that adjusts according to the absolute value, based on the pixel position of the absolute value of the difference bulking down means increases or decreases in said difference, separation means for separating each pixel position in the pixel block is a processing unit of an image compression If a storage of該分by means for accumulating and storing the absolute value of the difference obtained by fractionation, and a plurality of difference storing means respectively corresponding to each pixel position in the pixel block, integrated for storing the said difference storage means comparing means for comparing the absolute value with each other by the differential, a detection means for detecting the absolute value of the largest accumulated the difference by the comparison of the comparison means, the maximum that detecting means detects accumulated the difference the difference for storing the absolute value Based on the pixel position in the pixel block憶means corresponds, the part specifying means for specifying a position to start the compression process, the image compression unit which performs image compression in pixel block from where it was identified relevant section specifying means It is characterized by providing.

The image compression apparatus according to the present invention is characterized in that the signal is a luminance signal and / or a color difference signal related to an image.

  The imaging device of the present invention includes an imaging unit that outputs a captured image and the image compression device, and the calculation unit calculates the image output by the imaging unit.

The processing location specifying method of the present invention prepares a plurality of difference storage means respectively corresponding to each pixel position in a pixel block which is a processing unit of image compression, and specifies a processing location specifying a location where image compression processing is started. In the method, a step of calculating an absolute value of a difference between signals of each of a plurality of pixels forming an image between adjacent pixels, and an absolute value of the calculated difference in the pixel block based on a pixel position related to the difference . a step of separating each pixel position, a step of storing the absolute value by integrating the difference storing means corresponding to the pixel position according to said difference of fractionated difference, accumulated a difference of said difference storage means for storing performing a comparison of the absolute value between the maximum and detecting the accumulated absolute value of the difference, difference storing hand stores the absolute value of the largest accumulated the difference detected by the comparison There based on the pixel position in the pixel block corresponding, characterized in that it comprises the steps of identifying where to start the compression processing against the image.

Processing site identification device of the invention, said difference plurality of pixels forming an image and a calculating means for calculating between pixels adjacent the absolute value of the difference signal, each of which has the absolute value of the difference which the calculation means has calculated An increase / decrease unit that increases / decreases in accordance with an absolute value of the image, and a classification that classifies the absolute value of the difference increased / decreased by the increase / decrease unit into each pixel position in a pixel block that is a unit of image compression based on the pixel position related to the difference And a storage for accumulating and storing the absolute values of the differences sorted by the sorting means, a plurality of difference storage means corresponding to the pixel positions in the pixel block, and an accumulation stored by the difference storage means comparing means for comparing the absolute value with each other by the differential, a detection means for detecting the absolute value of the largest accumulated the difference by the comparison of the comparison means, the maximum that detecting means detects accumulated the difference the difference for storing the absolute value憶means, characterized in that it comprises a point specifying means for specifying a position to start the compression processing against the image based on the pixel position in the pixel block corresponding.

Computer program of the present invention is a computer program for causing an image compression to the computer, operation manual stage a computer, a plurality of pixels forming an image is computed between pixels adjacent the absolute value of the difference signal, each of which has , the absolute value of the difference the calculating means is calculated, based on the pixel position according to said difference, separation means for separating each pixel position of the image compression processing pixel block which is a unit, each pixel position of the pixel block A plurality of difference storage means for accumulating and storing the absolute values of the differences sorted by the sorting means, comparison means for comparing the absolute values of the accumulated differences stored by the difference storage means, Detection means for detecting the absolute value of the maximum integrated difference by comparison of the comparison means, and storing the absolute value of the maximum integrated difference detected by the detection means Based on the pixel position in the pixel block divided storage means corresponding performs image compression in pixel block units from point specifying means to identify where to start the compression process, and identified relevant section identifying unit locations relative to the image It functions as an image compression means.

The computer program according to the present invention is a computer program for causing a computer to specify a place to perform image processing. The computer calculates an absolute value of a difference between signals of a plurality of pixels forming an image between adjacent pixels. operation manual stage, said calculating means based on the pixel position of the absolute value to the difference amount of the difference computed, separating means for separating each pixel position of the image compression processing pixel block which is a unit pixel of the pixel block respectively corresponding to the position, the plurality of difference storing means separating means for accumulating and storing the absolute value of the difference fractionated, compared hand stage said difference storage means for comparing the absolute value between the accumulated the differential store , detecting means for detecting a maximum accumulated absolute value of the difference by the comparison of the comparison means, the maximum accumulated the stored absolute value of the difference is detecting means has detected Based on the pixel position in the pixel block that difference storage means corresponds, characterized in that to function as a part specifying means for specifying a position to start the compression processing against the image.

  In the present invention, the difference between the signals of each pixel forming the image is calculated and compared, so that the contrast difference between adjacent pixels becomes clear. In addition, in order to identify a place where image processing such as compression is performed according to the comparison result, a suitable place for image processing is obtained. Furthermore, it is possible to suppress the degree of distortion caused by the compression by performing image compression on the location thus obtained.

  In the present invention, the result of calculating the signal difference between the pixels is separately stored in a plurality of difference storage means each corresponding to the number of pixel units related to image processing such as compression, so the calculated contrast difference Are arranged for each pixel unit. In addition, the maximum difference is detected by comparing the contrast difference in an organized state, and the location where distortion due to compression occurs can be determined from the pixel unit corresponding to the difference storage means for storing the maximum difference. . Therefore, by specifying the location to be compressed so that the determined location matches the boundary of the pixel unit related to the compression processing, it is possible to prevent the image quality from being deteriorated due to block noise, mosquito noise, or the like.

Further, in the present invention, by appropriately increasing / decreasing the calculated contrast difference, the difference between the pixels becomes clear, and the comparison between the differences and the maximum difference detection are facilitated.
Furthermore, in the present invention, since the difference storage means accumulates and stores the differences, the comparison for each number of pixel units can be made equally with respect to the entire image without being affected by a large difference locally. Overall comparison judgment becomes possible.

In the present invention, since a luminance signal and / or a color difference signal is used as a signal related to contrast, distortion is suitably applied to a process of conversion from an RGB system component signal to a YCbCr system component signal by filter processing. Processing to suppress can be performed.
Furthermore, in the present invention, since the processing for suppressing the distortion as described above is performed on the captured image, it is possible to suppress the degree of distortion in the compressed image generated by the image compression processing in the imaging apparatus.

  In the present invention, the difference of the signal of each pixel of the image is calculated and compared, the location to be compressed is identified according to the comparison result, and the identified location is image-compressed so that the contrast is reduced. The compression process can be performed so as not to be affected by the difference, and the quality of the compressed image can be improved compared to the conventional case by suppressing the occurrence of distortion.

  In the present invention, the result of calculating the signal difference between the pixels is separately stored in a plurality of difference storage means each corresponding to the number of pixel units related to image compression, whereby each difference of the signal in the image is converted into a pixel. Organize by unit. Further, since the maximum difference is detected by comparing with the contents stored in each difference storage means, it can be easily specified based on the storage state in which the pixel units related to the maximum difference are arranged. Therefore, by specifying the location to be compressed and performing image compression so that the pixel position that becomes the maximum difference based on the specified pixel unit coincides with the boundary of the pixel unit related to the compression process, the maximum difference is caused. The occurrence of such distortion can be eliminated, and deterioration of the image quality due to clear block noise and mosquito noise from the compressed image can be prevented.

In the present invention, by appropriately increasing or decreasing the calculated difference, the difference of each difference between pixels becomes clear, and comparison and detection of differences can be easily performed.
Furthermore, in the present invention, the difference storage means accumulates and stores the differences, and the comparison means compares the accumulated differences with each other, so that the entire difference is not affected by a large difference generated locally. Can be compared in a uniform manner, and can contribute to the improvement of the overall quality of the compressed image.

In the present invention, by using a luminance signal and / or a color difference signal as a signal relating to contrast, processing relating to distortion suppression can be performed smoothly in accordance with a signal to be filtered.
Furthermore, in the present invention, by performing the above-described processes on the captured image, it is possible to suppress the influence of distortion on the compressed image generated by the image compression process in the imaging apparatus.

In the present invention, it is possible to obtain a suitable image range for image processing by calculating a difference between signals of each pixel of the image and performing comparison, and specifying a compression target portion according to the comparison result. it can.
Further, in the present invention, the result of calculating the signal difference between the pixels is separately stored in a plurality of difference storage means respectively corresponding to the number of pixel units related to image processing, and the maximum in comparison between stored contents Since the difference is detected and the location where the image processing is performed is specified based on the number of pixel units corresponding to the maximum difference, an image range that avoids the influence of the image processing can be provided.

  FIG. 1 is a block diagram showing an internal configuration of the imaging apparatus 10 according to the first embodiment of the present invention. The imaging apparatus 10 of the present embodiment corresponds to a so-called digital camera, and the internal configuration thereof is an imaging processing unit 12 that generates an image signal related to a captured image, an image compression unit 13 that performs image compression processing, and an image display. It is roughly divided into a monitor unit 14. In addition, the imaging device 10 can store an image file related to the image captured in the external memory 15.

  The imaging processing unit 12 includes an imaging element 12a and an A / D conversion unit 12b. The imaging element 12a performs photoelectric conversion based on a condensed image through an optical system including an imaging lens (not shown) and generates an analog image signal related to the image. The A / D conversion unit 12b performs a process of removing a noise component included in the analog image signal, performs a signal amplification process as necessary, converts the digital image signal, and converts the digital image signal into an image compression unit. 13 to send. Note that the image signal sent out by the imaging processing unit 12 has one color per pixel and a signal amount per pixel of 10 bits according to the color filter array of the imaging element 12a, and is formed by a plurality of pixels. The size of the image to be displayed is, for example, an image range 6a of 1344 × 971 as shown in FIG.

  The image compression unit 13 corresponds to an image compression device, and is different from the image compression unit 3 in the conventional imaging device 1 of FIG. 12 in that the contrast position detection unit 11 is provided. The image compression unit 13 has a configuration in which the processing units 13a and 13b, etc., and the memory 13e and the MPU 13f are connected by an internal bus 13h. I do.

  The sequential processing unit 13a performs various correction processes such as black level correction, white balance correction, and gamma correction on the image signal as part of the camera signal processing. Note that the black signal level in the 52 × 5 optical black portion 7 included in the captured image signal shown in FIG. 13A is not usually “0”, but the optical level is sequentially corrected by the black level of the processing unit 13a. By subtracting the average value of the signal level of the black portion 7 from the output level of the entire imaging processing unit 12, the black color of the optical black portion 7 is reproduced as the reference level “0”. Further, when the black level correction is completed, the optical black portion 7 becomes unnecessary, and thus the processing unit 13a sequentially performs a process of deleting the optical black portion 7.

  Therefore, the image signal processed by the sequential processing unit 13a is compressed to 8 bits in signal amount per pixel, and the image size becomes, for example, a first processing range 6b of 1292 × 966 as shown in FIG. . The sequential processing unit 13a outputs the processed image signal to the memory 13e.

  The filter processing unit 13b reads the image signal sequentially processed by the processing unit 13a from the memory 13e, and performs RGB interpolation synchronization processing, RGB-YCC matrix processing, edge correction processing, and the like as other camera signal processing. The filter processing 13b converts the image signal into a total of three signal components of the luminance signal Y, the red color difference signal Cr, and the blue color difference signal Cb by each processing. By this conversion, the signal amount per pixel of the signal component related to the image signal becomes 16 bits (the luminance signal Y is 8 bits, both color difference signals Cr and Cb are 8 bits).

  Further, the filter processing unit 13b of the present embodiment applies finite impulse response filter processing (hereinafter referred to as FIR processing) to RGB interpolation simultaneous processing, edge correction processing, and the like, and effective filtering with a small number of delay elements. Is realized. Since the FIR processing is performed between adjacent pixels, there is no adjacent pixel among the pixels located in the peripheral edge portion of the image, and therefore effective filtering processing cannot be performed. Therefore, the filter processing unit 13b reduces the image size of the image signal by processing related to filtering. For example, by performing secondary FIR processing in the horizontal and vertical directions, the effective portion is reduced by two pixels in the horizontal and vertical directions. Then, the image size is set to the second processing range 6c of 1288 × 962 as shown in FIG.

  The filter processing unit 13b sequentially outputs the luminance signal Y of each pixel to the contrast position detection unit 11 with pixel position information added among the three signal components related to the converted image signal. The processed image signal is output to the memory 13e.

  FIG. 2 is a block diagram illustrating an internal configuration of the contrast position detection unit 11. The contrast position detection unit 11 corresponds to a processing location identifier, and includes a horizontal contrast position detection unit 20 and a vertical contrast position detection unit 30 for performing processing separately for each pixel in the horizontal and vertical directions. Since the basic configurations of the contrast position detection units 20 and 30 are the same, the horizontal contrast position detection unit 20 will be described as a representative.

The horizontal contrast position detection unit 20 includes a one-pixel delay unit 21, a horizontal difference unit 22, a horizontal conversion unit 23, a horizontal selector unit 24, integration units 25a to 25h, and a horizontal detection unit 26.
The horizontal contrast position detection unit 20 receives the luminance signal Y related to the horizontal pixels sequentially output from the filter processing unit 13 b by the horizontal difference unit 22 and the one-pixel delay unit 21. The horizontal subtractor 22 calculates the difference between adjacent horizontal pixels with respect to the luminance signal Y received by itself as the calculating means and the luminance signal Y received by the one-pixel delay unit 21 and delayed by one pixel. In addition, since a negative value is generated depending on the calculation result, the horizontal difference unit 22 also performs processing for converting the calculated result into an absolute value, and outputs the absolute value difference to the horizontal conversion unit 23. Note that the horizontal contrast position detection unit 20 adds pixel position information indicating what kind of pixel position is the difference between the pixel positions to be output.

  The horizontal conversion unit 23 increases or decreases according to the difference in order to nonlinearly convert the difference from the horizontal difference unit 22 as an increase / decrease unit. Specifically, the horizontal conversion unit 23 stores data related to the curve K having input / output characteristics as shown in the graph of FIG. The obtained difference (input) is non-linearly transformed and output. In the graph of FIG. 3, the horizontal axis indicates the signal level (0 to 255) of the input difference, and the vertical axis indicates the signal level (0 to 255) of the output difference (conversion value). By such non-linear conversion, compression conversion is performed so that the luminance is low when the input difference is small, and expansion conversion is performed so that the luminance is high when the input difference is large.

  As described above, the horizontal conversion unit 23 appropriately increases or decreases the difference, thereby facilitating detection processing in the horizontal detection unit 26 described later. In addition, the increase / decrease in the difference in the horizontal conversion unit 23 can be performed not only by nonlinear conversion based on the curve K but also based on straight lines, bent lines, and the like. Further, the horizontal conversion unit 23 outputs the converted difference to the horizontal selector unit 24.

  The horizontal selector 24 separates the difference output from the horizontal converter 23 into 8 pixels that are the number of pixel units related to the JPEG compression process based on the information about the pixel position added to the difference. Function as. For example, if the first pixel from the left in the horizontal direction in the processing target image is set to “0” pixel, the pixel of JPEG compression processing is between the eighth “7” pixel and the ninth “8” pixel. Since the unit boundary is located, the difference between the “7” pixel and the “8” pixel output from the horizontal conversion 23 is classified into a 0th group called “8n + 0” (n is an integer, the same applies hereinafter).

  Continuing the description with the above example, the horizontal selector 24 changes the difference between the “8” pixel and the “9” pixel to the first group “8n + 1”, and the difference between the “9” pixel and the “10” pixel. To a second group called “8n + 2”, a difference between “10” pixels and “11” pixels to a third group called “8n + 3”, and a difference between “11” pixels and “12” pixels to “8n + 4”. The fourth group is called “8n + 5”, the difference between “12” pixels and “13” pixels is the fifth group, and the difference between “13” pixels and “14” pixels is “8n + 6”. The difference between the “14” pixel and the “15” pixel is divided into the seventh group called “8n + 7”, and the difference between the “15” pixel and the “16” pixel is again sorted into the “8n + 0” zeroth group. To do.

  The above-described classification forms are generally summarized as follows. The difference between the “8n−1” pixel and the “8n + 0” pixel is sorted in the 0th group of “8n + 0”, and the difference between the “8n + 0” pixel and the “8n + 1” pixel is sorted in the first group of “8n + 1”. The difference between the “8n + 1” pixel and the “8n + 2” pixel is sorted into the second group of “8n + 2”, and the difference between the “8n + 2” pixel and the “8n + 3” pixel is sorted into the third group of “8n + 3”. The difference between the “8n + 3” pixel and the “8n + 4” pixel is sorted into the fourth group of “8n + 4”, and the difference between the “8n + 4” pixel and the “8n + 5” pixel is sorted into the fifth group of “8n + 5”. The difference between the “8n + 5” pixel and the “8n + 6” pixel is classified in the sixth group “8n + 6”, and the difference between the “8n + 6” pixel and the “8n + 7” pixel is classified in the seventh group “8n + 7”. It is.

  Since there is no pixel on the left side of the leftmost “0” pixel, only the difference between the “0” pixel and the “1” pixel is classified into the first group for the “0” pixel. Further, since the rightmost pixel does not exist on the right side, the difference regarding the rightmost pixel is classified into only the seventh group.

  The zeroth integration unit 25a to the seventh integration unit 25h each store the difference sorted by the horizontal selector unit 24, and each integration unit 25a to 25h is provided corresponding to the 0th group to the 7th group. The difference sorted into each group is stored. That is, the 0th integration unit 25a stores the difference sorted into the 0th group. Hereinafter, the first integration unit 25b is the first group, the second integration unit 25c is the second group, and the third integration unit 25d is the third group. The group, the fourth integration unit 25e stores the difference sorted into the fourth group, the fifth integration unit 25f stores the difference into the fifth group, the sixth integration unit 25g stores the sixth group, and the seventh integration unit 25h stores the difference sorted into the seventh group. .

  Each of the integrating units 25a to 25h integrates and stores the differences that are sorted and output by the horizontal selector unit 24. This integration can be performed sequentially every time it is output, or it can also be integrated collectively at the end after the difference relating to the entire image is output.

  The horizontal detection unit 26 corresponds to a comparison unit that compares the differences stored in the integration units 25a to 25h, a detection unit that detects the maximum difference, and each integration unit 25a to 25h that stores the detected maximum difference. It functions as a location specifying means for specifying a location related to the compression processing based on the number of pixel units.

  Specifically, the horizontal detection unit 26 compares the differences integrated by the integration units 25a to 25h, and detects the maximum difference by this comparison. Further, the horizontal detection unit 26 specifies which of the integration units 25a to 25h that stores the detected maximum difference, and sets the numerical value related to the group to which the specified integration units 25a to 25h correspond to the portion to be compressed. The offset position is output to the MPU 13f shown in FIG. The numerical value relating to the above group means an integer “m” in “8n + m (m is an integer of 0 to 7)” corresponding to each group. For example, the first integration unit 25b stores the maximum difference. In this case, since the first integrating unit 25b corresponds to the first group, the integer “1” is output from “8n + 1” related to the first group. The horizontal detection unit 26 determines whether or not the differences with respect to all the pixels of the image have been accumulated in the integration units 25a to 25b, and then performs the above-described processing such as comparison and maximum value detection.

  2 includes a one-line delay unit 31, a vertical difference unit 32, a vertical conversion unit 33, a vertical selector unit 34, integration units 35a to 35h, and a vertical detection unit 36. And an integer of the offset position relating to the portion to be compressed in the vertical pixel is output to the MPU 13f. The one-line delay unit 31 receives the luminance signal Y related to the pixel delayed by one line because the pixel related to the luminance signal Y to be processed is a vertical pixel.

  The MPU 13f shown in FIG. 1 performs a process of cutting out the range to be compressed by the JPEG processing unit 13c from the filtered image stored in the memory 13e, in addition to the control of the processing units 12 and 13 and the like. Although the extraction of the compression target range has conventionally been performed from a fixed location, the MPU 13f of the present embodiment offsets the extraction location based on the integers related to the horizontal and vertical pixels output from the contrast position detection unit 11. And adjust.

  For example, when the contrast position detection unit 11 outputs an integer “1” for the horizontal pixel and an integer “1” for the vertical pixel, the MPU 13f is the origin coordinate pixel that is the upper left vertex of the image stored in the memory 13e. A process of cutting out a range of a required image size using the coordinate pixel (1, 1) offset by “1” in the right direction and “1” in the downward direction from (0, 0) as the upper left vertex of the image to be JPEG compressed. Do. By performing such cutout adjustment by the MPU 13f, even if a difference in contrast occurs within an 8 × 8 pixel unit related to the compression processing as in the image of FIG. 14A in a part of the image, FIG. a) As shown in the image shown in (d), a portion where a difference in contrast from the background portion is caused by the patch coincides with a boundary of 8 × 8 pixel units. Note that the image in FIG. 6D has the number of horizontal and vertical pixel units (8 × 8) related to the JPEG compression process (two-dimensional discrete cosine transform: DCT) with respect to the image in FIG. The broken lines are shown in a grid pattern, and the broken lines do not exist in the actual image.

  The JPEG processing unit 13c shown in FIG. 1 corresponds to an image compression unit, and performs JPEG compression processing on the image cut out by the MPU 13f as described above, and converts it into a JPEG image stream signal. Specifically, the JPEG processing unit 13 performs a series of JPEG compression processing such as level shift processing, two-dimensional discrete cosine transform processing, quantization, zigzag sequencing, and Huffman coding on an image. The JPEG processing unit 13c outputs the compressed image to the memory 13e and stores it in the memory 13e. The MPU 13f outputs the compressed image stored in the memory 13e to the external memory interface unit 13g with various necessary data added as a JPEG image file.

The external memory interface unit 13g outputs a JPEG image file to the external memory 15 and stores it in the external memory 15 when the external memory 15 which is a semiconductor memory such as a memory card conforming to each standard is set. I do.
The bus management unit 13d of the image compression unit 13 manages the internal bus 13h. The monitor unit 14 shown in FIG. 1 receives the luminance signal Y processed by the filter processing unit 13b and the image signals converted into the color difference signals Cr and Cb, and displays an image.

Next, a series of processes until the image captured by the above-described imaging device 10 is compressed and stored in the external memory 15 will be described based on the first flowchart of FIG.
First, the image compression unit 13 receives an image signal related to an image captured by the imaging processing unit 12 (S1). The image compression unit 13 temporarily stores the received image signal in the memory 13e. Next, the sequential processing unit 13a of the image compression unit 13 reads the image signal from the memory 13e, and performs various correction processes such as level correction and gamma correction (S2). By this correction processing, the optical black portion is deleted. The sequential processing unit 13a temporarily stores the corrected image signal in the memory 13e.

  Then, the filter processing unit 13b of the image compression unit 13 reads the image signal from the memory 13e and performs the filter process (S3). The image size is reduced by this filtering process. Further, the filter processing unit 13b temporarily stores the filtered image signal in the memory 13e, and sends only the luminance signal Y included in the image signal to the contrast position detection unit 11.

  The contrast position detection unit 11 performs compression processing to specify the compression portion that outputs an integer related to the extraction portion specification to the MPU 13f by performing each process on the luminance signal Y (S4). Also, the MPU 13f sends the range cut out from the image stored in the memory 13e to the JPEG processing unit 13c with the part specified by the integer from the contrast position detection unit 11 as the upper left vertex, and the JPEG processing unit 13c is cut out. The required image compression processing is performed on the range (S5).

  The JPEG processing unit 13c outputs the compressed image subjected to the compression process to the memory 13e. Finally, the MPU 13f adds the necessary data as the JPEG image file to the compressed image stored in the memory 13e and adds the external memory interface unit 13g. To the external memory 15 (S6). As a result, the JPEG image file relating to the compressed image is stored in the external memory 15. It should be noted that the step (S4) of specifying the compression point and the step (S5) of the image compression process described above are the processing contents mainly corresponding to the image compression method according to the present invention.

The second flowchart of FIG. 5 shows in detail the processing procedure related to the processing location specifying method performed by the contrast position detection unit 11 for the compression location specifying stage (S4) of the first flowchart of FIG.
The contrast position detection unit 11 in FIG. 2 first receives the luminance signal Y output from the filter processing unit 13b for the horizontal pixels by the one-pixel delay unit 21 and the horizontal difference unit 22 and converts them to the vertical pixels. On the other hand, the reception is performed by the one-line delay unit 31 and the vertical difference unit 32 (S10). Next, each of the difference units 22 and 32 calculates a difference with respect to the luminance signal Y between adjacent pixels (S11), and outputs the difference related to the calculation result to each of the conversion units 23 and 33.

  Each of the conversion units 23 and 33 performs non-linear conversion that appropriately increases or decreases the output difference (S12), and outputs the increased or decreased difference to each of the selector units 24 and 34. Each selector unit 24, 34 classifies the difference into pixel positions related to each difference, and outputs them to each integrating unit 25a-25h, 35a-35h (S13). The integration units 25a to 25h and 35a to 35h accumulate and store the output differences, and detect whether or not the difference obtained by the integration of the detection units 26 and 36 is stored for all pixels. It is determined whether the processing for the entire image has been completed (S14).

  If the processing for the entire image has not been completed (S14: NO), the step of outputting the difference from the luminance signal reception (S10) to each integrating unit 25a to 25h, 35a to 35h (S13) for the entire image. Repeat until done. On the other hand, when the processing for the entire image is completed (S14: YES), the detection units 26 and 36 detect the maximum value from the integrated values stored in the integration units 25a to 25h and 35a to 35h (S15), The integers for horizontal and vertical pixels corresponding to the integrating units 25a to 25h and 35a to 35h storing the maximum value are output to the MPU 13f (S16).

  When the MPU 13f accepts an integer obtained by the process in the second flowchart of FIG. 5 described above, the MPU 13 cuts out the image as shown in FIGS. 6A and 6D, for example, and a difference in contrast occurs in the image. Are made to coincide with the boundary relating to the pixel unit of the compression processing.

  Further, the low-compressed image in FIG. 6B is a JPEG-compressed image obtained by adjusting the cutout portion in FIG. 6A at a low compression rate, and the high-compressed image in FIG. The image shown in (a) is JPEG compressed at a high compression rate. The low-compressed image in FIG. 6B has a smaller difference in the situation around the patch than the conventional low-compressed image shown in FIG. 14B, but the high-compressed image in FIG. Compared to the conventional high-compression image shown in c), it can be seen that there is no distortion around the patch and a good image is obtained.

  This is because, by adjusting the cut-out position based on the processing of the contrast position detection unit 11, the location where the contrast difference is large coincides with the boundary of 8 × 8 pixel units related to JPEG compression, so that the two-dimensional discrete cosine transform is performed. This is because there is no difference in contrast within the processing unit and the harmonic component of the image is eliminated, and the influence of removing the high frequency component by the quantization of the JPEG processing unit 13c is eliminated.

  Note that by adjusting the cutout range as described above, the size of the cutout image may be smaller than in the case where the conventional cutout location is fixed. Therefore, the resolution of the monitor unit included in the digital camera does not affect the image having a large number of pixels. Further, if the number of pixels is such that the entire image can be accommodated in the monitor unit, the above-described influence does not become obvious. Furthermore, by adjusting the cutout range, the center of the image and the center of the optical axis related to imaging may not match, but since the offset dimension is within half the number of pixel units at the maximum from the matched state, There is no particular problem with an imaging device or the like.

  Further, the imaging apparatus 10 according to the first embodiment is not limited to the above-described form, and there are various modifications. For example, the image compression unit 13 converts the JPEG processing unit 13c into an MPEG compression unit that performs image compression processing using two-dimensional discrete cosine transform, or H.264. It can be replaced with a moving image compression unit such as H.262. When the JPEG processing unit 12c is replaced with an MPEG compression unit, the imaging device 10 is equivalent to a digital video camera. When the moving image compression unit such as 262 is replaced, the imaging device 10 corresponds to a camera compatible with a videophone or the like. The imaging device 10 of the first embodiment can also be applied to a mobile phone device and a fixed phone device having an imaging function.

  Furthermore, the image compression unit 13 shown in FIG. 1 can be separated from the imaging processing unit 12 and the monitor unit 14 to form a single image compression apparatus. In this case, an external image input unit is provided in the image compression apparatus having the same configuration as that of the image compression unit 13 in FIG. 1, and a device that outputs an image signal is connected to the external signal input unit. Can be performed by the image compression apparatus according to the present invention.

  Furthermore, the number of integrating units 25a to 25h and 35a to 35h included in the horizontal and vertical contrast position detecting units 20 and 30 is not limited to eight, and the pixels in the horizontal and vertical directions of the image compression process are not limited to eight. It is possible to increase or decrease appropriately according to the number of units. For example, as described above, when an MPEG compression unit or a moving image compression unit is used instead of the JPEG processing unit 13c, the compression processing is performed in units of 16 pixels. Sixteen integration units are provided. In this case, the selectors 24 and 34 are also separated for every 16 pixels, and classified from the 0th integration unit corresponding to the 0th group “16n + 0” to the 15th integration unit corresponding to the 15th group “16n + 15”. Will output the difference.

  In the above description, the component signal processed by the contrast position detection unit 11 is the luminance signal Y. However, the luminance signal Y can be processed by combining both or one of the color difference signals Cr and Cb. Furthermore, the luminance signal Y may not be used and processing may be performed using both or one of the color difference signals Cr and Cb. In such a case, the filter processing unit 13b outputs the component signal to be used to the contrast position detection unit 11. Further, when the ratio of each signal component is so-called “YCC422”, it is preferable to increase the number of horizontal pixels of the integration unit for the luminance signal Y from 8 to 16 because the matching of each color difference signal is improved. is there. Furthermore, when the ratio of each signal component is so-called “YCC420”, the matching of each color difference signal can be improved by increasing the number of vertical pixels of the integration unit to the luminance signal Y to 16.

  As described above, when the color difference signals Cr and Cb are used in addition to the luminance signal Y, a contrast position detection unit for luminance signals and a contrast position detection unit for color difference signals Cr and Cb are provided. Also good.

  Further, the image compression unit 13 shown in FIG. 1 can be formed into a single chip except for the memory 13e, and can be formed into a single chip except for the memory 13e and the contrast position detection unit 11. is there. In the latter case, a semiconductor chip (for example, LR38668 manufactured by Sharp Corporation) that sequentially performs processing such as the processing unit 13a and the filter processing unit 13b can be applied. Note that only the contrast position detecting unit 11 having the structure shown in FIG. When the configuration of the image compression unit 13 is simplified, the conversion units 23 and 33 may be omitted.

  FIG. 7 is a block diagram showing an internal configuration of the image compression apparatus 50 according to the second embodiment of the present invention. The image compression device 50 is substantially the same in hardware as the internal configuration of the image compression unit 3 in the imaging device 1 shown in FIG. 12, and is different from the internal bus 53h in the external input unit 53i and the external output unit 53j. Are connected, the flash memory 53k is connected to the MPU 53f, and the external memory interface unit 3g is omitted. In the image compression apparatus 50 of the present embodiment, the MPU 53f performs the processing of the contrast position detection unit 11 in the image compression unit 13 of the first embodiment shown in FIG.

  An MPU 53f having an internal memory unit 51 is applied to the image compression device 50, and the required address portion of the internal memory unit 51 is used as a register, and the 0th horizontal buffer 58a to the 7th horizontal buffer 58h and the 0th vertical buffer 59a to 5th shown in FIG. A seventh vertical buffer 59h is provided. The horizontal buffers 58a to 58h and the vertical buffers 59a to 59h correspond to the integrating units 25a to 25h of the horizontal contrast position detecting unit 20 and the integrating units 35a to 35h of the vertical contrast position detecting unit 30 shown in FIG. Is.

  A program for causing the MPU 53f to perform processing related to the processing units 21 and 22 of the horizontal contrast position detection unit 20 and the processing units 31 and 32 of the vertical contrast position detection unit 30 is stored in the flash memory 53k. Yes. The portions other than those described above of the image compression apparatus 50 are the same as those of the image compression unit 13 of the first embodiment, and the sequential processing unit 53a, filter processing unit 53b, JPEG processing unit 53c, and bus management unit 53d are the first embodiment. Processing similar to that of the form is performed.

  The program stored in the flash memory 53k is divided into a horizontal pixel processing portion and a vertical pixel processing portion. The horizontal pixel processing portion processes the luminance signal Y related to the horizontal pixel output from the filter processing portion 53b, and the vertical pixel processing portion processes the luminance signal Y related to the vertical pixel output from the filter processing portion 53b, and the program contents of each portion Are approximately equivalent.

The horizontal and vertical pixel processing section includes a one-pixel delay means for processing the MPU 53f for horizontal and vertical pixels (one-line delay means for vertical pixels, the same applies hereinafter), a calculation means, an increase / decrease means, a sorting means, and a specifying means. The contents to function as are programmed.
The MPU 13f sequentially receives the luminance signal Y related to the horizontal and vertical pixels output from the filter processing unit 13b based on the activation of the program. The situation related to the reception is that the luminance signal Y related to the adjacent pixel is calculated by means of a calculation unit and one pixel. Each is accepted as a delay means.

  The processing contents related to the calculation means are equivalent to the processing of each of the difference units 22 and 32 shown in FIG. 2, and the luminance signal Y received as the calculation means and the one-pixel delay output of the luminance signal Y received as the one-pixel delay means It is prescribed to calculate the difference between the two. In addition, for the calculation of the difference, contents for converting the calculation result into an absolute value are also defined.

  The increase / decrease means performs processing equivalent to that of each of the conversion units 23 and 33 in FIG. 2, and the contents for performing nonlinear conversion on the absolute difference based on the curve K in the graph shown in FIG. 3 are defined. In the program, data relating to the curve K is also defined in advance. Specifically, when the difference to be processed is small, it is converted into a low luminance value, and when the difference is large, it is converted into a high luminance value. By performing non-linear conversion in this way, it becomes possible to reliably detect a high luminance value in the case where there are a large number of low luminance values in the processing described later.

  The sorting means corresponds to each selector unit 24, 34 in FIG. 2, and sorts the converted difference for each number of pixel units related to image processing such as compression based on pixel position information related to the difference. It is prescribed to process. In this embodiment, each of the differences is sorted and output to the buffers 58a to 58h and 59a to 59h shown in FIG. 8 corresponding to the 0th group to the 7th group, respectively, like the selector units 24 and 34 of the first embodiment. To do.

Each of the buffers 58a to 58h and 59a to 59h accumulates and stores the output differences, and the correspondence between each buffer 58a and the like and each group is the same as in the first embodiment.
Further, the specifying unit corresponds to each of the detection units 26 and 36 in FIG. 2, and functions as a comparison unit, a detection unit, and a location specifying unit. That is, the specifying unit compares the difference accumulated in each buffer 58a and the like to detect the maximum value, specifies which buffer stores the maximum value, and according to the number of pixel units related to the specified buffer. Processing to determine the integer.

  Further, a part to be cut out from the image processed by the filter processing unit 53b stored in the memory 53e is specified based on the integer determined by the MPU 53f as the specifying means, and a required dimension is cut out from the specified part. When the program adjusts the cutout range in software, when starting the processing related to each means described above, the buffers 58a to 58h and 59a to 59h are initialized and all the pixels are stored in the buffers 58a to 58h and 59a to 59h. It is specified that the MPU 13f confirms whether or not the minute difference is stored.

  The processing for the clipped range is the same as in the first embodiment, and the JPEG processing unit 53c performs image compression on the clipped range, and stores the generated compressed image in the memory 53e.

FIG. 9 is a third flowchart showing a series of processing procedures related to a method for specifying a processing location for compressing the MPU 53f based on the above-described program.
In the third flowchart, the image compression apparatus 50 of FIG. 7 sequentially processes the image signal received from the external input unit 53i by the processing unit 53a and the filter processing unit 53b, and the filtered image signal is stored in the memory 53e. In addition, when the luminance signal Y is output from the filter processing unit 53b to the MPU 53f, the MPU 53f sequentially performs the contents defined in the program.

  First, when starting the processing, the MPU 53f initializes the horizontal buffers 58a to 58h and the vertical buffers 59a to 59h and deletes the stored contents (S20). Next, the MPU 53f receives the luminance signal Y related to the horizontal and vertical pixels from the filter processing unit 53b (S21), and calculates a difference between adjacent pixels as a calculation means (S22). Further, the MPU 53f performs non-linear conversion as an increase / decrease means in order to increase / decrease the calculated difference as appropriate (S23).

  Then, the MPU 53f sorts each difference for each number of pixel units based on the pixel position information included in the difference converted as the sorting means (S24). Then, the respective buffers 58a to 58h and 59a to 59h corresponding to the sorted differences are stored (S25), and the stored differences are integrated.

  In this state, the MPU 53f detects the stored contents of the buffers 58a to 58h and 59a to 59h to determine whether the buffers 58a to 58h and 59a to 59h have stored the differences relating to all the pixels (S26). When the differences relating to all the pixels are not stored (S26: NO), the process returns to the step (S21) of accepting the luminance signal, and each difference is stored from the accepting step (S21) of the luminance signal until the difference relating to all the pixels is memorized. Repeat step (S25).

  Further, when the differences relating to all the pixels are stored in each of the buffers 58a to 58h and 59a to 59h (S26: YES), the MPU 53f uses the accumulated differences stored in the buffers 58a to 58h and 59a to 59h as comparison means. The comparison is made (S27), the maximum difference is detected as a detecting means, and the buffer storing the maximum difference is specified (S28). Further, the MPU 53f specifies the cutout part of the image to be compressed as the part specifying unit by obtaining an integer corresponding to the number of pixel units corresponding to the specified buffer (S29).

  Note that the MPU 53f cuts out the required dimensions from the image stored in the memory 53e as described above with the cut-out location specified as the upper left vertex, and the JPEG processing unit 53c generates the cut-out range by compressing it. The compressed image is temporarily stored in the memory 53e. The compressed image stored in the memory 53e is read by the MPU 53f in response to an external request or the like, and is output to the outside via the external output unit 53j.

  In this way, the MPU 53f specifies the processing location corresponding to the compression based on the program stored in the flash memory 53k, so that the image processed by the JPEG processing unit 53c has a contrast difference related to at least the maximum luminance difference. Since the pixel portion coincides with the boundary of the number of pixel units in the compression process, the occurrence of large distortion due to JPEG compression is prevented. In addition, the image compression apparatus 50 according to the second embodiment has a simplified configuration as long as the contrast position detection unit 11 does not exist as compared with the image compression unit 13 of the first embodiment.

  Note that the image compression apparatus 50 according to the second embodiment is not limited to the above-described form, and various modifications can be applied. For example, the program for specifying the processing location may be stored not in the flash memory 53k but in the memory 53e or the internal memory 51. Further, each of the buffers 58a to 58h and 59a to 59h shown in FIG. 8 may be provided with a storage area in the memory 53e or the flash memory 53k. Further, a memory for each buffer is newly provided in the internal bus 53h. You may make it connect. Note that the flash memory 53k can be omitted if the above-described program, various data, and the like are not stored in the flash memory 53k.

  In the second embodiment, the same modifications as in the first embodiment can be applied. For example, the JPEG processing unit 53c can be replaced with a compression unit using two-dimensional discrete cosine transform such as an MPEG compression unit. The number of the buffers 58a to 58h and the number of the buffers 59a to 59h may be 16 each for horizontal and vertical. Furthermore, in addition to the luminance signal Y, each of the color difference signals Cr, Cb, etc. Processing may be performed. It should be noted that the deformation processing related to the above-described processing location identification can be easily handled by appropriately changing the program stored in the flash memory 53k.

  Furthermore, the image compression apparatus 50 can be combined with an image pickup unit that outputs an image signal and a monitor unit that displays an image, thereby forming an image pickup apparatus such as a digital camera and a digital video camera as a whole. It can also be applied to a mobile phone equipped with an imaging function.

FIG. 10 is a block diagram showing an internal configuration of a computer 60 corresponding to the image compression apparatus according to the third embodiment of the present invention.
The computer 60 has a general configuration, and a CPU 60a, a memory 60b, an external storage medium access unit 60c, a monitor connection unit 60d for connecting a monitor device 61, an external connection unit 60e, and a hard disk 60f are connected by an internal bus 60g. is doing. The external storage medium access unit 60c corresponds to an access unit for a floppy disk, an access unit for a semiconductor memory, an access unit for an optical disk such as a CD and a DVD, and the external connection unit 60e is a serial connection unit, A USB connection unit, an IEEE 1394 connection unit, a LAN port, a communication unit such as a modem, and the like correspond to this.

  In the computer 60, the CPU 60a also performs each processing related to the sequential processing unit 53a, the filter processing unit 53b, and the JPEG processing unit 53c in the image compression apparatus 50 of the second embodiment shown in FIG. 7 based on the computer program stored in the hard disk 60f. Thus, the contents for specifying the processing location related to compression are also included in the computer program.

  Therefore, the computer program causes the CPU 60a to function so as to sequentially perform the processes related to the processing unit 53a, the filter processing unit 53b, and the JPEG processing unit 53c, and, similarly to the second embodiment, a one-pixel delay unit, a calculation unit, an increase / decrease unit The contents that function as the sorting means and the specifying means are programmed. The processing when the CPU 60a functions as the processing units 53a to 53c is the same as the processing performed by the processing units 53a to 53c in the second embodiment.

  Further, the computer program portion of the content for specifying the processing location is substantially the same as the program stored in the flash memory 53k of the image compression apparatus 50 of the second embodiment, but when the CPU 60a starts image processing, FIG. A process of providing each of the buffers 58a to 58h and 59a to 59h in the memory 60b as shown in FIG. The number of buffers provided by the CPU 60a is the number of pixel units related to image compression as in the second embodiment. When JPEG compression is performed, a total of 8 horizontal buffers for horizontal pixels and a total of 8 vertical buffers for vertical pixels are used. Provide a buffer.

  FIG. 11 shows an example of an algorithm for the computer program part related to the processing location identification of the third embodiment. In this algorithm, x is the number of horizontal pixels of the original image, y is the number of vertical pixels of the original image, i is the horizontal position of the pixel to be processed (i is an integer from 0 to x−1), and j is the target of processing. Pixel vertical position (j is an integer from 0 to y-1), bh [k] is a horizontal buffer (k is an integer from 0 to 7), bv [l] is a vertical buffer (l is an integer from 0 to 7), g (j, i) is the signal level of the pixel at the vertical position j and the horizontal position i, c (j, i) is a primary variable for holding the signal level, m is a horizontal cut-off offset value, n is a vertical cut-off offset value,% Is an operator for calculating the remainder, read () is a function for reading the signal level of the pixel at the designated position from the memory 60 b, and non_linear_transform () is a non-linear conversion function.

  In the algorithm of FIG. 11, first, each buffer that accumulates and stores the difference in contrast between pixels is initialized in (Sentence 1) and (Sentence 2). The horizontal buffer is set to “0” in (Sentence 1) using the for loop statement, and the vertical buffer is set to “0” in (Sentence 2). Next, in (Sentence 3) to (Sentence 15), contrast difference accumulation between pixels is performed between pixels of the entire original image. This has a double structure in the horizontal and vertical directions using a for loop statement.

  Specifically, (Sentence 3) is a vertical for loop sentence, and (Sentence 5) is a horizontal for loop sentence. (Sentence 4) is performed when the vertical position to be analyzed by the remainder calculation specifying the vertical buffer is determined. (Sentence 6) to (Sentence 8) read the target pixel, (Sentence 9) calculates the contrast difference between adjacent horizontal pixels, and converts it into an absolute value. (Sentence 13) An absolute value is obtained by calculating a difference in contrast between adjacent vertical pixels. (Sentence 10) and (Sentence 14) perform non-linear transformation on the calculated difference, and the input / output characteristics of this non-linear transformation conform to the curve K of the graph shown in FIG. (Sentence 11) is a remainder operation for designating a horizontal buffer, and (Sentence 12) and (Sentence 15) are for accumulating and storing the non-linearly converted difference in each designated buffer.

  When the contrast difference integration is stored in the pixels of the entire original image in the processes of (Sentence 3) to (Sentence 15) described above, the buffer that stores the next maximum integration value is specified. This is done by detecting the contents of all buffers using a for loop statement. (Sentence 16) to (Sentence 19) are processes for horizontal pixels, and (Sentence 20) to (Sentence 23) are processes for vertical pixels. (Sentence 17) and (Sentence 21) are for comparing magnitude relations between adjacent buffers using an if sentence, and in (Sentence 18) (Sentence 19) (Sentence 22) (Sentence 23), a large variable m, A process of holding n is performed.

(Sentence 24) specifies the buffer that finally holds the maximum difference integrated value, and uses the values of the variables m and n as return values for the offset value related to extraction.
By specifying the cut-out location from the image stored in the memory 60b based on the values of the variables m and n obtained as described above, the boundary between pixels that cause a strong contrast difference and the pixel unit related to JPEG compression are matched. It becomes possible to make it. Note that the algorithm in FIG. 11 describes basic processing contents, and is not completed as source code.

Note that the computer 60 accepts an image to be processed from the outside through a storage medium set in the external storage medium access unit 60c or the external connection unit 60e.
Therefore, when the computer 60 starts processing related to image compression for an image received based on the computer program described above, first, each buffer is provided in the memory 60b and the sequential processing unit of the second embodiment is applied to the received image. The CPU 60a performs processing corresponding to 53a and the filter processing unit 53b, temporarily stores the filtered image in the memory 60b, and performs processing based on the above-described algorithm of FIG. A processing location related to compression is specified.

Then, the CPU 60a reads out an image from the memory 60b, cuts out a required range with the location specified as described above as the upper left vertex, and performs JPEG compression processing on the cut out range. The compressed image obtained by the compression process is stored in the memory 60b or the hard disk 60f. The compressed image stored in this way can be output to the external storage medium access unit 60c under the control of the CPU 60a in response to an operation request from the outside, and can also be output to the outside via an external connection unit. It is.
The processing performed by the computer 60 according to the third embodiment can be applied to various modifications according to the second embodiment, and various modifications can be applied by appropriately changing the computer program stored in the hard disk 60f. Can be easily handled.

1 is a block diagram illustrating an internal configuration of an imaging apparatus according to a first embodiment of the present invention. It is a block diagram which shows the internal structure of the contrast position detection part of 1st Embodiment. It is a graph which shows the curve which concerns on a nonlinear transformation. It is a 1st flowchart which shows the process sequence which concerns on an image compression method. It is a 2nd flowchart which shows the process sequence which concerns on a process location identification method. (A) is a schematic diagram of an uncompressed image whose cutout range is adjusted, (b) is a schematic diagram of a low-compressed image, (c) is a schematic diagram of a high-compressed image, and (d) is a compression process in which a contrast difference is compressed It is the schematic of the image which shows the state which corresponds to the boundary of an 8x8 pixel unit concerning. It is a block diagram which shows the internal structure of the image compression apparatus which concerns on 2nd Embodiment of this invention. It is the schematic which shows a horizontal buffer and a vertical buffer. It is a 3rd flowchart which shows the processing content of the program which concerns on a process location specification. It is a block diagram which shows the internal structure of the computer which concerns on 3rd Embodiment of this invention. It is an example which shows the algorithm which concerns on a process location specification. It is a block diagram which shows the internal structure of the conventional imaging device. (A)-(c) is the schematic which shows the range which the image dimension changes through each process of the imaged image. (A) is a schematic diagram of an image before compression, (b) is a schematic diagram of a low-compression image, (c) is a schematic diagram of a high-compression image, and (d) is an 8 × 8 pixel whose contrast difference is related to compression processing. It is the schematic of the image which shows the state which does not correspond to the boundary of a unit.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Imaging device 11 Contrast position detection part 12 Imaging processing part 13 Image compression part 13a Sequential processing part 13b Filter processing part 13c JPEG processing part 13d Bus management part 13e Memory 13f MPU
13g External memory interface unit 13h Internal bus 14 Monitor unit 15 External memory 20 Horizontal contrast position detection unit 21 1 pixel delay unit 22 Horizontal difference unit 23 Horizontal conversion unit 24 Horizontal selector units 25a to 25h 0th integration unit to 7th Integration unit 26 Horizontal detection unit 30 Vertical contrast position detection unit 31 1-line delay unit 32 Vertical difference unit 33 Vertical conversion unit 34 Vertical selector unit 35a to 35h 0th integration unit to 7th integration unit 36 Vertical detection unit 58a to 58h 0th horizontal buffer to 7th horizontal buffer 59a to 59h 0th vertical buffer to 7th vertical buffer

Claims (10)

An image compression method for performing image compression by preparing a plurality of difference storage means respectively corresponding to each pixel position in a pixel block that is a processing unit of image compression,
Calculating an absolute value of a difference between signals of each of a plurality of pixels forming an image between adjacent pixels;
Sorting the absolute value of the calculated difference for each pixel position in the pixel block based on the pixel position related to the difference;
Integrating and storing the absolute value of the sorted difference in the difference storage unit corresponding to the pixel position related to the difference;
A step of comparing absolute values of accumulated differences stored in the difference storage means;
Detecting the absolute value of the largest accumulated difference by the comparison;
A step difference storing means for storing the absolute value of the largest accumulated the difference detected is based on the pixel position in the pixel block corresponding to identify where to start compression processing against the image,
An image compression method comprising: compressing an image in units of pixel blocks from a specified location. Further comprising the step of increasing or decreasing in accordance with the absolute value of the calculated difference to the absolute value of said difference, the step of fractionation claim 1 Symbol placement of image separate the absolute value of the difference step of the increase or decrease is increased or decreased Compression method. A computing means for computing an absolute value of a difference between signals of each of a plurality of pixels forming an image between adjacent pixels;
Classification means for classifying the absolute value of the difference calculated by the calculation means into each pixel position in a pixel block which is a processing unit of image compression based on the pixel position related to the difference;
A plurality of difference storage means corresponding to each pixel position in the pixel block, for storage for accumulating and storing the absolute values of the differences sorted by the sorting means;
Comparison means for comparing absolute values of accumulated differences stored in the difference storage means;
Detecting means for detecting an absolute value of the maximum accumulated difference by comparison of the comparing means;
Point specifying the difference storing means for detection means for storing the absolute value of the largest accumulated the difference detected based on the pixel position in the pixel block corresponding to identify where to start compression processing against the image Means,
An image compression apparatus comprising: an image compression unit that compresses an image in units of pixel blocks from a location specified by the location specifying unit. A computing means for computing an absolute value of a difference between signals of each of a plurality of pixels forming an image between adjacent pixels;
A adjusting unit that said calculating means is increased or decreased according to the absolute value of said difference absolute value of the difference computed,
Based absolute value of the difference bulking down means increases or decreases to the pixel position according to said difference, and sorting means for sorting to each pixel position in the pixel block is a processing unit of an image compression,
A plurality of difference storage means corresponding to each pixel position in the pixel block , for storage for accumulating and storing the absolute values of the differences sorted by the sorting means;
Comparing means for comparing absolute values of accumulated differences stored in the difference storing means;
Detecting means for detecting an absolute value of the maximum accumulated difference by comparison of the comparing means;
A part specifying unit that the difference storing means for detection means for storing the absolute value of the largest accumulated the difference detected based on the pixel position in the pixel block corresponding to identify where to start the compression process,
An image compression apparatus comprising: an image compression unit that compresses an image in units of pixel blocks from a location specified by the location specifying unit. The image compression apparatus according to claim 3 or 4 , wherein the signal is a luminance signal and / or a color difference signal related to an image. An imaging unit that outputs the captured image;
Wherein an image compression apparatus according to any one of claims 3 to 5,
The imaging device is characterized in that the calculation means calculates the image output by the imaging unit. In a processing location identification method for preparing a plurality of difference storage means respectively corresponding to each pixel position in a pixel block which is a processing unit of image compression, and identifying a location where image compression processing is started ,
A step of calculating an absolute value of a difference between signals of each of a plurality of pixels forming an image between adjacent pixels;
Sorting the absolute value of the calculated difference into each pixel position in the pixel block based on the pixel position related to the difference;
Integrating and storing the absolute value of the sorted difference in the difference storage unit corresponding to the pixel position related to the difference;
A step of comparing absolute values of accumulated differences stored in the difference storage means;
Detecting the absolute value of the largest accumulated difference by the comparison;
Further comprising the steps of difference storing means for storing the absolute value of the largest accumulated the difference detected is based on the pixel position in the pixel block corresponding to identify where to start compression processing against the image Characterized processing location identification method. A computing means for computing an absolute value of a difference between signals of each of a plurality of pixels forming an image between adjacent pixels;
A adjusting unit that said calculating means is increased or decreased according to the absolute value of said difference absolute value of the difference computed,
Based absolute value of the difference bulking down means increases or decreases to the pixel position according to said difference, and sorting means for sorting to each pixel position in the pixel block is a processing unit of an image compression,
A plurality of difference storage means corresponding to each pixel position in the pixel block , for storage for accumulating and storing the absolute values of the differences classified by the classification means;
Comparing means for comparing absolute values of accumulated differences stored in the difference storing means;
Detecting means for detecting an absolute value of the maximum accumulated difference by comparison of the comparing means;
Largest accumulated a difference of place specifying means for specifying a position to start the compression process the absolute value difference storing means for storing pairs in the image based on the pixel position in the pixel block corresponding to the detection means detects And a processing location specifying device. In a computer program for causing a computer to perform image compression,
Computer
Arithmetic hand stage in which a plurality of pixels forming an image is computed between pixels adjacent the absolute value of the difference signal, each of which has,
Classification means for classifying the absolute value of the difference calculated by the calculation means into each pixel position in a pixel block which is a unit of image compression based on the pixel position related to the difference;
A plurality of difference storage means that respectively correspond to each pixel position in the pixel block and accumulate and store the absolute values of the differences sorted by the sorting means;
Comparison means for comparing absolute values of accumulated differences stored in the difference storage means,
Detecting means for detecting an absolute value of the maximum accumulated difference by comparison of the comparing means;
Largest accumulated the basis of the pixel position in the pixel block difference memory means for storing the absolute value of the difference is corresponding, part specifying means to identify where to start the compression process on the image detection means detects ,
A computer program for causing an image compression unit to perform image compression on a pixel block basis from a location specified by the location specifying unit . In a computer program for causing a computer to specify a location for image processing,
Computer
Arithmetic hand stage in which a plurality of pixels forming an image is computed between pixels adjacent the absolute value of the difference signal, each of which has,
Classification means for classifying the absolute value of the difference calculated by the calculation means into each pixel position in a pixel block which is a processing unit of image compression based on the pixel position related to the difference;
A plurality of difference storage means that respectively correspond to the pixel positions in the pixel block and accumulate and store absolute values of the differences sorted by the sorting means;
Comparative hand stage said difference storage means for comparing the absolute value between the accumulated the difference store,
Detecting means for detecting an absolute value of the maximum accumulated difference by comparison of the comparing means;
Point specifying the difference storing means for detection means for storing the absolute value of the largest accumulated the difference detected based on the pixel position in the pixel block corresponding to identify where to start compression processing against the image A computer program for functioning as a means.