CN100552712C - Image processing apparatus, image processing method, and distortion correcting method - Google Patents

Image processing apparatus, image processing method, and distortion correcting method Download PDF

Info

Publication number
CN100552712C
CN100552712C CN 200480021876 CN200480021876A CN100552712C CN 100552712 C CN100552712 C CN 100552712C CN 200480021876 CN200480021876 CN 200480021876 CN 200480021876 A CN200480021876 A CN 200480021876A CN 100552712 C CN100552712 C CN 100552712C
Authority
CN
China
Prior art keywords
distortion correction
image
range
distortion
data
Prior art date
Application number
CN 200480021876
Other languages
Chinese (zh)
Other versions
CN1830002A (en
Inventor
上野晃
古川英明
日暮正树
Original Assignee
奥林巴斯株式会社
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority to JP202664/2003 priority Critical
Priority to JP202493/2003 priority
Priority to JP202663/2003 priority
Priority to JP2003202493A priority patent/JP4772281B2/en
Application filed by 奥林巴斯株式会社 filed Critical 奥林巴斯株式会社
Publication of CN1830002A publication Critical patent/CN1830002A/en
Application granted granted Critical
Publication of CN100552712C publication Critical patent/CN100552712C/en

Links

Abstract

一种具有失真校正处理部(8)的图像处理装置,该图像处理装置具有失真校正范围计算部(12),其用于计算使前述失真校正处理部(8)进行失真校正处理所需要的输入图像范围。 The image processing apparatus having a distortion correction processing section (8), the image processing apparatus includes a distortion correction range calculation unit (12) for calculating the distortion correction processing section so that (8) input to the distortion correction processing required image range. 这样,可针对图像输出范围不多不少地输出通过失真校正所获得的校正图像(输出图像),可进行有效利用作为源数据的输入(摄影)图像数据的失真校正处理,可针对枕形失真、桶形失真以及曲弦失真等进行有效的失真校正处理。 In this way, neither more nor less output corrected image (output image) obtained by the distortion correction for the image output range, can be effectively utilized as the input source data distortion correction process (photographic) image data, for pincushion distortion can be , barrel distortion and song chord distortion effective distortion correction processing.

Description

图像处理装置、图像处理方法以及失真校正方法技术领域 An image processing apparatus, an image processing method, and a distortion correction TECHNICAL FIELD

本发明主要涉及供数字照相机等的电子摄像装置使用的图像处理装置和图像处理方法,具体涉及不使电路规模和数据传送量增大,即可实现失真校正功能,并且在进行小区域(例如块行)单位中的失真校正处理的情况下,可算出小区域的空间位置的图像处理装置、图像处理方法以及失真校正方法。 The present invention relates to an image processing apparatus and an image processing method for an electronic image pickup apparatus using a digital camera or the like, more particularly to the circuit scale and without increasing the data transfer amount, the distortion correction function can be realized, and performing a small region (e.g. block row) in the case where the distortion correction processing unit, the image processing device of the spatial position of the small region, an image processing method, and a method of distortion correction can be calculated.

背景技术 Background technique

不管是数字照相机,还是银盐照相机, 一般在照相机的光学系统中会出现畸变像差。 Whether it is a digital camera, a camera or a silver salt, typically in the optical system of the camera distortion aberration occurs. 并且,在目前出售的照相机中,大部分是可进行光学变焦的机型,在该情况下,从广角端到望远端,畸变像差的状态发生变化。 Further, in the camera currently available, most optical zoom models, in this case, from the wide angle end to the telephoto end, the distortion of the state changes. 在广角端桶形失真多,在望远端枕形失真多。 Barrel distortion at the wide angle end and more and more pincushion distortion at the telephoto end.

例如对格子状的被摄物体进行拍摄时,畸变像差被观测为桶形失真或枕形失真。 For example, when the subject is photographed lattice shape, distortion aberration is observed as barrel distortion or pincushion distortion.

图56是格子状的被摄物体,图57是发生了桶形失真的摄影图像, 图58是发生了枕形失真的摄影图像。 FIG 56 is a lattice-shaped object to be photographed, the photographic image of FIG. 57 is a barrel distortion has occurred, FIG 58 is a photographic image pincushion distortion occurs.

另外,在数字照相机中,对CCD等摄像元件的数据进行各种图像处理,之后把通过JPEG等压縮方式所压縮的数据记录在存储卡等的记录介质内。 Further, in the digital camera, the imaging data element such as a CCD performs various image processing, and the like after the compression by the JPEG compressed data is recorded in a recording medium such as a memory card.

图59示出在一般数字照相机中执行的图像处理步骤的概念。 FIG 59 shows a concept of the image processing steps performed in a general digital camera. 在预处理中对由CCD所取入的摄像信号进行像素缺陷处理、A/D转换等,把所获得的图像数据临时存储在SDRAM等帧存储器内。 CCD image pickup signal generated by the defective pixel taken in the pre-processing, A / D conversion, temporarily stores the image data obtained in the frame memory such as an SDRAM. 然后,通过图像处理对从帧存储器中读出的图像数据进行各种图像处理,进而通过JPEG处理进行图像压縮并记录在作为记录介质的存储卡等内。 Then, performs various image processing on image data read out from the frame memory by the image processing, and further performs image compression processing by JPEG and recorded in the memory card as a recording medium and the like.

图60示出现有数字照相机的图像处理装置的方框图(例如特开 A block diagram (FIG. 60, for example, Laid-Open image processing means has a digital camera illustrating the

42000-312327号公报中的现有例)。 Publication No. 42000-312327 conventional in embodiment).

在图60中,现有的图像处理装置使CPU104、预处理电路102、多个图像处理电路106-l〜106-n、 JPEG处理部107、帧存储器105以及作为记录介质的存储卡108与总线103连接。 In FIG. 60, the conventional image processing apparatus that the CPU 104, pre-processing circuit 102, a plurality of image processing circuits 106-l~106-n, JPEG processing unit 107, a frame memory 105 as a recording medium and the memory card 108 and bus 103. 然后,通过CPU 104的控制, 将来自CCD 101的摄像信号在预处理电路102中实施像素缺陷处理和A/D转换等,之后通过总线103临时存储在帧存储器105内。 Then, by the control of the CPU 104, the image pickup signal from the CCD 101 of the embodiment of pixel defects process and A / D conversion in the pre-processing circuit 102, after passing through the bus 103 is temporarily stored in the frame memory 105. 然后,从帧存储器105中读出图像数据,通过总线103输入到图像处理电路106-1 中来进行规定的图像处理,并再次通过总线103重新写入到帧存储器105 内。 Then, from the frame memory 105 is read out image data to predetermined image processing to the input image processing circuit 106-1 through the bus 103, and bus 103 again by re-writing the frame memory 105. 以下同样,通过总线103在帧存储器105和图像处理电路106-2〜106-n 之间依次进行数据交换,最后在JPEG处理部107中进行JPEG压縮处理, 把进行了图像处理的数据临时存储在帧存储器105内,把从帧存储器105 中读出的处理数据记录在存储卡等108内。 Hereinafter the same, via a bus 103 between the frame memory 105 and the image processing circuit exchange data 106-2~106-n sequentially, and finally performs JPEG compression processing in the JPEG processing unit 107, temporarily stores the image data has been processed in the frame memory 105, read out from the frame memory 105 to process data recorded in the memory card 108. 在以上各级的图像处理和JPEG处理中,以小区域(块行)为单位进行图像处理。 In the above image processing and JPEG processing levels in a small area (block line) of the image processing units.

另外,作为在图像处理电路的至少一个中进行图56〜图58所述的失真校正处理的方法,通过使用下述光学表达的失真校正式[式5]对摄像图像的畸变像差进行校正。 Further, as a method of distortion correction processing according to FIG. 56~ 58 in FIG least one image processing circuit, formal [Formula 5] The distortion aberration is corrected captured image by using the following Expression optical distortion correction. <formula>formula see original document page 5</formula> <Formula> formula see original document page 5 </ formula>

校正图像内的坐标摄影图像内的坐标失真中心位置失真校正系数 Distortion coordinates within a coordinate position of the center in the captured image correcting image distortion correction coefficient

然而,在使用光学表达的上述[式5]进行失真校正的情况下,校正图像产生无用的部分或者不足的部分。 However, in the case where the optical expression of [Formula 5] a distortion correction, a corrected image generating section unnecessary portion or insufficient.

图61和图62对使用上述[式5]来校正发生了桶形失真和枕形失真的摄影图像的动作进行说明。 FIGS. 61 and 62 using the above-mentioned [Formula 5] action barrel distortion and pincushion distortion correcting photographed image occurs will be described.

图61是使用[式5]对虚线所示的发生了桶形失真的源图像(图示左)进行了校正的情况,所校正的图像(图示右)从图像输出范围溢出而产生无用的部分。 FIG 61 using [5] of the source image (shown left) barrel distortion as shown in broken lines occur in the case where the corrected image (shown right) from the corrected image output range overflow wasteful section.

图62是使用[式5]对虚线所示的发生了枕形失真的源图像(图示左)进行了校正的情况,所校正的图像(图示右)比图像输出范围小而产生数据不足的部分。 FIG 62 is a [Formula 5] of the broken line in the source image (shown left) pincushion distortion in the case where correction is performed as shown in the image (right in the figure) of the corrected image output range less than the insufficient data is generated part.

对于这种图61和图62的不利之处,在作为先行申请的日本国特开平9-098340号公报中,在失真校正后,通过进行电子变焦来缩小因校正而扩大的部分,使其返回到原始位置,把因失真校正而丢失的图像信息抑制到最小限度。 For such disadvantages 61 and 62, as in the preceding Japanese Patent Application Laid-Open Publication No. 9-098340, after the distortion correction to reduce the correction-enlarged portion by the electronic zoom, to return to the original position due to the distortion correction image information lost minimized.

另一方面,通常,CCD等摄像元件的读出可具有以下2种方式:(1) 本拍摄用的所有像素读出,(2)透视图像用的疏化读出。 On the other hand, generally, the read image pickup element such as a CCD may have the following two ways: (1) all pixels of this captured with reading, (2) a perspective thinning readout image used. 由于通过疏化读出等从所拍摄的数据中取入了一部分的图像数据的长宽比不是l丄因而不能照原样进行准确的失真校正。 Since the thus accurate distortion correction can not be performed as it is read out by thinning data and the like from the captured image data taken into the aspect ratio is not part of Shang l.

并且,以往,公幵了进行作为该图像处理一部分的失真校正的先行技术。 Further, conventionally well-Jian preceding the distortion correction processing techniques as part of the image. 例如,具有特开平6-181530号公报和特开平10-224695号公报。 For example, Laid-Open No. 6-181530 has Publication and Laid-Open Publication No. 10-224695.

在特开平6-181530号公报中作了以下记载:在检测单元检测出拍摄时的摄像变焦镜头的摄像位置在畸变像差大的位置内的情况下,通过根据几何变形读出固体摄像元件的摄像数据,对由该摄像变焦镜头所产生的图像的几何失真迸行校正。 The following were described in JP-A-6-181530 in: detecting the imaging position of the zoom lens of the imaging time of shooting in the case where the detection unit within a large distortion aberration position, by reading the solid-state imaging device according to the geomorphing image data of the imaging geometry by the image distortion produced by the zoom lens into line correction.

在特开平10-224695号公报中作了以下记载:把由接收经由光学系统的光的固体摄像装置所拍摄的数据存储在随机存取视频存储器内,随机读出定时发生电路具有用于校正由光学系统产生的像差的像差校正数据,裉据该像差校正数据,按照规定顺序读出由固体摄像装置所生成的信号来生成视频信号,从而校正光学系统的畸变像差。 The following were described in JP 10-224695 in which: in the video random access memory, a random read from the solid-state imaging device receiving light via an optical system, the captured data is stored having a timing generating circuit for correcting a the aberration correction data generated by the optical system aberration, the aberration correction data according to Ken, a solid-state imaging device in accordance with the signals generated sequentially read out to generate a video signal, thereby correcting the distortion aberration of the optical system.

然而,在特开2000-312327号公报中,在帧存储器和各图像处理电路之间数据交换多,总线的数据传送量增大。 However, in Laid-Open Patent Publication No. 2000-312327, the data between the frame memory and the respective plurality of the image processing circuit switched data transfer load on the bus increases. 而且,对小区域(块行) 单位中的失真校正处理完全没有记载。 Further, the distortion correction processing for the small area (block line) in units of no description.

并且,在特开平9-098340号公报中,仅提及了桶形失真的校正,以距摄像中心最近的点为基准,在原理上只是桶形失真的校正。 Also, in JP-A-9-098340, the only mention of barrel distortion correction to the nearest point from the imaging center as a benchmark, in principle only barrel distortion correction.

6对于枕形失真,所校正的图像比图像输出范围小而产生图像数据不足的部分。 6 for the pincushion distortion, the corrected image output range is smaller than the image generated by the image data portion insufficient. 而且,为了校正图22所示的曲弦失真,失真校正式需要4次项以上,然而很难分析求出用于最大限度使用源图像数据的校正倍率M。 Further, in order curved string 22 shown in FIG distortion correction, distortion correction formula requires four or more terms, however, is difficult to analyze the correction of magnification for obtaining maximum use of the source image data M.

在特开平6-181530号公报中描述了进行失真校正处理的内容,然而如图36所示,对于CCD的摄像面上的像素(由黑圆点表示的纵横格子点)31,本来应当是直线的被摄物体像由于光学镜头的畸变像差而弯曲成像(由符号32表示)时,畸变像差校正电路内的行存储器需要大容量缓冲存储器。 In JP-A-6-181530 describes the contents of the distortion correction process, however, as shown in FIG. 36, for the pixel (vertical and horizontal lattice points represented by black circles) of the imaging surface of the CCD 31, a straight line should have been the subject image distortion due to aberration of the optical lens to bend imaged (represented by the symbol 32), the line memory circuit in the distortion aberration correction requires a large memory capacity of the buffer. 并且,可使用缓冲存储器的容量进行处理的图像尺寸受到限制。 Further, the available capacity of the buffer memory size of the image processing is limited. 而且,没有关于缓冲存储器内所保持的数据的空间位置的确定方法等的记载。 Moreover, there is no description of a method to determine the spatial position of the data within the buffer memory remains. 而且,在特开平6-181530号公报中,对小区域(块行)单位中的失真校正处理完全没有记载。 Further, in JP 6-181530, the distortion correction processing for the small area (block line) in units of no description.

另外,在特开10-224695号公报中,进行随机存取处理不需要缓冲存储器,然而一般对SDRAM等存储器进行随机存取时,需要转送时间。 Further, in Laid-Open Patent Publication No. 10-224695, the process does not require random access buffer memory, however, when the SDRAM is generally a random access memory, etc., required transfer time.

因此,本发明的第l目的是提供有效利用所拍摄的源数据进行失真校正处理,可针对图像输出范围没有浪费且有效地输出所校正的图像, 可进行能应对枕形失真、桶形失真以及曲弦失真的失真校正处理的图像处理装置和图像处理方法。 Accordingly, a first object of the present invention, l is the source data to provide efficient use of captured distortion correction process, can be efficiently without waste and outputs the corrected image for the image output range, you may be able to cope with pincushion distortion, barrel distortion and the image processing apparatus and an image processing method chord song distortion distortion correction processing.

本发明的第2目的是提供可算出小区域(例如块行)的空间位置, 并可实现小区域单位中的失真校正处理的图像处理装置和失真校正方法。 The second object of the present invention to provide a small area can be calculated (e.g. block line) spatial position, and to achieve an image processing apparatus and method for distortion correction in units of small areas of the distortion correction processing.

本发明的第3目的是提供可对进行了疏化取入的图像数据执行准确的失真校正,另一方面,也能对通过数字变焦等切取了任意区域的数据进行准确的失真校正的图像处理装置和失真校正方法。 A third object of the present invention is to provide an accurate distortion correction is performed performs a thinning of the image data taken, on the other hand, can be accurate distortion correction processing on the image data arbitrary taken by digital zooming area cut distortion correction means and methods.

本发明的第4目的是提供不使总线的传送量和存储器的容量大幅增大,即可实现失真校正处理的图像处理装置和图像处理方法。 A fourth object of the present invention is to provide not the transfer amount and the memory capacity of the bus is increased significantly, the image processing apparatus and an image processing method of the distortion correction processing can be realized.

发明内容 SUMMARY

本发明的图像处理装置是具有失真校正单元的图像处理装置,该图 The image processing apparatus according to the present invention is an image processing apparatus includes a distortion correction unit, FIG.

像处理装置还具有:失真校正范围计算部,其计算由前述失真校正单元 The image processing apparatus further includes: a distortion correction range calculation unit that calculates the distortion correcting unit by the

7进行失真校正处理的输入图像范围,前述失真校正范围计算部具有:坐标生成部,其生成内插坐标;失真校正坐标变换部,其输出把规定的失真校正式应用于前述所生成的内插坐标而变换的坐标;以及校正范围检测部,其根据前述所变换的坐标位置计算前述输入图像范围。 7 performs distortion correction processing input image range, the distortion correction range calculation unit comprises: coordinate generation unit, which generates the interpolation coordinates; a distortion correcting coordinate-transformation unit, outputs the correction formula is applied a predetermined distortion in the generated interpolated converted coordinates; and a correction range detecting unit, which calculates the range of the input image according to the coordinates of the transformed position.

这样,通过设置用于计算进行失真校正处理的输入图像范围的失真校正范围计算部,计算与输出图像(即失真校正后的图像)的输出范围的一部分或全部相当的区域在输入图像(即摄像图像)内占据的范围, 从而可针对应输出的输出范围不多也不少地输出由失真校正所获得的校正图像,不会使校正图像相对于输出范围溢出或不足。 Range of the input image this way, by providing the distortion correction process for calculating the distortion correction range calculation unit, a part or all of the corresponding region of the output range of the output image is calculated (i.e., distortion-corrected image) in the input image (i.e., imaging occupy within the image) in the range, so as to be no less outputs corrected image obtained by the distortion correction for the output range to be output, without the correction with respect to the image output range is insufficient or an overflow.

在该结构中,生成内插坐标是指,为了求出失真校正前的摄像图像(输入图像)的位置,事先生成应输出的校正后的校正图像(输出图像) 的位置。 In this configuration, the interpolation coordinates generation means, to determine the position of the captured image (input image) before the distortion correction, the corrected image to be generated in advance of the corrected output (output image) position. 把失真校正式应用于该内插坐标,生成(变换)并输出失真校正前的图像位置(坐标),根据所变换的校正前的坐标位置计算应进行失真校正处理的输入图像范围。 The distortion correction formula is applied to the interpolation coordinates, generate (conversion) and outputs the image before distortion correction position (coordinates) calculated from the coordinate position before correction of the conversion range of the input image should be distortion correction processing.

或者,优选的是,前述失真校正范围计算部针对通过坐标变换所生成的坐标,根据与前述输出图像范围的4边的各边对应的像素的坐标的最大值和最小值以及与前述输出图像范围的4顶点对应的坐标中的至少任意一项计算前述输入图像范围。 Alternatively, it is preferable that the distortion correction range calculation unit for coordinate generated by the coordinate transformation according to maximum and minimum coordinates of the respective side edges 4 of the output image area and the corresponding pixels of the output image range at least any one of the input image calculated range corresponding to the coordinates of the four vertexes. 在该结构中,4顶点是指失真校正后的输出图像中的各边的顶点。 In this configuration, the apex 4 refers to the apex of the edge of the output image after the distortion correction.

或者,优选的是,前述失真校正范围计算部针对进行失真校正处理的多个输入信号依次重复范围计算来计算前述输入图像范围。 Alternatively, it is preferable that the distortion correction range calculation unit for a plurality of input signals are sequentially repeated distortion correction range calculation process calculates the range of the input image.

在该结构中,在使用多个信道的情况下,在对多个输入信号同时进行失真校正处理期间,针对多个输入信号,分别对作为下一处理对象的块行依次重复进行失真校正范围计算,从而可使用1个失真校正范围计算部进行多个信道的范围计算。 In this configuration, in the case where a plurality of channels, a plurality of input signals during a distortion correction process simultaneously for a plurality of input signals, respectively, as the next processing target block line sequentially repeated distortion correction range calculation , enabling the use of a distortion correction range calculation unit for calculating a range of the plurality of channels.

优选的是,通过重复处理进行前述范围计算,决定使失真校正后的图像范围相对于输入图像范围在规定范围内的校正倍率M。 Preferably, the calculation processing is repeated by the aforementioned range, the range of the determined image distortion correction range with respect to the input image magnification is corrected within a predetermined range M.

根据该结构,通过把校正倍率M决定为合适值,也能应对曲弦失真。 According to this configuration, by correcting the magnification M is determined as an appropriate value, it is possible to deal with distortions chord song.

优选的是,前述失真校正范围计算部在前述失真校正单元执行失真校正中,计算进行下一失真校正的图像输入范围。 Preferably, the distortion correction unit performs the distortion correction range calculation in the distortion correction unit calculates the distortion-corrected image for the next input range.

在该结构中,由于在失真校正执行中计算进行下一失真校正的图像输入范围,因而当1个块行的失真校正结束时,可知进行下一失真校正处理的图像输入范围,可针对下一块行依次顺利进行失真校正处理。 In this configuration, since the next input image distortion correction range calculation performed in the distortion correction, so that when one block line distortion correction ends, an image input range of the next known distortion correction processing for the next block may be followed by the smooth line distortion correction processing.

本发明的图像处理方法是进行失真校正处理的图像处理方法,在进行前述失真校正处理时,计算进行失真校正处理的输入图像范围,其特征在于,包括以下步骤:生成内插坐标;输出把规定的失真校正式应用于所生成的内插坐标而变换的坐标;以及根据所变换的坐标位置计算前述输入图像范围。 The image processing method of the present invention is the distortion correction processing of an image processing method, when performing the distortion correction processing, the distortion correction processing calculation range input image, characterized by comprising the steps of: generating the interpolation coordinates; the predetermined output distortion correction formula applies coordinate transformation and interpolating the generated coordinate; and calculates the coordinate range of the input image transformed position.

这样,在进行失真校正处理时,通过计算进行失真校正处理的输入图像范围,计算与失真校正后的图像输出范围的一部分或全部相当的区域在输入图像内占据的范围,从而可针对图像输出范围不多也不少地输出由失真校正所获得的校正图像,不会使校正图像相对于图像输出范围溢出或不足。 Thus, during the distortion correction processing by the distortion correction processing calculates the input range image, a portion or all of the region corresponding to the image output range occupied by calculating the distortion correction in the image input range, output range so that the image may be directed to no less outputs corrected image obtained by the distortion correction, the correction is not an image relative to the image output range is insufficient or an overflow.

附图说明 BRIEF DESCRIPTION

图1是示出本发明的第1实施方式的图像处理装置的整体结构的方框图。 FIG. 1 is a block diagram of the overall configuration of an image processing apparatus according to a first embodiment of the present invention.

图2是示出失真校正处理部的结构的方框图。 FIG 2 is a block diagram showing a configuration of the distortion correction processing section.

图3〜图5是失真校正处理部中的坐标变换的概念图,图3是示出摄像图像数据的图,图4是示出校正图像的图,图5是对内插处理进行说明的图。 FIG 3 ~ FIG. 5 is a conceptual diagram of distortion correction coordinate conversion processing section, FIG. 3 is a diagram showing an image of image data, FIG. 4 is a diagram showing a corrected image, FIG. 5 is a diagram illustrating the interpolation process of .

图6是示出图像数据的读出顺序的图。 FIG 6 is a sequence of the read image data shown in FIG.

图7是对第1数据顺序变换部中的数据顺序变换处理进行说明的图。 FIG 7 is a diagram illustrating the data sequence of the first sequence of data conversion processing in the conversion unit. 图8是对第2数据顺序变换部中的数据顺序变换处理进行说明的图。 FIG 8 is a diagram for explaining the second data sequence order of the data conversion in the conversion processing unit. 图9是示出块行和失真校正处理所需要的存储容量(缓冲容量)的关系的图。 FIG 9 is a block diagram illustrating the relationship between the line and the distortion correction processing required storage capacity (buffer capacity) of FIG.

图IO是示出块行宽度的设定方法的图。 FIG IO is a diagram illustrating a method of setting the line width of the block of FIG.

图11是示出失真校正范围计算部的结构的方框图。 FIG 11 is a block diagram showing a configuration of a distortion correction range calculation portion. 图12和图13是说明失真校正范围计算部中的计算进行失真校正处理时的输入图像范围的动作的图,图12是示出校正图像上的范围的图, 图n是示出源数据上的范围的图。 12 and FIG. 13 is calculated distortion correction range calculation section view illustrating an operation of the input image range when the distortion correction processing, FIG. 12 is a diagram showing a range in the corrected image, FIG n is shown on the source data FIG range.

图14是对失真校正处理动作进行说明的流程图。 FIG 14 is a flowchart of the operation of the distortion correction processing is described. 图15是对在图14的步骤Sll中的校正倍率M的计算方法进行说明的流程图。 FIG 15 is a flowchart illustrating the method of calculating the correction step Sll magnification M in FIG. 14 is performed.

图16是示出图11的失真校正范围计算部的结构的方框图。 FIG 16 is a block diagram showing a configuration of a distortion correction range calculation unit 11 of FIG. 图17是图16的数据输出的时序图。 FIG 17 is a timing chart of data output 16 of FIG.

图18对图17的动作进行了改善,是流水线中的数据输出的时序图。 18 FIG. 17 is the operation of the improvement, is a timing chart of data output from the pipeline. 图19是说明计算进行3个信道的失真校正处理时的输入图像范围的方法的图。 FIG 19 is a diagram showing a method of calculation of the input image range when the three channels of the distortion correction processing.

图20是对3个信道时的失真校正处理动作进行说明的流程图。 20 is a flowchart of the distortion correction processing operation when the three channels illustrated. 图21是示出本发明的第2实施方式的图像处理装置的整体结构的方框图。 FIG 21 is a block diagram showing an overall configuration of an image processing apparatus according to a second embodiment of the present invention.

图22是示出发生了曲弦失真的摄影图像的图。 FIG 22 is a diagram showing chord song photographic image distortion occurs.

图23和图24是示出校正前的摄像图像数据和校正后的校正图像的关系的图,图23是示出校正前的摄像图像数据的图,图24是示出校正后的校正图像的图。 23 and FIG. 24 shows a relationship between a corrected image after shows captured image data and the corrected before correction, FIG. 23 is a view showing the captured image data before the correction, FIG. 24 is a corrected image of the shown correction Fig.

图25是图23和图24的失真校正处理的流程图。 FIG 25 is a flowchart of FIG. 24 and 23 the distortion correction process in FIG.

图26和图27是示出在数字变焦的情况下,校正前的摄像图像数据和校正后的校正图像的关系的图,图26是示出校正前的摄像图像数据的图,图27是示出校正后的校正图像的图。 26 and FIG. 27 is a diagram illustrating the case of a digital zoom, showing the relationship between the captured image data before the correction and the corrected image after the correction, FIG. 26 is a view showing the captured image data before the correction, FIG. 27 is a diagram showing corrected image after correction.

图28和图29是示出在从摄像数据中切取某区域的情况下,校正前的摄像图像数据和校正后的校正图像的关系的图,图28是示出校正前的摄像图像数据的图,图29是示出校正后的校正图像的图。 28 and FIG. 29 is a diagram illustrating the case where the cut out a region from the image data, showing a relationship between a corrected image of the captured image data and the corrected before correction, FIG. 28 is a view showing the captured image data before the correction in FIG. FIG 29 is a diagram showing a corrected image corrected FIG.

图30〜图32是对CCD的监视模式中的行疏化进行说明的图,图30 是对将摄像时的数据在纵方向进行3行疏化并从CCD中取入到存储器内的状态进行说明的图,图31是示出被取入到存储器内的状态的摄影图像数据的图,图32是示出对图像纵横非对称进行了校正的校正图像的图。 FIG 30~ FIG. 32 is a CCD line monitoring mode diagram illustrating thinning, FIG. 30 is a data line 3 of the imaging will be thinned in the longitudinal direction from the CCD and taken into the state in the memory DESCRIPTION oF FIG, 31 is a diagram illustrating a state is taken into the memory of the photographed image data, and FIG. 32 is a diagram illustrating an image aspect asymmetric correction was corrected image of FIG.

10图33和图34是对YC422形式的数据中的颜色数据的横方向的疏化进行说明的图,图33是示出亮度数据的图,图34是示出颜色数据的图。 10 and FIG. 34 FIG. 33 is a cross-direction thinning color data of the data in the form of YC422 diagram illustrating, FIG. 33 is a diagram showing luminance data, FIG 34 is a diagram illustrating the color data of FIG. 图35是对图像的中心偏移进行说明的图。 FIG 35 is a center of the image offset illustrated in FIG. 图36是示出CCD的像素和摄影图像的关系的图。 FIG 36 is a diagram illustrating the relationship between the captured image and the pixels of the CCD is shown. 图37是示出本发明的第3实施方式的图像处理装置中的失真校正处理部的详细结构的方框图。 FIG 37 is a block diagram showing the configuration of a distortion correction processing section of the image processing apparatus shown in the third embodiment of the present invention. FIG.

图38是内插电路中的内插运算的表象图。 38 is interpolated in the interpolation operation circuit representation of FIG. 图39是对失真校正处理部中的内部存储部进行说明的图。 FIG 39 is an internal storage unit in the distortion correction processing unit will be described in FIG. 图40和图41对图39进行补充说明,是对在2端口SRAM内写入数据的状态进行说明的图,图40是示出写入顺序的图,图41是示出图40所示的写入顺序的数据被写入到16个2端口SRAM上的何处的图。 FIGS. 40 and 41 in FIG. 39 supplement, is a state of writing the data in the two-port SRAM diagram illustrating, FIG. 40 is a diagram illustrating the writing order, and FIG. 41 is a diagram illustrating shown in FIG. 40 sequential write data is written to FIG 16 where the two-port SRAM. 图42是用于求出计算图38的校正坐标位置所需要的Do的说明图。 FIG 42 is a diagram for explaining the correction obtained Do is calculated coordinate position of the 38 required in FIG. 图43是示出错误处理例的图。 FIG 43 is a diagram showing an example of error handling. 图44是示出另一错误处理例的图。 FIG 44 is a diagram showing another example of the error processing. 图45是示出另一错误处理例的图。 FIG 45 is a diagram showing another example of the error processing. 图46是对失真校正处理所需要的缓冲量进行说明的图。 FIG 46 is a buffer amount required for processing the distortion correction will be described in FIG. 图47是对失真校正处理所需要的缓冲量进行说明的图。 FIG 47 is a buffer amount required for processing the distortion correction will be described in FIG. 图48和图49是对失真校正处理所需要的缓冲量进行说明的图,图48是对preULB和postULB的定义进行说明的图,图49是对跨过失真中心时的缓冲器开放量进行说明的图。 48 and FIG. 49 is a buffer amount required for the distortion correction processing described in FIG. 48 is a definition preULB and postULB will be described in FIG, 49 is open when the amount of the buffer will be described across the center of the distortion Fig.

图50是对伴随UL处理的开放量计算方法进行说明的图。 FIG 50 is attendant UL opening amount calculating process will be described in FIG.

图51是对伴随UL处理的开放量计算方法进行说明的图。 FIG 51 is a process that accompanies the opening amount of the UL calculation method described in FIG.

图52是对伴随图51的处理的流水线处理的空闲进行说明的图。 FIG 52 is a view of the pipeline processing that accompanies the idle process 51 is illustrated in FIG.

图53是对伴随UL处理的开放量计算方法进行说明的图。 FIG 53 is attendant UL opening amount calculating process performed. FIG.

图54是对伴随图53的处理的流水线处理的空闲进行说明的图。 FIG 54 is a view of the pipeline processing that accompanies the idle process 53 is illustrated in FIG.

图55是对伴随UL处理的开放量计算方法进行说明的图。 FIG 55 is attendant UL opening amount calculating process will be described in FIG.

图56是示出格子状的被摄物体的图。 FIG 56 is a diagram showing a lattice-like subject FIG.

图57是示出发生了桶形失真的摄影图像的图。 FIG 57 is a diagram illustrating the barrel distortion of the captured image has occurred.

图58是示出发生了枕形失真的摄影图像的图。 FIG 58 is a diagram showing a captured image pincushion distortion occurs. 图59是示出一般数字照相机的图像处理步骤的概念的图。 FIG 59 is a diagram illustrating a concept of an image processing step shows a general digital camera.

图60是示出现有数字照相机的图像处理装置中的方框结构的图。 FIG 60 is a diagram showing a block configuration of a digital camera image processing apparatus shown in occurs.

图61是对使用[式5]校正发生了桶形失真的摄影图像的动作进行 FIG 61 is a use of [5] of the captured image motion barrel distortion correction occurs for

说明的图。 Illustrated in FIG.

图62是对使用[式5]校正发生了枕形失真的摄影图像的动作进行说明的图。 FIG 62 is the use of [Formula 5] operation of the photographic image pincushion distortion correction diagram explaining occurred.

具体实施方式 Detailed ways

参照附图对发明的实施方式进行说明。 Referring to the drawings, embodiments of the invention will be described.

图1是示出本发明的第1实施方式的图像处理装置的整体结构的方框图。 FIG. 1 is a block diagram of the overall configuration of an image processing apparatus according to a first embodiment of the present invention. -

在图1的图像处理装置中,通过进行与总线3连接的各部的控制的CPU 4的控制,在预处理电路2中对来自CCD 1的摄像信号进行像素缺陷处理和A/D转换等,把所获得的图像数据通过总线3临时存储在帧存储器5内。 In the image processing apparatus of FIG. 1, the respective units controlled by the control bus with the CPU 4 in the connector 3, the pixel defects of the imaging signals from the CCD 1 in the pre-processing circuit 2 in the processing and A / D conversion, the the image data obtained by the three bus temporarily stored in the frame memory 5. 帧存储器5由SDRAM等构成,是存储图像处理前的数据和图像处理后的数据的存储器。 The frame memory 5 is composed of an SDRAM, a memory data stored before and after the image data processing and image processing. 然后,把从帧存储器5中读出的图像数据通过总线3输入到第1数据顺序变换部6中。 Then, the read out from the frame memory 5, the image data input to the first data order conversion unit 6 via the bus 3. 第1数据顺序变换部6如后面图7说明的那样,具有多个、这里为2个可存储块单位的数据的存储器。 As the first data sequence converting unit 6 as described later in FIG. 7, having a plurality of, here, two memory data may be stored in units of blocks. 第1数据顺序变换部6从帧存储器5中按行方向读出并存储数据后,再按列方向依次读出数据,并输出到图像处理电路7中。 After the first data converting unit 6 sequentially read out in the row direction and stores data from the frame memory 5, then the column direction are sequentially read out and output to the image processing circuit 7.

图像处理电路7对所输入的数据进行规定的图像处理,并传送到作为下级的失真校正单元的失真校正处理部8中。 The image data processing circuit 7 is inputted predetermined image processing, and transmitted to the distortion correction unit as lower distortion correction processing unit 8. 失真校正处理部8对所输入的数据进行失真校正处理,并传送到下级的第2数据顺序变换部9 中。 Distortion correction processing on the input data of 8 unit performs the distortion correction process, and transmitted to the lower order of the second data converting unit 9. 第2数据顺序变换部9如后面图8说明的那样,具有多个、这里为2 个可存储块单位的数据的存储器。 Second data sequence converting section 9 explained later in FIG. 8, having a plurality of, here, two memory data may be stored in units of blocks. 第2数据顺序变换部9从失真校正处理部8中按列方向读出数据而进行存储后,再按行方向依次读出数据, 并传送到JPEG处理部10中。 After the second data sequence converting unit 9 from the distortion correction processing section 8 stores the read data in the column direction, then the row direction are sequentially read out, and transferred to the JPEG processing unit 10. 然后,在JPEG处理部10中进行JPEG压縮处理,把处理数据临时存储在帧存储器5内,把从帧存储器5中读出的处理数据记录在存储卡等11内。 Then, the JPEG compression processing in the JPEG processing unit 10, the processed data is temporarily stored in the frame memory 5, read out from the frame memory 5 for processing data recorded in the memory card 11.

12失真校正处理部8如图2所示,构成为具有:内插坐标生成部81, 其生成失真校正后的校正图像的位置(称为内插位置,X, Y)、和与其对应的失真校正前的源图像的位置(X,, Y');作为缓冲存储器(以下 8 distortion correction processing section 12 shown in Figure 2, is configured to have: an interpolation coordinate generation unit 81 that generates the position correction of the distortion-corrected image (referred to as an interpolation position, X, Y), and the corresponding distortion the position of the source image before the correction (X ,, Y '); as a buffer memory (hereinafter,

简称缓冲器)的内部存储部82,其临时存储来自失真校正处理部8的前级块的电路的图像数据的一部分;存储器控制部83,其进行针对该内部存储部82的写入和读出控制;以及内插运算部84,其根据所变换的失真校正前的图像位置坐标(X', Y')进行图像处理,以进行数据的失真校正。 Internal storage unit referred to as buffer) 82 temporarily stores the distortion correction from the image data processing unit of a circuit portion of the previous stage block 8; the memory control unit 83, which was 82 for writing and reading out the internal storage unit control; the distortion correction and interpolation operation unit 84 that performs image processing in accordance with the position of the image before distortion correction coordinates of the transformed (X ', Y'), for the data.

内插坐标生成部81如图2所示,由以下部分构成:内插位置生成部811,其生成内插坐标(X, Y);失真校正坐标变换部812,其输出把规定的失真校正式[式1](后述)应用于所生成的内插坐标(X, Y)而变换的校正前的坐标(X', Y');以及选择器813,其可有选择地输出来自内插位置生成部811的内插坐标(X, Y)和来自失真校正坐标变换部812 的变换坐标(X', Y')。 The interpolation coordinates generating unit 81 shown in Figure 2, consists of the following parts: an interpolation position generation unit 811, which generates interpolated coordinates (X, Y); a distortion correction coordinate conversion unit 812 which outputs the predetermined distortion correction formula [formula 1] (described later) applied to the generated interpolation coordinate (X, Y) coordinates before the correction conversion of (X ', Y'); and a selector 813, which can be selectively output from the interpolator the interpolation section 811 generates a coordinate position (X, Y) and the coordinate transformation from the distortion correction coordinate conversion unit 812 (X ', Y'). 失真校正处理部8中的内插位置生成部811、失真校正坐标变换部812、选择器813以及存储器控制部83根据存储有控制数据的控制寄存器85内所设定的针对各块的设定值来进行动作。 Interpolating distortion correction processing section 811 in position generation section 8, the distortion correction coordinate conversion unit 812, a selector 813 and a memory control unit 83 set values ​​for each block in the control register 85 in accordance with the set data is stored to carry out the action. 并且, 可从CPU参照处理结果的状态等。 Then, the state of the processing result and the like can be referred to by the CPU.

在失真校正处理部8内,如图1所示,设置有失真校正范围计算部12,其计算由失真校正处理部8进行失真校正处理的输入图像范围。 In the distortion correction processing unit 8, shown in Figure 1, it is provided with a distortion correction range calculation unit 12, which calculates the range of the input image distortion correction process by the distortion correction processing unit 8. 失真校正范围计算部12如图11说明的那样,构成为具有:坐标生成部91, 其生成内插坐标;失真校正坐标变换部92,其输出把规定的失真校正式应用于所生成的内插坐标而变换的坐标;以及校正范围检测部93,其根据所变换的校正前的坐标位置计算前述输入图像范围。 11 as described distortion correction range calculation unit 12 in FIG configured having: coordinate generation unit 91, which generates the interpolation coordinates; a distortion correcting coordinate-transformation unit 92, outputs the correction formula is applied a predetermined distortion in the generated interpolation converted coordinates; and a correction range detection unit 93, which calculates the coordinate position before correction of the conversion range of the input image.

在以上构成的本发明的图像处理装置中,从上述第1数据顺序变换部6到JPEG处理部10,不通过总线3,而通过与该总线3不同的信息传递路径进行连接,使得可进行流水线处理,将图像数据按二维像素排列中的规定块单位进行传送和处理。 In the image processing apparatus according to the present invention the above configuration, from the first data sequence converting unit 6 to the JPEG processing unit 10, without passing through the bus 3, and by connecting the three different information transmission path of the bus, such that a pipeline processing the image data transfer and processing two-dimensional pixel arrangement according to a predetermined block unit. 这样通过总线3的数据传送仅为以下传送,即:从帧存储器5到第1数据顺序变换部6的传送,从JPEG处理部10到帧存储器5的传送,以及从帧存储器5到存储卡11的传送,因此,与在帧存储器和各图像处理电路之间进行数据交换的现有例(图60) 相比,可大幅降低通过总线3的数据传送量,可大幅减轻总线3的负荷。 Thus the following is only transmitted via the data transfer bus 3, namely: a transmission frame from the memory 5 to the first data order conversion unit 6, is transmitted from the frame memory 10 to the JPEG processing unit 5, and from the frame memory 5 to the memory card 11 conventional embodiment (FIG. 60) as compared to the transmission, and therefore, the data exchanged between the respective frame memory and the image processing circuit, can be greatly reduced by the amount of transferred data bus 3, can significantly reduce the load bus 3. 并且,在该图1所示的例中,进行图像处理的图像处理电路7仅设置1 The image processing circuit and, in the embodiment shown in FIG. 1, the image processing is provided only 7 1

个,然而可以是多个。 One, however, may be more. 而且,在图中,在图像处理电路7后设置有失真校正处理部8,然而可以是相反的结构。 Further, in the figure, after the image processing circuit 7 is provided with the distortion correction processing unit 8, however, the structure may be reversed.

在由初级的图像处理电路7和作为第2级的图像处理电路的失真校正处理部8构成的图像处理电路部中,构成为:在各图像处理电路7、 8 的前级或内部配置有作为流水线寄存器的未作图示的小容量存储器,各图像处理电路7、 8通过该小存储器进行流水线处理动作。 By the image processing circuit of the primary image processing circuit 7, and a second stage of the image processing circuit of the distortion correction processing section 8 constituted, configured to: in the image processing circuit 7, the first stage or internal 8 is arranged as (not shown) of the small memory pipeline register, each of the image processing circuit 7, 8 pipeline processing operation by the small memory. 这些小容量存储器是为了进行如下动作而设置的,即:在各图像处理电路7、 8中进行空间图像处理的情况下,存储图像处理所需要的周边数据,以及必须将图像数据按块单位读出并进行重新排列等,来进行处理。 These small-capacity memory for the following operation provided, namely: in the case, the 8 spatial image processing of each image processing circuit 7, the periphery of data storage of the image processing required, and the need to read image data in units of blocks and rearranging the like, for processing.

图3〜图5是失真校正处理部8中的坐标变换的概念图。 FIG 3 ~ FIG. 5 is a conceptual diagram of distortion correction coordinate conversion processing unit 8. 图3示出作为源数据的摄像图像数据,图4示出校正图像,图5示出针对图4的校正图像的坐标位置(X, Y),变换到图3的源数据的坐标上的坐标位置P处的数据的坐标(X,, Y,)(该坐标P位于不一定与实际构成源数据的多个像素的位置准确一致的坐标位置)。 Coordinates on the graph 3 shows the captured image data as source data, Figure 4 shows a corrected image, Figure 5 shows for the coordinate position corrected image of FIG. 4 (X, Y), converted to the source data in Figure 3 of the coordinate data at the position P (X ,, Y,) (P is the coordinate position of the plurality of pixels constituting the actual data are not necessarily accurate and consistent source coordinate position). 该P点处的坐标位置(X', Y,)使用该P点的周边像素16点的坐标来计算,并且P点处的图像数据使用其周围的前述16点的像素数据进行内插运算。 A coordinate position at the point P (X ', Y,) using the surrounding pixels of the point P 16 of the coordinates of a point is calculated, and the image data at the point P using the pixel data of surrounding the point 16 interpolation operation is performed. 在内插运算部84中对点P位置的数据进行处理,以便根据其周围16点的像素值(亮度数据) 进行内插运算。 Including point P of the data processed in the interpolation calculation unit 84, an interpolation operation for pixel value (luminance data) of around 16:00. 在图2的内插位置生成部811中生成内插坐标是指,在图4的校正图像侧指示哪个像素位置(X, Y)。 Interpolation position generating unit in FIG. 2, the interpolation coordinate generation means 811, indicating which pixel positions (X, Y) in the corrected image of FIG. 4 side.

使用[式l]可算出对应于失真校正后的像素位置(X, Y)的失真校正前的像素位置(X', Y')。 Use [Formula L] can be calculated corresponding to the pixel position (X ', Y') before the pixel position (X, Y) of the distortion-corrected distortion correction. 然而,如前所述,失真校正前的像素位置(X,, Y')不限于与源图像数据上的像素位置对应的整数值。 However, as described above, the pixel position (X ,, Y ') before distortion correction is not limited to integer values ​​corresponding to the pixel positions on the source image data.

根据[式l],把图4的坐标(X, Y)如图3所示变换成坐标(X,, Y,)。 The [Formula l], the coordinates (X, Y) as shown in FIG 4 is converted into coordinates (X ,, Y,) 3. 这样,可知道可以生成源数据中的哪个坐标位置的数据,该位置(X,, Y')的数据可使用进行16点内插处理的内插式,根据周围16点的己知像素值(图5的黑圆点的输入图像数据)来计算。 Thus, data may be generated which may be known coordinate positions of the source data, the position (X ,, Y ') data may be interpolated using the interpolation processing 16:00, known pixel values ​​around 16:00 of ( black circles in the input image data of FIG. 5) is calculated.

14[式l] 14 [Formula L]

<formula>formula see original document page 15</formula>校正图像内的坐标摄影图像内的坐标失真中心位置中心偏移校正范围校正倍率抽样比 <Formula> formula see original document page 15 </ formula> distortion coordinate position of the center coordinates of the center of the captured image within the corrected image offset correction sampling range correction magnification ratio

失真校正系数 Distortion correction coefficient

[式1]中的Z是从失真中心(Xd, Yd)到现在所关注的点(X, Y) 的距离。 [Formula 1] Z is the distance of the point (X, Y) is now of interest from the distortion center (Xd, Yd). 使用[式1]计算相对于校正图像的点(X, Y)发生失真的源图像的坐标(X,, Y')。 Using [Formula 1] calculated with respect to the point (X, Y) coordinates of the corrected image of the distorted source image (X ,, Y '). 对[式1]进行补充说明,M是用于在使用光学系统的数据进行逻辑校正时,对校正后图像溢出或不足的现象进行校正的校正倍率。 Of [Formula 1] supplementary explanation, when M is for using data of the optical system is logically correct, insufficient or an overflow phenomenon corrected image for correcting magnification is corrected. Sx, Sy是用于对疏化取入等、纵横空间抽样间隔不同的现象进行校正的抽样比。 Sx, Sy are taken for thinning the like, spaced vertical and horizontal spatial sampling different phenomena corrected decimation ratio. X。 X. ff, Y。 ff, Y. ff是对由于失真校正处理而在失真校正处理后使被摄物体位置偏移拍摄时的位置的现象进行校正的中心偏移值。 ff is the central offset value due to the distortion correction processing is shifted when the position of the imaging subject position after the distortion correction processing for correcting the phenomenon.

在本发明的[式1]中,通过考虑高次项(具体地说,是Z4、 Z6、…), 还可应对更复杂的曲弦失真(参照图22)。 In the present invention, [formula 1] by considering the higher order terms (specifically, Z4, Z6, ...), may be more complicated to deal with a curved chord distortion (see FIG. 22). 并且,在虽然中心一致但只有摄像元件的一部分取入数据的情况下,或者在光轴中心和摄像元件中心发生偏移的情况等下,必须考虑失真中心(Xd, Yd)。 And the like in the case, although the center line in the case where only a part of the imaging device but the data is taken, offset occurs in the optical axis or center of the imaging element and the center, the center must be considered distortion (Xd, Yd). 而且,为了对由于失真校正处理而使被摄物体位置偏移拍摄时的位置的现象迸行校正,必须要考虑中心偏移校正值(X。ff, Y。ff)。 Further, since the distortion correction process in order to subject the position misalignment phenomenon of an object photographed into line correction, the center must consider the offset correction value (X.ff, Y.ff). 在疏化取入的情况下,或者在对由亮度数据(Y)和颜色数据(Cb, C》构成的YC图像进行处理的情况 In the case of thinning fetched, or in the case of the data processed by the luminance (Y) and color data (Cb, C "YC image including

下,空间抽样根据X、 Y而不同。 , The spatial sampling according to X, Y are different. 因此,导入抽样比(Sx, Sy)作为系数, Thus, introducing the sampling ratio (Sx, Sy) as a coefficient,

对坐标变换进行校正。 Coordinate transformation is corrected.

在图61的桶形失真中,为了进行失真校正,必须设定M〈1来进行少许縮小。 Barrel distortion in FIG. 61, in order distortion correction, is set to be M <1 be slightly reduced. 相反,在图62的枕形失真的情况下,必须进行放大,所以设定M〉1。 In contrast, in the case of FIG. 62 pincushion distortion must be amplified, so the set M> 1.

在现有例所述的[式5]中,考虑校正桶形失真/枕形失真来进行描述。 In the conventional embodiment described in [5], in view of correcting barrel distortion / pincushion be described.

下面,参照图6〜图8对图像数据的写入和读出的处理和顺序进行说明。 Next, referring to FIG. 6 ~ FIG. 8 and the writing process and the sequentially read image data will be described.

图6是对从本实施方式中的帧存储器中读出图像数据的顺序进行说明的图。 FIG 6 is sequentially read out from the image data of the present embodiment will be described in the frame memory of FIG.

图像数据通常在行方向进行扫描来写入,并在读出时按行方向读出, 全部读出1行的图像数据,然后全部读出相邻行的图像数据,通常重复进行这种操作。 The image data is typically written to the row scanning direction, and read out in the row direction at the time of reading, read all the image data of one line, and reads out all the image data of the adjacent rows, this operation is usually repeated.

相比之下,本发明的图像处理装置针对在行方向进行扫描而写入的图像数据,以在列方向的某恒定长度为单位,把行方向的图像数据依次输入到图像处理部7中,然后,把相邻列依次输入到图像处理部7中, 重复输入到图像的右端,把这样获得的小区域(矩形状的图像数据)称为块行(BL)。 Contrast image data, the image processing apparatus according to the present invention for scanning in the row direction are written to a constant length as a unit in the column direction, the row direction of the image data of sequentially input into the image processing section 7, then, the adjacent columns are sequentially input to the image processing section 7, a repeated input to the right image, the small area (rectangular image data) thus obtained is referred to as block lines (BL).

参照图7和图8对用于可这样读出图像数据的第1数据顺序变换部6和第2数据顺序变换部9的结构进行说明。 7 and 8, the structure used for the first data sequence converting unit 6 and the second data sequence converting unit can thus read image data will be described with reference to FIG 9. 图7和图8是示出第1和第2数据顺序变换部的结构的方框图。 7 and FIG. 8 is a block diagram showing a configuration of the first order conversion unit and the second data.

第1数据顺序变换部6如图7所示,具有多个、这里为2个可存储块单位的图像数据的存储器,该2个存储器6a和存储器6b可由写入侧和读出侧的各开关交替切换写入和读出。 The first data sequence converting unit 6 shown in Figure 7, having a plurality of, here, the image memory 2 can store data in units of blocks, the two memories 6a and 6b may be a memory write side and the read side of the respective switches alternately switching writing and reading. 即,帧存储器5通过写入侧的开关可切换地与该存储器6a和存储器6b连接,并且图像处理部7也通过读出侧的开关可切换地与这些存储器6a和存储器6b连接,当帧存储 That is, the frame memory 5 can be switchably connected through the write switch-side memory with the memory 6a and 6b, and the image processing section 7 also reads out side switch switchably connected to these memories 6a and 6b through the memory, when the frame memory

16器5与存储器6a和存储器6b中的一方连接时,该存储器6a和存储器6b 中的另一方可切换成与图像处理部7连接。 16 5 is connected to one of the memories 6a and 6b in a memory, the memory 6a and 6b, the memory can be switched to the other one connected to the image processing section 7. 即,存储器6a、 6b被切换成不与帧存储器5和图像处理部7的双方同时连接,从而可交替地进行写入和读出。 That is, the memory 6a, 6b is switched simultaneously with the parties not connected to the frame memory 5 and the image processing section 7, thereby alternately write and read.

存储在帧存储器5内的帧图像的一部分以块单位按行方向读出,并被存储在一个存储器、这里例如存储器6a内。 Portion of the frame image stored in the frame memory 5 in units of blocks in the row direction is read out, and stored in a memory, for example, where the memory 6a.

与此同时,从存储器6b中将已从帧存储器5中读出并存储的块单位的图像数据按列方向(纵方向)依次读出,并输出到图像处理部7中。 At the same time, the image data from the frame memory 5 in units of blocks is read out from and stored in the memory 6b in the column direction (vertical direction) is sequentially read out, and output to the image processing unit 7.

当从帧存储器5向存储器6a的写入、以及从存储器6b向图像处理部7的读出结束时,写入侧的开关和读出侧的开关被切换,然后,开始从帧存储器5向存储部6b写入下一块单位的图像数据,并开始从存储器6a向图像处理部7读出块单位的图像数据。 When from the frame memory 5 is written to the memory 6a, 6b and the end of the memory to the image processing unit 7 reads out from the writing-side switches and the read switch is switched side, and then, starts from the frame memory 5 to the memory portion 6b write image data in a unit, and starts the image data from the memory block unit 6a to the read image processing section 7.

第2数据顺序变换部9如图8所示,也与上述第1,数据顺序变换部6大致相同地构成,并且进行大致相同的动作。 Second data converting section 9 sequentially 8, also in the first, data order conversion unit 6 is substantially the same configuration, and performs substantially the same operation.

艮口,第2数据顺序变换部9构成为具有:存储器9a,存储器9b,写 Burgundy port, the second data sequence converting unit 9 is configured to include: a memory 9a, the memory 9b, writing

入侧开关,以及读出侧开关。 The switch side, and the read-side switch.

并且,在该第2数据顺序变换部9动作时,从失真校正处理部8向存储器9a和存储器9b中的一方进行列方向(纵方向)的写入,从存储器9a和存储器9b中的另一方进行行方向(横方向)的读出,并输出到JPEG处理部10中。 Then, when the second data sequence converting operation unit 9, 8 write column direction (longitudinal direction) to one of the memory 9a and 9b in the memory from the distortion correction processing unit, a memory from other memory 9a and 9b of be read in the row direction (lateral direction) out, and output to the JPEG processing unit 10.

图9示出块行和失真校正处理所需要的存储容量(缓冲容量)的关系。 Figure 9 shows a storage capacity (buffer capacity) and the distortion correction block line processing required relationship. 虚线框上发生失真的4条实线在失真校正后的输出数据中是应当成为直线(纵l行)的数据,但在源数据中却发生失真。 4 The solid line distortion that occurs in the dashed box on the distortion-corrected output data is the data should be a straight line (vertical line l), but the distortion has occurred in the source data. 4条直线中的远离图像中心的左端直线失真最大。 4 the left end of the straight line away from the center of the image of the maximum linear distortion. 黑圆点表示与失真校正后的输出数据的像素位置对应的失真校正前的位置。 Position before the distortion correction corresponding to the pixel positions of the black circles indicates the distortion correction and output data. 在前述的内部存储部82中,把在摄像后发生失真的输入数据在块行内具有最大失真量的横方向的宽度上考虑16点内插而在左右具有余量的范围确保为失真校正处理所需要的缓冲量。 Considering the width in the internal storage unit 82, the input data distortion occurs in the image having a transverse direction of the maximum amount of distortion in the block line interpolation 16:00 and has a range of balance about ensuring the distortion correction process desired amount of cushioning. 即,表示只要有这种缓冲容量,就能进行失真校正的缓冲量,换句话说,当进行了失真校正时,可生成正确的直线形状的缓冲量。 That is, as long as this represents a buffer capacity, the buffer amount of distortion correction can be carried out, in other words, when the distortion correction, the proper amount of cushioning linear shape may be generated. 图io示出块行宽度的设定方法。 FIG io illustrates a method of setting the line width of the block. 示出根据失真校正处理的对象位置 Object position shows a distortion correction processing according to

对块行宽度进行可变设定的示例。 Block line width is variably set of examples. 弯曲的虚线表示输入侧的失真数据, 越是远离图像中心,即越是靠向外侧,失真程度就越大。 The curved dashed line shows the input side of the distortion data, farther from the center of the image, i.e. more toward the outside, the greater the degree of distortion. 因此,当针对帧存储器上的输入数据设定块行宽度时,该宽度被设定成,越是远离图像中心就越大,越是接近中心宽度就越窄。 Thus, when the line width is set for an input data block on the frame memory, the width is set to be, the greater the more away from the center of the image, the closer the center of the narrower width. 这样,当从帧存储器5通过总线3向图像处理电路进行数据传送时,只要数据宽度可以变窄,就能减少此时的总线专有时间。 Thus, when the data transfer to the image processing circuit 5 from the frame memory 3 through a bus, the data width can be narrowed as long as, at this time can be reduced proprietary bus time.

一般,在图像中心由失真引起的变形量小。 In general, in the center of the image deformation caused by the distortion amount is small. 根据失真校正处理对象的位置,变更块行宽度(在图示的纵方向进行扫描的宽度)的设定值。 Distortion correction processing according to the position of the object, change the block row (width in the vertical scanning direction shown) of the set value. 由于式1包含有高次项,因而不能分析求出块行部分的处理所需要的输入范围。 Since the Formula 1 contains higher order terms, and therefore can not process the input range of the analysis block obtains the required line portions. 块行宽度的设定是根据失真校正范围计算部12的处理结果,并通过由CPU对该结果实施规定运算来设定的。 Block line width is set according to the processing result of the distortion correction range calculation unit 12, and the CPU performs predetermined calculation by the result set.

图11示出失真校正范围计算部12的结构。 FIG 11 shows a structure of the distortion correction range calculation unit 12. 失真校正范围计算部12 Distortion correction range calculation unit 12

构成为具有:坐标生成部91,其生成内插坐标(X, Y);失真校正坐标 Configured having: coordinate generation unit 91, the interpolation coordinates (X, Y) within the generating; distortion correction coordinates

变换部92,其输出把规定的失真校正式(例如[式l])应用于所生成的内插坐标(X, Y)而变换的坐标(X', Y');以及校正范围检测部93, 其根据所变换的校正前的坐标位置(X', Y,)计算失真校正处理所需要的输入图像范围。 Conversion unit 92 which outputs the predetermined distortion correction formula (e.g. [Formula L]) applied to the generated interpolation coordinate (X, Y) and the coordinate transformation (X ', Y'); and a correction range detection unit 93 , which calculates the coordinate position before the distortion correction by the transformation (X ', Y,) required for the correction processing range of the input image. 这样,失真校正范围计算部12由坐标生成部91、失真校正坐标变换部92以及校正范围检测部93构成,由CPU 4通过控制寄存器94控制动作,并且通过寄存器95来取得范围计算结果。 Thus, the distortion correction range calculation unit 12 by the coordinate generating section 91, the distortion correction coordinate conversion section 92 and a correction range detection unit 93 constituted by a CPU 4 calculates the results using the control register 94 controls the operation of, and acquires the range of 95 through the register. 失真校正范围计算部12是针对失真校正处理功能而附加的,作为考虑失真变形可算出图像数据的输入范围的支持功能起作用。 Distortion correction range calculation unit 12 for processing the distortion correction function in addition, as with the distortion deformation can be calculated from the input range image data support function.

下面参照图12和图13,对失真校正范围计算部12中的计算失真校正处理所需要的输入图像范围的动作进行说明。 Referring now to FIGS. 12 and 13, the operation of the distortion correction processing of the input image ranges required 12 will be described in the distortion correction range calculation unit calculates.

采用在校正图像内指定范围,并在输入图像上进行坐标变换的方法来进行。 The method of using the corrected image within the specified range, and coordinate transformation is performed on the input image. 首先,如图12所示,决定校正图像上的规定范围(例如5行) 的像素位置(格子点位置),使用[式1]如图13所示在输出侧的源数据上进行坐标变换,从而知道输出侧数据上的5行的数据范围。 First pixel position, as shown in FIG 12, a predetermined range determined on the corrected image (e.g. 5 rows) (grid positions), using the coordinate transformation as shown on the output side of the source data [Formula 1] 13, 5 so as to know the scope of data lines on the output side of the data. 处理设定通过硬件的寄存器设定来进行。 Setting process performed by the hardware register settings. 在图12的校正图像上,例如起点被设定 In the corrected image of FIG. 12, the starting point is set e.g.

18成,从点(Xst, Yst)开始按间隔AW和AH的顺序生成格子点,在纵向取 18 into the order starting AW and AH interval grid points from the point of generation (Xst, Yst), taken in the longitudinal direction

几点,在横向取几点,在纵向成为H。 Points, points taken in the transverse direction, in the longitudinal direction becomes H. ut像素,在横向成为W。 ut pixels in the transverse direction becomes W. ut像素。 ut pixels. However

后,变换成源数据上的位置的结果是,在发生失真的线上形成与校正图像的格子点对应的交点坐标,然而可知与该块行对应的发生失真的范围, After converted into a position on the source data result, a corrected image of the lattice points corresponding to the intersection coordinates on a line distortion that occurs, however, we found that distortion occurs corresponding to the range block row,

即外接的图示粗虚线为输入图像范围。 I.e., the external input image shown thick broken line range. 实际上,如果只跟踪图12和图13 的黑圆点所示的矩形状的周围4边上的黑圆位置的坐标,则对于规定该位置范围的4边,根据最大值和最小值便可知道输入图像范围。 In fact, if the coordinates of the rectangular track only around the black circle shown in FIGS. 12 and 13 positions of the black circles edge 4, for the four sides of the predetermined range of positions, according to the maximum and minimum values ​​can be You know the input image range. 例如, E.g,

实际检测上边收敛在YTm狀,YT幽之间、下边收敛在YBmax, YBmJn之间、 左边收敛在Xf, Xi^之间、右边收敛在XR^x, X^n之间的数据,并存储在结果存储用寄存器95内。 The actual detection of convergence between the upper YTm like, quiet YT, lower convergence between YBmax, YBmJn, left convergence between Xf, Xi ^, the right data convergence between XR ^ x, X ^ n and stored in storing the result register 95. 块行处理需要把内插所需要的像素与 Block line processing requires the interpolation pixel and the desired

(XLmin〜XRmax, YTmin〜YBmax)相加后的范围。 (XLmin~XRmax, YTmin~YBmax) the added range. 另外,在图13中,在输出侧的源数据上,把各顶点的变换后的位置设定为(X, TL, Y, TL)、 (X, Further, in FIG. 13, on the output side of the source data, the position of each vertex after the conversion is set to (X, TL, Y, TL), (X,

tr, Y, tr)、 (X, bl, Y, bl)、 (X' 肌,Y, br)。 tr, Y, tr), (X, bl, Y, bl), (X 'muscles, Y, br).

坐标生成部91使用[式2]生成校正图像所需要的坐标位置(X, Y), 使用[式l]变换成(X,, Y,)。 Coordinate generation unit 91 using a coordinate position (X, Y) [Formula 2] required to generate the corrected image, using [Formula L] is converted to (X ,, Y,). 把上边的变形范围设定为YT函〜YT画, 把下边的变形范围设定为YBmin〜YBmax,把左边的变形范围设定为XLmin〜 X^ax,把右边的变形范围设定为Xr^〜X^w把各顶点的变换后的位置设定为(X, TL, Y, TL)、 (X, TR, Y' TR)、 (X, BL, Y, BL)、 (X, BR, Y, BR),在校正范围检测部93中对这些变形范围和位置进行检测,并存储在结果存储用寄存器95内。 The deformation of the top of the range is set to YT letter ~YT painting, the deformation range below the set YBmin~YBmax, the deformation range on the left is set to XLmin~ X ^ ax, the deformation range on the right is set to Xr ^ ~X ^ w to the position of each vertex after the conversion is set to (X, TL, Y, TL), (X, TR, Y 'TR), (X, BL, Y, BL), (X, BR, Y, BR), detects such variations in the correction range and position range detection unit 93, and stored in the register 95 storing the result. [式2] [Formula 2]

、=XST +k '厶W , = XST + k 'W Si

k和《是整数 and k "is an integer

图14是对失真校正处理动作进行说明的流程图。 FIG 14 is a flowchart of the operation of the distortion correction processing is described. 首先,在步骤Sll 中,决定失真校正[式l]的校正倍率M。 First, in step Sll, the decided distortion correction [Formula L] correction magnification M. 校正倍率M的决定方法在后面使用图15的流程图进行描述。 The method of correcting the magnification M is determined to be described later using a flowchart in FIG. 15. 然后,在步骤S12中,使用所决定的校正倍率M,如图10〜图13的说明所述,考虑内插所需要的部分来计算块行处理所需要的输入范围,在步骤S13中设定该计算的输入范围和失 Then, in step S12, the correction using the determined magnification M, as shown in the description of FIG. 10~ 13, the inner portion need be considered to calculate the input lines necessary for processing the range block is inserted, is set in step S13 the calculated input range and loss

19真校正处理所需要的设定值。 19 True desired setting correction processing. 然后,在步骤S14中,同时执行各块行的失真校正处理和下一块行的输入范围计算。 Then, in step S14, is calculated while performing distortion correction processing for each block line and a lower line of the input range. 之后,执行各块行的失真校正处理和下一块行的输入范围计算,直到进行图像的最下段的块行的失真校正处理为止。 Thereafter, the distortion correction process performed in each block line and the input range of one line calculated, block distortion until the lowermost line of the image correction processing. 在步骤S15中判定针对输出图像整体的失真校正处理是否结束,从而重复步骤S14,并继续进行,直到对图像整体的失真校正处理结束为止,然后结束该处理。 In step S15, it is determined for the entire output image distortion correction process is ended, thus repeating the step S14, and continues until the end of the distortion correction processing until the entire image, and then ends the process.

这样,由于在失真校正执行中计算进行下一失真校正的图像输入范围,因而当1个块行的失真校正结束时,可以知道进行下一失真校正处理的图像输入范围,针对下一块行可以没有延迟且依次顺利地进行失真校正处理。 Thus, since the next input image distortion correction range calculation performed in the distortion correction, so that when one end of the block line distortion correction, distortion of the image can be known for the next input level correction processing, for the next one line can not and sequentially delay distortion correction process smoothly.

在上述步骤Sll中,当适当决定校正倍率M时,无论对图61和图62所述的桶形和枕形的哪种失真图像进行校正处理,都能恰好进入图像输出范围。 In the step Sll, when correcting the magnification M is appropriately determined, regardless of the distorted image of FIG. 61 and FIG. 62 of the barrel and pincushion correction processing of which, can be entered just the image output range.

图15示出对图14的步骤S11的校正倍率M的计算方法进行说明的流程图。 15 shows a flowchart of a method of correcting the magnification M is calculated in step S11 of FIG. 14 will be described. 首先,在步骤S21中设定lO作为校正倍率M的初始值。 First, as an initial value set lO M is the magnification correction in step S21. 然后, 使用该M和失真校正[式l]对输出图像的例如4边进行坐标变换,从而在失真校正范围计算部12中计算输入图像范围(步骤S22),判定是否在源图像范围内(步骤S23)。 Then, using the M and the distortion correction [Formula L] of the four sides, for example, an output image coordinate transformation, so that the input image range (step S22) in the distortion correction range calculation unit 12 calculates, determines whether the source range (step S23). 当在步骤S23中如图61所示那样超出输出范围时,使M减少AM (步骤S24),之后返回到步骤S22,计算输入图像范围,再次进行步骤S23的范围内判定。 When in step S23 shown in Figure 61 as the output beyond the range of the AM M reduction (step S24,), returns to step S22 after the calculation of the input range image, the range of the determination of step S23 again. 即,逐渐改变M值,在步骤S23中判定是否在源图像范围内,当在源图像范围内时,转移到步骤S25。 That is, the value of M is gradually changed, it is determined whether a range within the source image, the source image when within range, proceeds to step S25, in step S23. 在步骤S25中判断前述的输入图像范围是否是收敛在源图像范围内的最大M。 Determining whether the aforementioned image input in step S25 falls within the range of the source image is the maximum range M. 当在源图像范围内不是最大时,使M增加AM (步骤S26),之后返回到步骤S22,执行步骤S23〜26。 When not within the maximum range of the source image, M is incremented so that AM (step S26), returns to step S22 after the step S23~26. 当在步骤S25中,在源图像范围内是最大时,把此时的M决定为校正倍率。 When the result in step S25, the source image within the maximum range, the M at this time is determined as the correction of magnification. 此时,在范围计算中使用的区域是校正图像整体。 In this case, the area used in calculating the range of the entire image is corrected.

另外,图15的流程可以采用其他方法。 Further, the flow of FIG. 15 may be employed other methods. 例如,当使用将校正倍率M For example, when the correction magnification M

的最大值设为Mn^、最小值设为lVUn,并对该范围依次进行2分割来逼 The maximum value of Mn ^, minimum value of lVUn, and sequentially dividing the range of 2 to force

近最佳M的方法时,可以减少计算量。 Near optimum when M is, the amount of calculation can be reduced.

20另外,在图14的步骤S14中,与失真校正处理部8的校正运算相比, 失真校正范围计算部12的范围计算运算的运算量少。 Further operation amount 20, as compared in step S14, the correction operation of step 14 of FIG distortion correction processing unit 8, the distortion correction range calculation unit 12 to calculate the operation range. 因此,通过按时序使用校正运算的乘法器,可削减电路规模。 Thus, by using the correction calculation by the multiplier sequence, the circuit scale can be reduced.

当进行失真校正处理时,观察[式1]可知,为乘法非常多、而且在硬件方面电路规模非常大的计算。 When the distortion correction process, observe [Formula 1] and found to be very much multiplication, and a very large circuit scale in computing hardware. 一般情况下,乘法器使硬件的电路规模增大,因而优选的是尽可能减少乘法器。 In general, the hardware multiplier so that the circuit scale is increased, it is preferable that the multiplier to reduce as much as possible. 设置乘法器控制部进行乘法器控制,在取定时的同时,通过进行流水线处理,不降低整体的处理速度,可縮小电路规模。 The multiplier is provided for the control unit controls the multiplier, while taking timing, by pipeline processing, without reducing the overall processing speed, circuit scale can be reduced.

图16示出图11的失真校正坐标变换部92的部分的具体结构。 FIG 16 shows a specific configuration of the portion 11 of the distortion correcting coordinate-converting unit of FIG. 92. 除此 In addition

以外的部分与图11相同。 Than the same portions in FIG. 11. 在失真校正坐标变换部92中构成为:尽管进 In constituting the distortion correction coordinate conversion unit 92: intake although

行[式1]那样的乘法多的计算,但通过在实现定时的同时,按时序对此时的多个乘法进行处理,可减少乘法器数。 Multi-row multiplication calculating [Formula 1] as, but at the same time achieved by the timing of a plurality of time series of multiplication processing in this case, the number of multipliers can be reduced.

失真校正坐标变换部92构成为具有:ZH十算部921,其输入来自坐标生成部91的内插坐标(X, Y),运算[式1]的(2)来计算Z、乘法器控制部922,其输入来自ZH十算部921的Zz和来自乘法器923的y,输出a和P;乘法器923,其输入oc和(3,输出"乘法器控制部924,其输出与校正系数A、 B、 C、…中的任意一方相当的5;乘法器925,其输入oc 和5,输出与校正系数和Z的整数倍的积相当的s;以及校正坐标计算部926,其输入来自乘法器925的s和来自坐标生成部91的内插坐标(X, Y),输出校正前的源数据的坐标(X', Y')。 The distortion correcting coordinate-transformation unit 92 is configured to include: ZH ten calculation unit 921, the input from the interpolation coordinates (X, Y) coordinates generating unit 91 calculates [Formula 1] (2) to calculate Z, the control unit multiplier 922, which is input from the ten Operators ZH Zz portion 921 and y, a and P output from the multiplier 923; multipliers 923, input oc and (3 outputs "multiplier control unit 924 outputs the correction coefficient a either, B, C, ... of the corresponding 5; multiplier 925, the input oc and 5, corresponding to an integral multiple of the product of the output of the correction coefficient s and Z; and a correction coordinate calculation unit 926 that inputs from the multiplication 925 s and coordinates generating unit 91 from the interpolated coordinates (X, Y), coordinates of the source data before the output correction (X ', Y').

图17示出图16的数据输出的时序图。 FIG 17 shows a timing chart of the data output 16. 首先,ZH十算部921按照时钟CLK '1,的定时输入内插坐标(X, Y)。 First, ZH ten count unit 921 in accordance with clock CLK interpolation coordinates (X, Y) within a '1, the input timing. 然后,乘法器控制部922按照时钟CLK '2'的定时输出Z2,作为a和P,乘法器控制部924按照相同定时'2'输出失真校正系数A,作为5。 Then, the control unit 922 in accordance with clock multiplier the CLK '2' of the output timing of Z2, and P as a multiplier in accordance with the same timing as the control unit 924 '2' outputs the distortion correction coefficient A, as 5. 乘法器923按照下一时钟CLK '3,的定时,根据所输入的oc和P输出Z4,作为Y,乘法器925根据所输入的a和S按照相同定时'3'输出AZ2,作为s。 In accordance with the next clock multiplier 923 CLK '3, timing, according to the input and P oc output Z4, as Y, the multiplier 925 based on the input timing of the same a and S' 3 'AZ2, outputs, as s. 同样,按照接下来的时钟CLK '4,和'5,的定时,获得B和BZ4,作为5和s,按照时钟CLK 乂,和W的定时,获得C和CZ6,作为5和s。 Similarly, under the next clock CLK '. 4 and 5', timing, and BZ4 B obtained, and as 5 s, Yi in accordance with the clock CLK, and the timing of W, C and CZ6 by obtaining, as 5, and s. 结果,按照时钟CLK '8,的定时,从校正坐标计算部926中输出与内插坐标(X, Y)对应的源数据的坐标位置(X', Y')。 As a result, in accordance with the clock CLK '8, timing, interpolation coordinates (X, Y) output from the coordinate calculation unit 926 corrects the coordinate position corresponding to the source data (X', Y ').

在图16的结构中,乘法器只有2个,然而通过按时序使用乘法器, In the structure of FIG. 16, only two multipliers, however, by using time-series multiplier,

实际上执行相当于6个的计算。 Actually performs calculations equivalent to six. 结果可把乘法器减少到1/3。 The results can reduce the multiplier to 1/3. 而且,即使 Moreover, even

在考虑高次项的情况下,也可以进行同样的动作。 In the case of considering higher order terms can also perform the same operation.

另外,在观察图17的时序图时可知,在失真校正坐标变换处理中, Further, when viewed in the timing chart of FIG 17 seen in the distortion correction coordinate conversion process,

各时钟的各变量的输出数据具有时间间隙(空闲)。 Output of the variables of each clock with a time gap (idle). 因此,考虑通过流水线处理连续输出各变量的输出数据,使其没有时间空闲。 Thus, the continuous pipeline processing by considering the output of the variables, so no idle time.

图18对图17的动作进行了改善,示出流水线处理中的数据输出的时序图。 The operation of FIG 17 FIG. 18 is improved, a timing diagram showing data output from the pipeline processing. 按照图17的动作定时,在从内插坐标(X, Y)的输入起经过7 个时钟后,获得作为输出的源数据的坐标位置(X', Y'),即,在产生l 个结果前需要7个时钟的时间。 A timing according to operation of FIG. 17, after 7 clocks starting interpolation coordinates from the (X, Y) input to obtain a coordinate position as the source data outputted (X ', Y'), i.e., generating l Results it takes time before 7 clock. 相比之下,在图18中,由于在第10周期发生与第1周期相同的情况,因而使用9个周期进行3个像素处理。 In contrast, in FIG. 18, since the case 10 occurs in the first period of the first cycle is the same, thus using nine pixel processing cycle three. 即,看上去使用3个周期进行1个像素处理。 That is, using three cycles looks a pixel processing.

以上,对在失真校正范围计算部中仅对1个信道进行失真校正范围计算处理的示例作了描述,下面,说明使用与前述的1个信道所花时间相同的时间,在1个失真校正范围计算部中对3个信道进行失真校正范围计算处理的方法。 Above, for only one channel in the distortion correction range calculation unit for exemplary distortion correction range calculation process are described below, description of the same time period using the 1 channel spent in a distortion correction range the method of distortion correction range calculation process for calculating three channel section. 这里,如图19所示,在图12的校正图像上的范围设定中,把关注区域的4边的范围计算用像素疏化设定为例如1/3的像素数来进行运算处理,从而可按照与1个信道的处理时间大致相同的时间结束。 Here, as shown in FIG. 19, the range is set on the corrected image of FIG. 12, the range of the four sides of the region of interest calculated by the pixel thinning is set to 1/3 the number of pixels, for example, arithmetic processing is performed, whereby Following the end of the processing time may be a channel of substantially the same time. 疏化量通过寄存器设定来执行。 Thinning amount is performed by a register setting. 由于如图19所示只在疏化后的坐标(附上黑圆的位置)上进行运算,因而可利用与前述的处理时间相同的时间来对应多个信道,而且不增大电路规模。 As shown in FIG Since only the coordinates thinning (black circles attached position) on 19 calculates, thereby taking advantage of the processing time and the same time to a corresponding plurality of channels, but no increase in circuit scale. 这样,在使用例如彩色图像的情况下,即使针对R、 G、 B各方未设置失真校正范围计算部,只要通过1个电路部使3个信道动作,则1个失真校正范围计算部即可, 不增大电路规模,并且可在相同时间内完成。 Thus, for example, in the case where a color image, even for the R, G, B parties distortion correction range calculation unit is not provided, so long as the three channels by one operation of the circuit portion, the one portion to the distortion correction range calculation , without increasing the circuit size, and can be completed in the same time.

图20是对3个信道时的失真校正处理动作进行说明的流程图。 20 is a flowchart of the distortion correction processing operation when the three channels illustrated. 把图14所示的失真校正处理应用于3个信道。 The distortion correction processing shown in FIG. 14 is applied to three channels.

图20中的步骤S31〜S33和步骤S35与图14的步骤S11〜S13和步骤S15相同。 FIG step 20 S31~S33 and Steps S11~S13 S35 of FIG. 14 and step S15 of the same.

22在步骤S34中,3个信道中的各块行的失真校正处理和以下的输入范围计算同时执行,即:针对第l信道(例如,R)的下一块行的输入范 22 In step S34, the distortion correction processing and the input range of each block row 3 channels calculations performed simultaneously, namely: input of the next one line for the l channel (e.g., R) Fan

围计算一针对第2信道(例如,G)的下一块行的输入范围计算一针对第3信道(例如,B)的下一块行的输入范围计算。 Under a confining calculated for the second channel (e.g., G) of an input line calculation range for input range under a third channel (e.g., B) in a row.

这样,在使用多个信道的情况下,在对多个输入信号同时进行失真校正处理期间,针对多个输入信号,分别对作为下一处理对象的块行依次重复进行失真校正范围计算,从而可在1个失真校正范围计算部中进行多个信道的范围计算。 Thus, in the case where a plurality of channels, a plurality of input signals during a distortion correction process simultaneously for a plurality of input signals, respectively, as a next block line sequentially processed distortion correction range calculation is repeated, thereby range plurality of channels in a distortion correction range calculation unit calculates. 并且,由于在失真校正执行中计算下一次进行失真校正的图像输入范围,因而当1个块行的失真校正结束时,可知进行下一失真校正处理的图像输入范围,可没有延迟且依次顺利地对下一块行进行失真校正处理。 Further, since the calculated distortion correction execution once the input range image distortion correction, so that when one block line distortion correction ends, an image input range of the next apparent distortion correction process, sequentially and without delay can be smoothly distortion correction processing on the next one line.

如以上所述,根据本发明的第1实施方式,由于设置了用于计算进行失真校正处理的输入图像范围的失真校正范围计算部,因而可针对应输出的输出范围不多也不少地输出由失真校正所获得的校正图像。 As described above, according to the first embodiment of the present invention, since the input image is calculated for a range of the distortion correction range calculation unit distortion correction process, which can be output for the output range of no less output correction by the distortion correction image obtained. 可有效利用源数据进行失真校正处理,可进行能应对枕形失真、桶形失真以及曲弦失真的失真校正处理。 Data source can be effectively utilized for the distortion correction processing, you can be able to cope with pincushion distortion, barrel distortion and chord song distortion distortion correction processing. 并且,可算出各块行的处理所需要的输入范围,可使数据传送量縮减到最小限度。 And, processing the input range can be calculated for each row of blocks required, data transfer amount can be reduced to a minimum.

图21是示出本发明的第2实施方式的图像处理装置的整体结构的方框图。 FIG 21 is a block diagram showing an overall configuration of an image processing apparatus according to a second embodiment of the present invention. 图21具有从图1中删除了失真校正范围计算部12的结构。 FIG 21 has deleted the distortion correction range calculation unit 12 from the configuration in FIG. 图21 中的失真校正处理部的结构与图2相同。 The same structure as in FIG distortion correction processing section 212 in FIG.

失真校正坐标变换的[式1]对于整个画面中的图像即校正图像, 计算其坐标变换结果的(X', Y')的图像在整个画面中位于何处。 The image distortion correction coordinate conversion of [Formula 1] for the entire screen image i.e. corrected image, calculate the coordinate transformation result (X ', Y') is positioned where the entire screen.

然而,被输入到失真校正处理部8中的图像数据是以块行为单位。 However, 8 is input to the distortion correction processing unit image data is in units of blocks. 作为失真校正处理部8,必须知道所输入的图像数据的空间位置关系(二维空间中的位置)。 Spatial relationship as the distortion correction processing unit 8, must know the input image data (two-dimensional space location). 控制寄存器85指定在失真校正处理部8中进行处理的块行的位置。 Control register 85 to specify the position of the block line processing in the distortion correction processing unit 8.

图23和图24示出校正前的摄像图像数据和校正后的校正图像的关系。 23 and FIG. 24 shows the relationship between the captured image and the corrected image data after the correction before the correction. 图23示出摄影图像数据,图24示出校正图像。 FIG 23 shows a photographic image data, corrected image shown in FIG. 24. 图23和图24是关注于校正图像(输出侧图像)的某块行BL的某像素位置A的情况的概 23 and FIG. 24 is a schematic of a case where the pixel of interest in the position corrected image A (the output side image) of a block line BL is

23念图。 23 read the map. 校正图像(输出图像)内的位置A的坐标(X, Y)使用失真校正坐标变换的[式1 ]来变换成(输入图像)内的位置A'的坐标(X' , Y')。 A coordinate positions in the corrected image (output image) (X, Y) using the distortion correction coordinate conversion of [Formula 1] is transformed into the position of A in the (input image) 'coordinates (X', Y ').

坐标位置(X, Y)和(Xblst, Yblst)、以及坐标位置(X,, Y,)和(X, blst, Y' blst)表示距整个画面的原点的位置。 The coordinate position (X, Y) and (Xblst, Yblst), and the coordinate position (X ,, Y,) and (X, blst, Y 'blst) from the origin indicates the position of the entire screen. 这里,假定原点是整个画面的左上端部。 Here, assuming that the origin is the upper left portion of the entire screen.

被输入到失真校正块的图像数据以由符号BL表示的块行为单位。 Is input to the distortion correction block of a block of image data represented by the symbol behavior BL units. 因此,首先为了对校正图像的例如像素A进行失真坐标变换处理,必须变换成块行BL内的坐标系中的坐标位置(X', Y'),以便可进行最终的内插处理。 Thus, for example, pixel A first order image distortion correction coordinate transformation processing, the coordinate system must be converted into a block line BL in a coordinate position (X ', Y'), so as to be a final interpolation processing.

首先,图24的针对校正图像(输出图像)的各像素的坐标位置(X, Y)使用下述[式3]来生成。 First, the corrected image (output image) in FIG. 24 for each pixel coordinate position (X, Y) using the following [Formula 3] is generated. [式3] X=Xblst+mXAX Y=Yblst+nXAY [Formula 3] X = Xblst + mXAX Y = Yblst + nXAY

AX和AY是内插间距,当大于等于1.0时,生成縮小图像,当小于等于1.0时,生成放大图像。 AX and AY is an interpolation intervals, greater than or equal to 1.0, to generate reduced image, when less than or equal to 1.0, to generate an enlarged image. 坐标位置(Xblst, Yblst)相当于失真校正块的输出侧块行的左肩位置,即块行的起点位置。 Left shoulder position coordinate position (Xblst, Yblst) corresponding to the output side of the block distortion correction block line, i.e. the starting position of the block row. 当(m, n) = (0, 0) 时,X=Xblst, Y=Yblst,因而可根据[式3]计算坐标位置(Xblst, Yblst)。 When (m, n) = time (0, 0), X = Xblst, Y = Yblst, thus according to [3] calculated coordinate position (Xblst, Yblst). 从m二0, n二0开始,当使m、 n例如交替各增加l时,可生成块行BL 内的所有像素的坐标位置(X, Y)。 From m = 0, n 20 starts, when the m, n, for example, of alternating increase L, to generate the block row coordinate positions of all the pixels in BL (X, Y).

针对该BL,当根据[式l]变换BL内所有图像的坐标位置时,该坐标位置被映射到图23的所变形的阴影线区域BL"内。图23的BL, 是处理图24的BL所需要的输入数据的范围。(X' blst, Y' blst)在BL' 的左上位置,当考虑内插所需要的像素等时,与所变形的阴影线区域的左上位置不一致。BL'利用图l所说明的支持功能(失真校正范围计算部12)来设定。 Within for the BL, when the coordinate position of all the images in the [formula L] converting BL, the coordinate position is mapped to FIG. 23 hatched area BL "are deformed. BL in FIG. 23, a BL process of FIG. 24 a desired range of the input data. (X 'blst, Y' blst) at BL 'in the upper left position, when considering a pixel interpolation and the like required, and the hatched area in the upper left position is inconsistent deformation .BL' using illustrated in FIG. l support (distortion correction range calculation unit 12) to set.

使用[式1]对使用[式3]所生成的校正后的输出图像的坐标位置(X, Y)进行坐标变换,求出图23的校正前的输入图像的坐标位置(X', Y,)。 Using [Formula 1] the use of [Formula 3] coordinate position corrected output image generated (X, Y) coordinate transformation, obtains a coordinate position of the input image before the correction in FIG. 23 (X ', Y, ). 然后,通过获得坐标变换后的坐标位置(X', Y')和坐标位置(X'blSt, Y, blst)的差,可求出以块行(粗虚线的矩形)的左肩为原点的坐标 Then, the coordinate position (X ', Y') and the difference between the coordinate position (X'blSt, Y, blst) can be determined in block lines (thick broken line rectangle) as the origin of the left shoulder coordinates after coordinate conversion is obtained

位置(x,, y,)。 Position (x ,, y,).

结果,可根据该块行内的坐标位置(x', y'),进行块行单位中的内 As a result, the row block organization of the coordinate position of the block within the row (x ', y'), for

插处理。 Interpolation process.

由图23所示的输入图像侧的粗虚线所示的块行BL'由寄存器85设定,然而为了生成该值,需要考虑失真变形可算出图像数据的输入范围(失真校正范围)的支持功能。 The input side of the image shown by the thick broken line in FIG. 23 block line BL 'is set by the register 85, however, in order to generate this value, the distortion deformation can be calculated to consider the range of the input image data (distortion correction range) of the support . 该支持功能可以通过附加给失真校正处理部8的失真校正处理功能等来进行设置。 The support may be set by the distortion correction processing and the like attached to the distortion correction processing unit 8.

下面,参照图25对失真校正处理流程进行说明。 Next, with reference to FIG. 25 the distortion correction processing flow will be described. 图25是图23和图24的失真校正处理的流程图。 FIG 25 is a flowchart of FIG. 24 and 23 the distortion correction process in FIG.

首先,开始图24的施加了阴影线的位置的块行处理(步骤S41)。 First, the block line starts FIG applied hatching processing position (step S41) 24 a. 在前述[式3]中,通过设定为(m, n) = (0, 0)(步骤S42),求出图24的阴影线区域的左肩,即块行的起点坐标位置(Xblst, Yblst)。 In the [Formula 3], by setting (m, n) = (0, 0) (step S42), obtains the left shoulder of the hatched area in FIG. 24, i.e., the starting coordinate position of the block row (Xblst, Yblst ). 然后, 可根据[式3]求出校正图像的坐标位置(X, Y)。 Then, according to [3] calculated coordinate position corrected image (X, Y). 在该情况下,当使m 和n分别增加1时,可生成图24的阴影线所示的块行BL内的各像素的坐标位置(X, Y)(步骤S43)。 In this case, when the increase in n and m are 1, can generate the coordinate position of each pixel in FIG. (X, Y) (step S43) in the block BL hatched line 24 shown in FIG.

然后,使用[式l]把校正图像中的坐标系的位置(X, Y)变换成校正前的摄影图像中的坐标系的位置(X', Y')(步骤S44)。 Then, using [Formula L] in the position of the corrected image coordinates (X, Y) into a captured image before the correction of the position coordinates (X ', Y') (step S44).

然后,使用前述方法把摄影图像中的坐标系的位置(X', Y')变换成输入块行BL内的坐标系的位置(x', y')(步骤S45)。 Then, using the photographic method the position in the image coordinate system (X ', Y') into a position in the coordinate system of the input block line BL (x ', y') (step S45).

然后,根据该块行BL内的坐标位置(x', y')进行内插处理,通过使用坐标位置(X', Y')中的16点内插等,根据周边的多个像素数据算出内插数据,获得1个像素的校正图像数据(步骤S46)。 Then, the coordinate position within the block line BL (x ', y') interpolation processing by using the coordinate position ( 'Y,' X) over 16 other interpolation points, calculated from the data of a plurality of pixels surrounding interpolated data to obtain corrected image data of one pixel (step S46).

然后,判断对校正图像的坐标系的BL内所有像素是否执行了以上步骤(步骤S47),当未结束时,设定下一m和n (步骤S48),重复执行步骤S43〜S47,当通过步骤S47的判定,对所有像素的处理已结束时, 转移到步骤S49。 Then, it is determined whether or not performed on all the pixels in the coordinate system of the corrected image BL of the previous step (step S47), when not ended, the next set of m and n (step S48), repeat steps S43~S47, as by the determination of step S47, when the processing for all the pixels has been completed, proceeds to step S49.

在步骤S49中,判断针对整个画面(帧)内的所有块行BL的失真校正处理是否结束,当未结束时,选择下一输入块行BL,重复步骤S41〜步骤S49,当通过步骤S49的判定,对整个画面的处理已结束时,结束整个过程。 In step S49, judges whether all the blocks BL in the row (frame) distortion correction processing is finished for the whole screen, when not completed, select the next input block line BL, repeat steps S41~ step S49, step S49 is when determination processing has been completed for the entire screen, the process ends.

图26〜图29应该说是图25的变形例。 FIG. 26~ 29 should be said that modification of the embodiment 25 of FIG.

在进行数字变焦的情况下,或者在仅切取所关注的小区域的情况下, 可以仅使用摄像数据的一部分。 When performing digital zoom, or when only a small cut in the region of interest, may use only a portion of the image data. 通过有选择地从CCD取入摄像数据的一部分,可节约存储器。 By selectively taking in a part of the imaging data from the CCD, the memory can be saved. 在该情况下,为了以块行为单位进行失真校正, 需要所取入的数据在摄像元件上位于何处的信息(X, imgst, Y, imgst)。 In this case, in order to perform the distortion correction in units of blocks, the data needed taken where the information (X, imgst, Y, imgst) located on the imaging element.

图26和图27示出数字变焦的情况,从所拍摄的图像数据中仅取入中央部分。 26 and FIG. 27 shows a case of digital zoom, the image data captured only taken in the central portion. 与图25的情况的不同点是,如上所述,在预处理电路2中设定在整个画面中取入何处。 Different from the case of FIG. 25, as described above, the entire screen is set where the fetch in the pre-processing circuit 2. 根据该设定仅取入必要部分,该范围由图26 的E'表示。 Taken according to the setting only the necessary portion of the range 'is represented by E in FIG. 26.

首先,在图27中,设定与整个画面中的进行数字变焦的范围E的左肩对应的起点位置(Ximgst, Yimgst)。 First, in FIG. 27, the range E is set with the left shoulder of the entire screen corresponding to the digital zoom start position (Ximgst, Yimgst). 然后,设定块行BL,这里是由斑点图形表示的部分。 Then, setting block line BL, there is a portion indicated by speckle patterns. 此时,决定以取入范围E的左肩为原点的位置坐标(Xblst, Yblst),与[式2] —样通过使m和n变化,可生成块行BL内的各像素的坐标。 In this case, the decision was taken as the position range E of the left shoulder coordinate origin (Xblst, Yblst), and [Formula 2] - m and n by like changes may coordinates of each pixel in the block line BL generated.

结果,根据坐标位置(Ximgst, Yimgst)、 (Xblst, Yblst)以及(mXAX, nXAY)决定整个画面的坐标位置(X, Y)。 As a result, the coordinate position (Ximgst, Yimgst), (Xblst, Yblst) and (mXAX, nXAY) determines the position of the entire screen coordinate (X, Y).

摄影图像中的坐标位置(X', Y')可根据上述决定的坐标位置(X, Y)使用[式l]来求出。 Photographic image coordinate position (X ', Y') may be used [Formula L] in accordance with the determined coordinate position (X, Y) is obtained.

然后,求出摄影图像数据中的块行BL'中的坐标位置(x', y')。 'Coordinate position (x in', Y ') and then, the photographed image data is obtained in block line BL.

首先,根据所取入的摄像数据的摄像面上的设定范围E',知道左肩的坐标(X, imgst, Y, imgst)。 First, the setting range of the imaging surface of the imaging data taken E ', know the coordinates of the left shoulder (X, imgst, Y, imgst). 在以取入范围E,的左肩为原点的坐标中, 块行BL'的左肩的坐标位置(X, b!st, Y' blst)利用支持功能预先求出, 决定在块行BI/内的坐标的坐标位置(x', y')。 In order to get into the range E, the left shoulder to the origin coordinates, the block line BL 'of the left shoulder of the coordinate position (X, b! St, Y' blst) obtained in advance using a support functionality, determined at block row BI / in coordinate position coordinates (x ', y'). 结果,进行各块行的内插处理。 As a result, each interpolation process block lines.

图28和图29是从摄像数据中切取某区域,例如画面端侧的情况。 Figures 28 and 29 are cut out from the image data of a region, for example, where the side end of the screen. 这相当于在监视照相机等中取入摄像数据的端部分的情况。 This corresponds to the case where the end portion of the image data taken in a surveillance camera or the like. 图28和图29 的情况也附上与图26和图27相同的符号。 The case of FIGS. 28 and 29 are also attached with the same reference numerals in FIG 26 and FIG 27. 与图26和图27只是切取位 FIG 26 and FIG 27 is only cut position

26置不同,采用与图26和图27的情况相同的动作。 Different set 26, the same in the case of FIGS. 26 and 27 of operation. 另外,在通常的数字照相机中,由于CPU把切取区域设定在控制寄存器内,因而知道切取位置,可进行上述变换,然而在对记录后的数据进行失真校正的情况下,必须把切取区域的位置和尺寸同时记录在标记等内。 Further, in the conventional digital camera, since the CPU the cutout area is set in the control register, and thus to know cut position, can perform the conversion, however, in the case where data recording distortion correction, it is necessary to cut out areas simultaneously recording the position and size within the tag and the like. 此时,同时也记录校正系数。 In this case, the correction coefficient is also recorded. 另外,在作为摄像元件的CCD内大多准备有监视模式(不读出所拍摄的整个画面的模式)。 Further, as in the CCD image pickup element are prepared mostly monitor mode (read mode is not taken of the entire screen). 在监视模式中,行大多被疏化,即使在摄像元件中的二维空间上相同,有时在数据上尺寸也不同。 In the monitor mode, lines are thinned often, even on the same two-dimensional space of the image pickup element, may be different in data size. 并且,在预处理中, 有时横方向的像素也被疏化。 Further, in the pretreatment, the pixel horizontal direction may also be thinned. 并且,有时对来自CCD的摄像数据实施低通滤波器(LPF)等的滤波处理,或者进行内插处理以取入到帧存储器内。 And sometimes a low pass filter (LPF) for filtering process such image data from the CCD, or the interpolation process in order to get into the frame memory. 在该情况下,发生纵横抽样比不是整数的情况。 In this case, where the aspect ratio is not an integer sampling occurs. 而且,在JPEG中的YCbCr 的记录中,Cb和Cr成分的横方向大多以2倍间隔被抽样,作为数据成为横方向縮小的形式。 Further, in the recording in a JPEG YCbCr, Cb, and Cr components of the lateral direction at most twice the sampling intervals, the lateral direction becomes narrow as the data form. 这些数据在[式l]中使用Sx、 Sy进行校正。 The use of data Sx, Sy corrected in [Formula L] in. 图30示出在摄像时的数据取入中,实际从CCD所发送的数据仅在黑圆点的位置处,在纵方向上进行3行疏化的状态。 FIG 30 shows the data taken at the time of imaging, the actual position of the black dot at, for 3-line thinning in the vertical direction of the state data transmitted from the CCD only. 在监视模式中,像这样从CCD仅输出疏化后的数据,当把该数据取入到存储器中时,如图31所示由于疏化而成为在上下方向上被破坏的图像数据。 In the monitoring mode, only data output from the CCD after the thinning this manner, when the data is taken into the memory, as shown in FIG. 31 due to the thinning image data becomes vertically destroyed. 如图32所示, 即使取出在空间上距中心相同距离的数据,如图31所示,也发生图像在纵方向上相对于横方向输出较短的现象。 As shown in FIG. 32, taken in space even if the same distance from the center of the data shown in Figure 31, the image output short also occurred a phenomenon transversely to the longitudinal direction. 结果,在基于[式l]的失真校正中,距中心的距离越远,失真校正效果就越大,因而在纵方向和横方向上,失真校正效果发生变化。 As a result, the distortion correction based on [Formula L], the greater the distance from the center, the greater the effect distortion correction, and thus in the longitudinal direction and the lateral direction, the effect of the distortion correction is changed. 这样,以对图像纵横非对称进行校正为目的,设置了校正系数Sx和Sy。 Thus, in order to correct image aspect asymmetric for the purpose of setting the correction coefficient Sx and Sy. 在图31的情况下,由于必须在纵方向上伸展,因而设定Sx-l和Sy =4。 In the case of FIG. 31, since it is necessary extension in the longitudinal direction, the set Sx-l and Sy = 4. 根据该Sx和Sy的效果,即使纵横抽样比不同,也能进行准确的失真校正。 The effect of the Sx and Sy, even if the aspect ratio of the different samples, it is possible to accurately perform distortion correction. 并且,如前所述,在抽样比不是整数的情况下,Sx和Sy也可以设定为非整数。 Further, as described above, in the case where the sampling ratio is not an integer, to Sx and Sy may be set to a non-integer. 而且,在摄像元件的各像素的间隔纵横不同的情况下,也能进行校正(例如矩形像素)。 Further, each pixel in different vertical and horizontal spacing of the imaging element, the correction can be performed (e.g., rectangular pixels). 27另外,通常,当是JPEG时, 一般使用YCbCr的数据来记录,然而作为此时的设定,考虑图33和图34所示的情况。 27 Further, generally, when a JPEG, YCbCr generally used to record data, but as this time is set, consider the case shown in FIGS. 33 and 34. 例如,在YC422的形式的数据的情况下,与亮度数据相比颜色数据的分辨率在横方向为1/2。 For example, in the case of YC422 form data, color data as compared with the luminance data resolution in the lateral direction is 1/2. 在该情况下,针对颜色数据设定Sx^2和Sy二l,可以进行校正处理。 In this case, setting two Sy Sx ^ 2 and l for the color data, the correction processing may be performed. 另外,在失真中心和图像中心不一致的情况下,在校正前位于中心的被摄物体在校正后偏移中心。 Further, in a case where the image center and the center of distortion inconsistent, centrally located in the pre-correction in the corrected subject off center. 在[式l]中,通过使用中心偏移校正值(X。ff, Y。ff)进行校正,可使位于图像中心的被摄物体在校正前后不变化。 In [Formula L] by using the center offset correction values ​​(X.ff, Y.ff) correction, allows the center of the image of the subject does not change before and after correction. 图35是对图像中心偏移进行说明的图。 FIG 35 is a diagram for explaining the image center is offset. 图35示出校正图像。 35 shows a corrected image. 在前述的图4中,是失真中心和图像中心一致, 即CCD的中心(图像中心)与镜头光轴(失真中心) 一致的情况。 In the FIG. 4, center of the image distortion is consistent with the center, i.e., the same center of the CCD (image center) lens axis (the center of distortion) conditions. 图35是失真中心相对于CCD的中心偏移的情况。 FIG 35 is distorted with respect to the center of the center of the CCD shift. 在图35中,由于失真中心(Xd, Yd)与图像中心不同,因而在不考虑中心偏移而进行失真校正的情况下,输出范围H内的图像中心P偏移到Q点(由大的黑圆表示)。 In the case of FIG. 35, since the distortion center (Xd, Yd) different from the image center, so the distortion correction is performed without considering the shifted center, center of the image within the output range P H is shifted to the point Q (by a large black circles show). 对用户来说,当实际拍摄的图像和失真校正后显示的图像中心偏移时,感到非常不协调。 For users, when the center of the display image after image distortion correction and offset actual shooting was very inconsistent. 因此,为了使本来在失真校正前位于输出范围H 的中心的数据相对于输出范围处于中心位置,改变失真校正后的图像数据的切取方法。 Accordingly, in order to make the data center have been located in the output range of H before the distortion correction with respect to the center position in the output range, method of changing the cut out image after distortion correction data. 为了执行该方法,只要使用(X。ff, Y。ff)定义图像中心偏移,改变图像数据的切取方法即可。 To perform this method, as long as the use of (X.ff, Y.ff) define an image center offset change cut out to process the image data. 通过设定图像中心偏移(X。ff, Y。ff), 使输出范围从H移动到H,,可将本来位于图像中心P的被摄物体校正成, 即使在校正后的输出范围HT内也位于中心。 By setting the center of the image shift (X.ff, Y.ff), the output range from H to H ,, can move original center P of the image is corrected to the subject, the output range even after correction HT also in the center. 根据图35的实施例,可消除由于中央的被摄物体偏移而引起的不协调感。 According to the embodiment of FIG. 35, the sense of discomfort can be eliminated since the center of the subject caused by the offset. 以上所述,根据本发明的第2实施方式,可算出小区域(例如块行) 的空间位置,可实现小区域单位中的失真校正处理。 Described above, according to the second embodiment of the present invention, a small area can be calculated (e.g. block line) spatial position, the distortion correction processing can be realized in units of small areas. 并且,可对疏化取入的图像数据进行准确的失真校正,另一方面,还可对数字变焦等,切取了任意区域的数据进行准确的失真校正。 Then, distortion correction can be accurately performed on the image data thinning fetched, on the other hand, digital zoom, etc. may be excised arbitrary data area for accurate distortion correction. 本发明的第3实施方式的图像处理装置的整体结构与图21相同。 The overall configuration of an image processing apparatus according to a third embodiment of the present invention 21. 图37示出图2中的失真校正处理部8的详细结构。 37 shows a detailed configuration of the distortion correction processing section 28 of. 对图37的各部和图2的各部的对应关系进行说明。 Correspondence relation of each part and each part 37 of FIG. 2 will be described. 图37中的内插位置计算电路22与图2的内插位置生成部811对应,选择器24与图2 的选择器813对应,失真校正系数计算电路21和内插位置校正电路23 与图2的失真校正坐标变换部812对应。 Interpolation position in FIG. 37 calculates the interpolation position generating unit 811 corresponds to the selector circuit 22 of FIG. 2 selector 24 of FIG. 2 813 corresponds to the distortion correction coefficient calculation circuit 21 and the interpolation position correcting circuit 23 of FIG. 2 corresponding to the distortion correction coordinate conversion portion 812. 并且,图37中的2端口SRAM 26与图2的内部存储部(内部缓冲器)82对应,写入地址生成电路28、 缓冲器空闲容量监视电路29、数据可否发送判定电路30、缓冲器开放量计算电路31以及读出地址生成电路25与图2的存储器控制部83对应, 内插电路27与图2的内插运算部84对应。 Further, in FIG. 37 with the two-port SRAM 26 internal storage unit (internal buffer) corresponding to 82 of FIG. 2, the write address generating circuit 28, the free capacity of the buffer monitoring circuit 29, whether the transmission data determination circuit 30, a buffer open amount calculation circuit 31 and the read address generating circuit 83 corresponding to the memory control unit 25 of FIG. 2, the interpolation circuit 27 of FIG. 2 corresponds to the interpolation operation unit 84. 错误检测电路32在图2中未作图示,但被设置成与存储器控制部83连接。 The error detection circuit 32 in FIG. 2 not shown, but is provided connected to the memory control unit 83. 错误检测电路32具有以下功能,即:当在失真校正处理中失真量增大而发生超出后述preULB和postULB的设定值的失真时,作为错误(ERROR)传递给CPU4。 Error detection circuit 32 has a function, namely: when the amount of distortion in the distortion correction processing is increased and the distortion of said setpoint preULB postULB occurs beyond the post when passed as CPU4 to error (ERROR). 内插位置计算电路22在根据来自失真校正处理部8的后级电路的请求向后级电路返回了同意(接受请求)时,把该同意作为触发,计算1 单位行(以下称为1UL)的内插位置(Xp Y,)。 Interpolation position calculating circuit 22 at the subsequent stage circuit is returned consent (request accepted) a request from a subsequent circuit distortion correction processing unit 8, which agreed to as a trigger, a calculating unit row (hereinafter referred 1UL) of interpolation position (Xp Y,). 这里,1UL是指在前述块行处理中,对存储部进行写入和读出时,写入或读出在列方向排成一列的恒定数目的图像数据时的一个单位。 Here, 1UL means when the block line processing, the storage unit for writing and reading, when a unit of writing or reading in a row in the column direction of the image data of a constant number. 即,1UL是指块行(BL)上的在列方向排成一列的恒定像素数的数据。 That is, 1UL constant refers to the number of data of pixels arranged in a row in the column direction row block (BL) on. 内插位置校正电路23使来自失真校正系数计算电路21的失真校正系数F与内插位置(Xp Y,)相乘,计算失真校正前的源数据的坐标位置(X', Y')。 Interpolation position correction circuit 23 so that the distortion correction coefficient calculating circuit F from the interpolation position (Xp Y,) is multiplied by the distortion correction coefficient 21 calculates the coordinate positions of the source data before distortion correction (X ', Y'). 选择器24选择(X^ Y。和(X', Y'),在进行失真校正的情况下,选择输出(X,, Y'),在仅进行放大/縮小处理(调整尺寸) 的情况下,选择输出(Xp Y】)。2端口SRAM26是存储失真校正处理部8内的数据的缓冲器。读出地址生成电路25生成与内插位置对应的2端口SRAM 16内的地址(ADR),输出用于使来自2端口SRAM26的输出对准的控制信号, 与输出图像数据同步输出写入控制信号WE—N,输出用于告知图38和[式3]所示的Do在2端口SRAM上的何处位置的数据串控制信号。写入地址生成电路28根据写入控制信号WE,生成作为内部存储器的2端口SRAM 26的地址(ADDRESS),当1UL的数据输入完成时,29使其内部计数器(BLC)递增计数。数据可否发送判定电路30具有以下功能,即:根据BLC值、本电路的动作状态、下一UL开头坐标、以及来自失真校正处理部8的后级电路的请求(REQ)状态,判定是 The selector 24 selects (X ^ Y. and (X ', Y'), the case of performing the distortion correction, selects the output (X ,, Y '), performed only in the case of enlargement / reduction processing (resizing) of select the output (Xp Y]). SRAM26 port 2 is stored in the buffer for data distortion correction processing unit 8 read out address generating circuit 25 generates an address (ADR) 16 within the interpolation position corresponding to the two-port SRAM, output a control signal for output from the second port SRAM26 aligned with the output image data out of the write control signal WE-N, and an output 38 for informing the [formula 3] as shown in the 2-port SRAM Do where the control signal data string position. write address generating circuit 28 in accordance with the write control signal WE, generate two-port SRAM address (aDDRESS) of 26 as an internal memory, when the data input is completed 1UL, 29 so that the internal counter (BLC) incrementing the count data transmission availability determination circuit 30 has a function, namely: according to the BLC value, the operation state of this circuit, the beginning of the next UL coordinates, and a request from the rear-stage circuit of the distortion correction processing section 8 (REQ ) state is determined 处于可针对来自后级电路的REQ信号发送同意(GRANT—N)的状态,当可发送同意时,设定GRANT一N-1。内插电路27针对与内插位置对应的图像数据进行16点内插。缓冲器幵放量计算电路31计算当前处理中的UL开头坐标和下一处理预定的UL开头坐标的整数部的差作为缓冲器开放量(参照图50)。缓冲器空闲容量监视电路29保持存储在作为内部缓冲器的2端口SRAM 26内的数据(UL)量,当缓冲器有空闲时,把请求(REQ=1) 发送到失真校正处理部8的前级电路中。下面,对图37的失真校正处理部8的电路动作进行说明。首先,在作为内部存储部(内部缓冲器)的2端口SRAM26有空闲的情况下,当从缓冲器空闲容量监视电路29向前级电路发送作为数据请求的请求(REQ)时,缓冲器空闲容量监视电路29从前级电路接收作为接受请求的同意(GRANT),与该接收同时,使存储有可存储在2端口SRA In the REQ signal may be sent from the consent for the subsequent stage circuit (GRANT-N) state, can be transmitted when its consent GRANT a N-1. The interpolation circuit 27 16:00 the image data corresponding to the interpolation position interpolation. Jian buffer volume calculating circuit 31 calculates the difference UL at the beginning of the coordinates of the current process and the integer portion of the predetermined process at the beginning of the next UL coordinates as a buffer amount of opening (see FIG. 50). monitoring the free capacity of the buffer circuit 29 a holding data stored in the internal buffer of the two-port SRAM 26 (UL) the amount of free time when the buffer, the request (REQ = 1) is sent to the front stage circuit of the distortion correction processing section 8 in. next, distortion correction processing unit 37 of the circuit operation of FIG. 8 will be described. first, in the idle port 2 as an internal storage unit SRAM26 (internal buffer), when the free capacity of the buffer is transmitted from the monitoring circuit 29 forwards stage circuit (REQ) when the free capacity of the buffer monitoring circuit 29 as a preceding stage circuit receives a consent request data request (the GRANT) accepts the request, the receiver simultaneously with the port 2 is stored may be stored in SRA M26内的UL数的计数器(在电路29内)减l。根据1次请求和同意,以1UL为一动作单位进行数据传送。当上述计数器为O时,撤消请求。然后,数据从前级电路流入写入地址生成电路28,进行对2端口SRAM 26的写入。每输入1UL,写入地址生成电路18的内部计数器(BLC) 就递增计数。2端口SRAM 26使用4X4共计16个构成如图39所示的可同时进行读出和写入的2端口SRAM,以便能在内插电路27进行例如16点内插。 The counter of the number of UL in M26 (in the circuit 29) Save l. According to 1 request and agrees to 1UL as an operation unit of data transfer when said counter is O, retraction. Then, the data from the front stage circuit flows write address generating circuit 28, write to port 2 of the SRAM 26. 1UL each input, internal write address counter (BLC) circuit 18 to generate a count-port SRAM 26 .2 16 configured using 4X4 total 39 2 can simultaneously read and write port SRAM shown, to enable the interpolation circuit 27 performs interpolation, for example, 16:00. 这里,使用图39对2端口SRAM 26进行说明,然而存储器数和各存储器尺寸可以不同。 Here, FIG. 39 pairs of two-port SRAM 26 will be described, however, the memory size of each memory and the number may be different. 例如,在图39中,把块行宽度(UL长度)设定为96,然而可以更长,也可以更短。 For example, in FIG. 39, the line width of the block (UL length) is set to 96, but can be longer, or be shorter. 可以根据电路规模和校正性能的平衡来决定。 It may be determined according to the circuit scale and balance correction performance. 并且,4X4 (计16个)也是,如果是4点内插,则可以是2X2 (计4个)。 And, 4X4 (gauge 16) is, if the four-point interpolation is, it may be a 2X2 (count 4). 可以根据内插方式(进行几点内插)来决定。 It can be determined according to the interpolation mode (interpolation points). 图39是在由16个2端口SRAM构成的存储空间中,横方向是所输入的UL顺序,而纵方向是对各2端口SRAM所附的地址。 FIG 39 is a storage space 16 by the two-port SRAM configuration, the lateral direction is sequentially inputted UL, while the vertical direction the respective 2-port SRAM appended address. 设N为整数, 对处理的某时刻的状态进行说明。 Let N be an integer, the state of a process time will be explained. 作为在纵方向排列的4个各2端口SRAM的No. 0、 4、 8、 12内存储的数据是第4N、 4N+4、 4N+8、 4N 十12的UL数据,No. 1、 5、 9、 13内存储的数据是第4N+1、 4N+5、 4N+9、 4N+13的UL数据,No. 2、 6、 10、 14内存储的数据是第4N+ 2、 4N+6、 4N+10、 4N+14的UL数据,No.3、 7、 11、 15内存储的数据是第4N+3、 4N+7、 4N+11、 4N+15的UL数据。 As arranged in the longitudinal direction of each of the four 2-port SRAM No. 0, 4, 8, 12 stores the data in the first 4N, 4N + 4, 4N + 8, UL data 4N ten 12, No. 1, 5 9, the data 13 stored in the first 4N + 1, 4N + 5, 4N + 9, UL data 4N + 13 is, No. 2, 6, 10, the data 14 stored in the first 4N + 2, 4N + 6 , 4N + 10, UL data is 4N + 14, No.3, 7, 11, 15 stores the data in the first 4N + 3, UL data 4N + 7, 4N + 11, 4N + 15 in. 对于作为在横方向排列的4个各2端口SRAM的No. 0、 1、 2、 3, No. 4、 5、 6、 7, No. 8、 9、 10、 11以及No. 12、 13、 14、 15的各组,也进行同样的存储。 As for the horizontal direction arranged in each of four 2-port SRAM No. 0, 1, 2, 3, No. 4, 5, 6, 7, No. 8, 9, 10, 11 and No. 12, 13, 14, 15 of each group, and subjected to the same storage. 使用图40和图41,对在各2端口SRAM内写入数据的状态进行说明。 FIG 40 and FIG 41, the state of writing data within each of the 2-port SRAM will be described. 在图40和图41中,如图40所示,假定了宽度为8行(像素)的块行,数据以UL为单位按照"1、 2、…、i、 m"、 "5、 6、…、j、 n"、 "9、 10、…、k、 o"、…这样的顺序输入。 In FIG. 40 and FIG. 41, FIG. 40, it is assumed that the block line width of 8 rows (pixels), the data units according to UL "1, 2, ..., i, m", "5, 6, ..., j, n "," 9, 10, ..., k, o ", ... in this order input. 图41示出图40所示的写入顺序的数据被写入到4X4的16个2端口SRAMNo. 0〜No. 15上的何处。 Data is written in the order shown in FIG. 40 is shown in FIG 41 is written into the second port 16 SRAMNo. Where the 0~No. 15 4X4 in. 例如在No. 0、 No. 4、 No. 8、 No. 12内分散存储有图40的左侧纵一列的1UL 的图像数据(按照顺序l〜m所示的数据)。 For example, in No. 0, No. 4, No. 8, No. 12 dispersed within the stored image data (data in the order shown l~m) on the left side of FIG. 40 is a longitudinal of the 1UL. 另夕卜,针对图39所示的纵方向的2端口SRAM No. 0、 No. 4、 No. 8、 No. 12的各纵行所附的标记0〜 92、 1〜93、 2〜94、 3〜95与图40所示的各单位行(1〜m…、5〜n…、 9〜0…、n,,)的标记不同。 Another Bu Xi, for the longitudinal direction 39 shown in FIG. 2 port SRAM No. 0, No. 4, No. 8, No. 12 of each of the following markers 0~ wales 92, 1~93, 2~94 , each different row units (1~m ..., 5~n ..., 9~0 ..., n ,,) shown in FIG. 40 labeled with 3~95. 这是因为,在图40和图41中,为了知道对于4X4的16个图像数据的写入顺序,将16个块分别附上1〜16、 a〜p、 A〜P、…的符号。 This is because, in FIG. 40 and FIG. 41, in order to know the sequence for writing the image data 16 of 4X4, the blocks 16 are attached 1~16, a~p, A~P, ... symbols. 在图39的情况下,对该纵方向的2端口SRAMNo,O、 No. 4、 No. 8、 No. 12的各纵行所附的标记相当于纵方向的1个UL中的数据顺序。 In the case of FIG. 39, the second port SRAMNo the vertical direction, O, No. 4, No. 8, the sequence of data appended wale No. 12 corresponds to the longitudinal direction of the marker 1 in the UL. 在2端口SRAM No. 0、 No. 4、 No. 8、 No. 12的各方内交替分散记载1个UL像素数据No. 0〜95。 In the two-port SRAM No. 0, No. 4, No. 8, according to an alternate dispersion UL pixel data No. 0~95 No. 12 of the parties. 因此,对图39的纵方向所排列的4个各纵行附上表示是1个单位行(UL)的像素顺序(0〜95)的第几像素的标记。 Accordingly, the arrangement of the vertical direction in FIG. 39 each of the four wales are attached represent a marker line units (UL) sequentially pixel (0~95) of the first few pixels. 在图39的2端口SRAM的No. 0的纵方向存储有UL的4倍31数的像素(0、 4、 8…92),在2端口SRAM的No.4的纵方向存储有UL 的(4倍数)+1的像素(1、 5、 9…93),在2端口SRAM的No.8的纵方向存储有UL的(4倍数)+2的像素(2、 6、 10…94),在2端口SRAM 的No. 12的纵方向存储有UL的(4倍数)十3的像素(3、 7、 11…95)。 The UL 31 4 times the number of pixels (0, 4, 8 ... 92) in the longitudinal direction of the No. 2-port SRAM memory 39 of FIG 0, 2 in the longitudinal direction of the storage-port SRAM No.4 have the UL ( a multiple of 4) +1 pixels (1, 5, 9 ... 93), in the longitudinal direction of the 2-port SRAM memory has UL No.8 (4 times) the pixel (2, 6, 10 ... 94) +2 in the longitudinal direction of the 2-port SRAM memory has UL No. 12 (4 times) + 3 pixels (3, 7, 11 ... 95). 而且,对于2端口SRAM的No. 1、 2、 3, No. 5、 6、 7, No. 9、 10、 11 以及No. 13、 14、 15的各组,也进行同样的存储。 Also, for the 2-port SRAM No. 1, 2, 3, No. 5, 6, 7, No. 9, 10, 11 and No. 13, 14, 15 of each group, and subjected to the same storage. 再次返回到图39, 1UL的数据按照每一个像素依次被存储在排列于纵方向的4个2端口SRAM内。 Returning again to the data of FIG. 39, 1UL sequentially according to each pixel is stored in the vertical direction are arranged in four 2-port SRAM. 最初的UL数据被写入到No. 0、 4、 8、 12 (图39的4N的行)中。 The initial UL data is written to No. 0, 4, 8, (4N row of FIG. 39) 12. 下一UL数据被写入到No. 1、 5、 9、 13 (图39的4N+1的行)。 UL data is written to the next, 5, 9, (FIG. 39, row 4N + 1 in) 13 No. 1. 在图39的缓冲器形式中,采用以下结构,即:通过失真校正处理使内插位置无论怎样移动,都能通过一次存取同时抽出缓冲器内的任意坐标周边16点。 FIG form in the buffer 39, the following structure, namely: no matter how to move by interpolating distortion correction processing so that the position, can be withdrawn simultaneously by a single access to an arbitrary coordinate in the buffer surrounding 16:00. 数据可否发送判定电路30从后级电路接收请求(REQ—N),当处于发送下一UL数据的状态时,输出接受请求(GRANT一N)。 Determining whether the data transmission circuit 30 receives a request from the post-stage circuit (REQ-N), when in the state of the next UL data transmission, the output request is accepted (the GRANT a N). 该接受请求(GRANT一N)成为对自己自身的触发,内插位置计算电路22幵始动作, 在进行了1UL动作后,计算下一UL开头坐标,然后结束。 Accepts the request (the GRANT a N) to become their own trigger, the interpolation position calculation circuit 22 starts its operation, after performing the operation 1UL, calculated at the beginning of the next UL coordinates, and then ends. 内插位置计算电路22幵始动作,为了与输出内插位置(Xl, Yl) 同步,从数据可否发送判定电路30向失真校正系数计算电路21发送动作触发(trig)。 Interpolation position calculating circuit 22 starts its operation, to output the interpolated position (Xl, Yl) synchronized, determines whether the data transmission circuit 21 transmits an operation circuit 30 calculates a distortion correction coefficient to the trigger (trig). 失真校正系数计算电路21也与内插位置计算电路22 — 样,在进行了1UL动作后,计算下一UL开头坐标,然后结束。 The distortion correction coefficient calculation circuit 21 also calculates the position of the interpolation circuit 22 - comp, 1UL after performing the operation, at the beginning of the next UL calculates the coordinates, and then ends. 读出地址生成电路25根据所输入的内插坐标,把读出地址发布给16个2端口SRAM26的各方。 Read address generating circuit 25 based on the interpolation coordinate input, the read address to the release parties SRAM26 the second port 16. 下面,对读出和内插方法进行说明。 Next, the read and interpolation method will be described. 图38示出内插方法。 38 shows an interpolation method. 图38是内插电路27中的内插运算的图。 FIG 38 is an interpolation operation in the interpolation circuit 27 in FIG. 内插坐标位置P的坐标(X,, Y')使用前述[式l]已求出。 Interpolating coordinate position of the coordinates P (X ,, Y ') using the above [Formula L] was obtained. 为了求出该坐标中的像素值(亮度数据),根据坐标P (X', Y')的周边16点的像素数据Do〜Du来求出。 In order to determine the surrounding pixel value (luminance data) of the coordinates, the coordinate P (X ', Y') of the pixel data Do~Du 16 points is obtained. 其中,如果知道Do来自16个存储器中的哪个存储器,则根据相对于Do的位置关系可知D!、 D2-、 D15。 Wherein, if you know Do from the memory 16 in the memory which, it is known that D !, D2-, D15 according Do relative positional relationship. 如后所述,Do使用内插位置的坐标来求出。 As described later, the coordinates of the interpolation position is obtained using Do. 32图42示出从由16个2端口SRAM构成的缓冲器的读出例。 32 shows an embodiment of FIG. 42 from the read buffer 16 by the 2-port SRAM configuration. 现在, 假定内插位置(X', Y,)为X, =10.…,Y, =50.…(…是小数点以下的数字),则X' =10.…大于等于(4X0 + 10),因而在横方向,内插位置位于2端口SRAM的第4N+10的UL的稍微右侧。 Now, assuming that the interpolation position (X ', Y,) for the X, = 10. ..., Y, = 50. ... (... is a decimal number), then X' = 10. ... greater than or equal (4X0 + 10) UL slightly right, and thus in the transverse direction, the interpolation position is located within the first 2-port SRAM is 4N + 10. Y, =50.…大于等于(4X12+2),根据第4N+10的UL,存储有第50的像素的存储器(No. 10)输出图38中的像素D5,因而输出Do的存储器为其左上的No. 5。 Y, = 50. ... greater than or equal (4X12 + 2), according to UL first 4N + 10, the pixel is stored in D5 38 (No. 10) of the output pixel map memory 50, thus the output of the memory for Do No. 5 upper left. 由于与图38的D。 Since FIG 38 D. 〜D。 ~D. 对应的图像数据是图42的〇部,因而生成地址,以便输出这些像素数据。 The image data corresponding square portion in FIG. 42, thus generating an address to output the pixel data. 从No. O输出的图像数据与Do不对应。 Image data output from the No. O Do not correspond. 在图42的例中,来自No.5 的输出与Do对应。 In the embodiment of FIG. 42, No.5 from the output of the corresponding Do. 因此,为了识别从哪个存储器输出哪个数据,从读出地址生成电路25输出数据串控制信号,这样,进行内插运算的内插电路27识别Do从2端口SRAM26的何处输出,进行16点内插。 Accordingly, in order to identify which of the memory from which data output, the control circuit 25 outputs the data string is generated from the read address signal, so that, for the interpolation operation Do 27 recognizes the interpolation circuit 2 where the output port SRAM26 performs the 16:00 plug. 只要知道像素数据Do〜D,5,就能通过使用[式4]的内插式进行内插处理,求出校正坐标位置的像素数据作为D。 Pixel data only know Do~D, 5, the interpolation process can be carried out by using [Formula 4] is interpolated to determine the coordinate position corrected pixel data as D. ut。 ut. [式4]Dout-kxo (kyo Do + kyi D4 + ky2 D8 + ky3 D.12 ) 十kx! (kyq D|.+ky.| Ds十ky2 D3 + ky3 D13 ) + kx2 (ky0 D2 + kyi D"ky2 D!o +ky3. D|4 〉 + "3 (kyo D3 + ky| D7 + ky2 Dii +'ky3 D15 )在最后的UL数据输出后,缓冲器开放量计算电路31计算目前处理结束的UL开头坐标和下一UL开头坐标的差(参照图50),为了开放存储有不必要数据的缓冲器,把缓冲器开放量输出给缓冲器空闲容量监视电路29。然而,优选的是,跨过失真中心时的缓冲器开放量如图49所示, 开放也考虑了postULBl和postULB2的值变化的量。图48所示的preULB 和postULB针对形成UL的像素中的最初像素的坐标位置,针对行方向在前侧和后侧设置规定宽度的区域,将它们各自定义为preULB和postULB。在跨过失真中心时,缓冲器开放量不是通常的UL开头坐标的差,通过参照使参照值变化的postULB (ULB: Unit Line Buffer的简称,单位行缓冲器)的变化量来调节缓 [Formula 4] Dout-kxo (kyo Do + kyi D4 + ky2 D8 + ky3 D.12) ten kx! (Kyq D |. + Ky |. Ds ten ky2 D3 + ky3 D13) + kx2 (ky0 D2 + kyi D "ky2 D o + ky3 D |!. 4> +" 3 (kyo D3 + ky | D7 + ky2 Dii + 'ky3 D15) after the last UL data output buffer circuit 31 calculates an open amount calculation processing is ended current UL at the beginning of coordinates and the difference between the beginning of the next UL coordinates (see FIG. 50), to open an unnecessary buffer stores data, the amount of opening of the output buffer to the free capacity of the buffer monitoring circuit 29. However, it is preferred that the cross opening of the distortion amount of the buffer at the center shown in Figure 49, the opening also contemplated postULBl value change amount and postULB2 As shown in FIG. 48 and postULB preULB for forming the first pixel coordinate position UL in pixels, for front and rear sides in the row direction is provided in the region of a predetermined width, each of which is defined and preULB postULB. crossing distortion at the center, not the amount of buffer normally open beginning UL coordinate difference, by referring to the change in value of the reference postULB (ULB: unit line buffer short, line buffer unit) to adjust the amount of change slow 器开放量(调整量是postULBl — postULB2,参照图49)。当(UL开头坐标的差)+ (调整量)为负的情况下,存储负值。在进行处理并且缓冲器开放量超过负值之前,不进行缓冲器的开放。缓冲器开放量计算电路31计算前述缓冲器开放量,并且从前级电路向数据可否发送判定电路30发送下一UL处理还需要多少数据。当通过前述缓冲器开放量的计算发现缓冲器有空闲时,缓冲器空闲容量监视电路29向前述前级电路进行请求。数据可否发送判定电路30根据来自写入地址生成电路28的内部计数器(BLC)和缓冲器开放量计算电路31的输入以及preULB值,判断是否可发送下一UL数据。数据可否发送判定电路30针对后级电路的请求返回接受请求(GRANT—N)。当被输入到读出地址生成电路25中的坐标超出块行(BL)的左端(参照图43)或右端,或者超出块行(BL)的上下端(参照图44),或者 An open amount (adjustment amount postULBl - postULB2, with reference to FIG. 49). When the (head difference coordinate UL) + (adjustment amount) is negative, the processing is performed and the negative buffer storage opening amount exceeds a negative value. prior, not open buffer buffer opening amount calculating circuit 31 calculates the opening amount of the buffer, the previous stage circuit and the data circuit 30 determines whether transmission of the next UL transmission data processing much needed. when the buffer through the open calculating the amount of free time found buffers, monitoring the free capacity of the buffer circuit 29 makes a request to the pre-stage circuit data transmission determines whether the internal counter circuit (BLC) 28 and the amount of opening of the buffer circuit 30 from the write address generated in accordance with preULB value calculation circuit 31 inputs and determines whether to send the next UL data transmission data decision circuit 30 returns whether to accept the request (GRANT-N) a request for a subsequent circuit. when the input to the read address generation circuit 25 the row coordinate is outside the block (BL) of the left (see FIG. 43) or right, or beyond the block lines (BL) of the upper and lower ends (see FIG. 44), or 真量超出preULB和postl几B的设定值(参照图45)的情况下,错误检测电路32输出错误。这些错误输出通过判别坐标来输出。在图43 和图44的情况下,输出错误,但处理继续进行。在图45的情况下,也输出错误,但处理继续进行。 The case where the amount exceeds the set value true postl preULB and several B (see FIG. 45), the error detection circuit 32 outputs an error. These error output by the discriminating output coordinates. In the case of FIGS. 43 and 44, the output error, However, processing continues. in the case of FIG. 45, and also outputs an error, but the process continues. '在图45的preULB和postULB的情况下, 不仅输出错误,而且存储超出量,并设定在寄存器内,从而可作为数据由CPU4取得。 'In the case of FIG. 45 and postULB preULB, not only an error output, and storage overrun, and is set in the register, so that data can be acquired by a CPU4. 这样,在块行处理中,在相对于输出侧的图像对输入侧的图像进行内插计算时,有时从输入图像范围溢出。 Thus, in the block line processing, when the image with respect to the output side of the input side is an image interpolation calculation, sometimes overflows from the input range image. 在该情况下,在输入范围内没有数据的部分生成内插数据,从错误检测电路32输出表示不能进行内插的错误。 In this case, no data is generated within the portion of the input range interpolated data, the interpolation error can not be expressed from the error detection circuit 32 outputs. 如图46所示,图像失真相对于失真中心坐标是对称的。 As shown in FIG. 46, for distortion of the image distortion is symmetrical center coordinates. 并且,失真一般越是接近中心就越小,越是远离中心就越大。 And distortion generally the closer the smaller centers, the greater the farther from the center. 因此,通常,如果考虑BL两端的UL变形来确保缓冲器,则在所有UL中,在内插处理中数据不会不足。 Therefore, in general, in consideration of deformation of the ends of the UL buffers to ensure BL, then all the UL, the interpolation processing data is not insufficient. 因此,相对于进行内插的最初坐标(由图中块行BL内的图像数据的最上面的X符号表示的开头坐标),在前方(图示右侧)设置规34定的区域preULB,在后方(图示左侧)设置规定的区域postULB。 Thus, the initial coordinates relative to the interpolation (line beginning from the figure the uppermost block X at the image data within the coordinates indicated BL), the front (right side shown) provided 34 gauge given area preULB, in rear (left side shown) disposed in a predetermined region postULB. 支持功能相当于图1的"失真校正范围计算部"的部分。 Support portion corresponding to the "distortion correction range calculation unit" in FIG. 在图12中, 关注区域为BL,然而可通过失真校正的坐标变换如图13所示那样变形。 In FIG. 12, the region of interest as BL, however, may be transformed as shown by the distortion correction coordinates deformed as shown in Fig. 此时,根据X, TL、 Y, TL、 XLmin、 XL鍾决定preULBl/postULBl (参照图48)。 At this time, X, TL, Y, TL, XLmin, XL clock decision preULBl / postULBl (see FIG. 48). 同样,根据X, TR、 Y' TR、 XRmin、 Xr^x决定preULB2/postULB2。 Also, according to X, TR, Y 'TR, XRmin, Xr ^ x decision preULB2 / postULB2. 而且,在UL处理中,当通过失真中心时,光学失真的弯曲方向为反方向。 Further, the processing in the UL, when distorted by the center, the bending direction of the optical distortion in the reverse direction. 这样,如图46所示,在失真中心左右,preULB和postULB的值不同。 Thus, as shown in Figure 46, the distortion around the center, and postULB of different values ​​preULB. 为了使该值左右都能对应,如图47所示,在右侧采用最大的postULB,在左侧采用最大的preULB,这样,尽管在左侧或右侧的一侧的处理时不需要,但也必须大量(即,扩大缓冲器)确保另一侧的数据, 使缓冲器浪费。 In order to make this value can correspond to the left and right, as shown in FIG 47 using the largest right postULB, using the maximum preULB the left, so, although not necessarily at the left or right side of the process, but It must also be large (i.e., expansion buffer) on the other side to ensure that data, so the buffer waste. 因此,preULB和postULB不是预先使用相同大的值来决定双方,而是把preULB和postULB作为变量来分别输入值。 Therefore, preULB and postULB not use the same pre-determined value larger parties, but to preULB and postULB as variables to input values. 当经过了失真中心坐标时,改变这些值来改变确保量。 When after a distortion center coordinates, change these values ​​to ensure that the amount of change. 即,通过在失真中心前后使preULB和postULB变化,可以在内部缓冲器的使用上消除浪费,可使用小缓冲器进行比较大的失真校正。 That is, by preULB and postULB change before and after the center of distortion, can eliminate waste in the use of an internal buffer, the buffer can be used for a relatively large small distortion correction. 把pre和postULB设定在寄存器85内,然而为了生成这些值,需要考虑失真变形可算出图像数据的输入范围的支持功能。 The pre and postULB set in the register 85, however, in order to generate these values, the distortion deformation can be calculated to consider the range of the input image data support. 该支持功能可以通过附加给失真校正处理部8的失真校正处理功能等来进行设置。 The support may be set by the distortion correction processing and the like attached to the distortion correction processing unit 8. 存储器控制部83进行如下的控制:使前述区域preULB和postULB在1UL处理中不被其他处理所改写。 The memory control unit 83 performs control: making the region preULB postULB not be overwritten, and other processing 1UL process. 前述区域preULB和postULB的设定可以从CPU 4设定到寄存器85中(参照图2),也可以使用CPU自动进行计算来设定。 And postULB preULB aforementioned region can be set from the CPU is set to the register 85. 4 (see FIG. 2), may also be used to set the CPU automatically calculated. 前面对在失真校正处理中进行16点内插作了描述,然而如图48所示,通过失真校正可使内插位置在lUL处理期间移动a (其中,a〉0)。 Been described in the foregoing interpolation 16:00 in the distortion correction process, however, shown in Figure 48, the distortion correction can be moved during the interpolation position lUL process a (wherein, a> 0). 由于进行16点内插的关系,作为输入图像范围,pre侧和post侧全都需要把内插所需要的像素的间隔&和W与上述间隔a相加后所得的范围。 As a result of the relationship between 16-point interpolation, as an input image range, pre and post side to side all require obtained after range desired pixel spacing and W & adding the above-described interpolation interval a. 当越过失真中心时,通过改变pre和postULB的值,把缓冲器縮小到必要最小限度。 When distortion across the center, by changing the value of pre and postULB, the buffer is reduced to a bare minimum. 并且,在前述的"失真校正计算部"中可以考虑内插量来输出结果。 Further, in the above-described "distortion correction calculation unit" may be considered to output the interpolated amount. 失真校正处理部8具有的内部存储部(内部缓冲器)82在图39的情况下,最大为在纵向上是96个像素,在横向上是16行(像素)。 Distortion correction processing section 8 has an internal storage unit (internal buffer) 82 in the case of FIG. 39, in the longitudinal direction is a maximum of 96 pixels, 16 rows in the transverse direction (pixels). 使用该内部存储部82进行失真校正。 The internal storage unit 82 using the distortion correction. 块行处理的数据按照至少每1UL向图示右方向进行扫描。 Data processing block line is scanned in the illustrated right direction at least every 1UL. 由于有时几UL —起开放,因而并不一定按照每1UL 进行扫描。 Because sometimes a few UL - from the opening, and thus not necessarily be scanned at every 1UL. 内部存储部82由于最初从空状态开始,因而输入16行来进行失真校正处理。 The internal storage unit 82 as the initial start with an empty state, and therefore the input line 16 performs distortion correction processing. 当处理这些单位行时,左侧行产生不需要的数据(1至几UL的数据)。 When these processing units row, left for unwanted data (UL data 1 to several). 另外,在处理后不一定不需要,当放大率大时,有时也需要几UL,直到不需要时为止。 In addition, after the treatment is not required necessarily, when a large magnification, sometimes take a few UL, until when not needed. 对于不需要的数据,开放缓冲器(即, 容许改写),以输入新数据。 For unnecessary data buffer open (i.e., allowing rewriting), to enter new data. 进行扫描的图像偏移到右侧。 Image scanning shifted to the right. 由于只有最大16行读入尺寸,因而开放不需要的数据区域,依次改写新数据。 Since only the maximum size read line 16, and thus does not need to open a data area, the new data are sequentially rewritten. 不需要的缓冲区域全部一起开放。 Buffer area need not all open together. 该开放量也许是1行,也许是5行。 The amount may be open one line, maybe five rows. 例如, 最初的3行的数据进入缓冲器,对以后处理是不需要的,因而进行开放, 接受下一数据,并改写下一数据。 For example, the first three lines of data into the buffer, for later processing is not required, and thus an open, accept the next data, and the next data is rewritten. 另外,如果不决定下一UL开头坐标就不知道在1UL处理结束后可开放的缓冲量(参照图50)。 Further, if the beginning of the next decision does not know the coordinates UL buffer amount after the end of treatment may 1UL open (see FIG. 50). 由于内部缓冲器小,因而为了尽早开放存储有未使用的数据的部分,计算下一UL开头坐标来取得开放量,通过开放缓冲器,可有效利用内部缓冲器。 Due to the small internal buffer, thereby to open as soon as the unused portion stored data, calculates the coordinates of the beginning of the next UL acquires the amount of opening by opening the buffer, the internal buffer can be effectively utilized. 然而,针对失真校正从后级电路块发来请求,返回针对该请求的同意,把1UL的数据输出到后级,并且在下一请求发来后返回同意,之后求出坐标,在按照上述进行动作时,缓冲器的开放延迟,与此同时,新数据的取入也延迟,结果,会发生流水线动作出现长的空闲。 However, for the distortion correction request sent from the post-stage circuit block, returns agreed for the request, the output data to the subsequent stage 1UL, and the next request sent agree to return, then finds the coordinates, as described above operates in the when open buffer delay, at the same time, loading new data is also delayed as a result, the operation appears pipeline long idle occur. 因此,如图51所示在进行先前失真校正处理(图示的坐标l〜n的处理)的阶段,与来自后级电路块的请求没有关系,当坐标n的处理结束时,运算下一开头坐标n+l来求出开放量,提前开放缓冲器。 Therefore, as shown in FIG. 51 performed earlier stage of the distortion correction processing (processing coordinate l~n illustrated) is not related to the request from the rear-stage circuit block, when the processing of the coordinates n, calculating the beginning of the next n + l coordinates to determine the amount of open, open buffer in advance. 这样,可从前级电路块较早地进行新数据输入,可填补流水线的空闲。 Thus, the front stage circuit block is a new data input earlier, to fill the pipeline may be idle. 然而,在图51的情况下,在进行了坐标n的处理后,计算像素n 十1的坐标,之后结束处理,因而如图52所示,尽管在缓冲器内下一UL 处理所需要的数据齐备,然而当来自后级的请求连续时,流水线发生空闲。 After However, in the case of FIG. 51, a process performed in the n coordinates, calculates the coordinates of the pixel n + 1, and thereafter terminates the process. Therefore, as shown in FIG. 52, although the data in the buffer required for processing the next UL available, however, when a request from the successive stage, the occurrence of an idle pipeline. 由于流水线的空闲依然存在,因而当请求连续时,有可能对装置整36体的动作速度产生不良影响。 Since the pipeline remains free, so that when the continuous request, it may adversely affect the operation speed of the entire apparatus body 36. 因此,作为图51的改善对策,如图53所示,具有尽早决定下一坐标来求出缓冲器开放量的方法。 Thus, a countermeasure to improve FIG. 51, 53, having the method determines the next coordinate is obtained as soon as opening amount of the buffer. 如图51所示,在比最后求出下一开头坐标更早的阶段取得下一UL开头坐标,当判断为必要数据齐备时,通过连续进行下一UL处理,来减小处理空闲。 51, the start of the next UL achieved than the last calculated coordinates in the coordinate beginning of the next earlier stage, when it is determined necessary data available, the next UL by continuous process to reduce the processing idle. 即,在1UL处理结束前把UL 的数据保持在寄存器等内,通过计算下一UL的开头坐标位置,预先求出下一UL处理的开放量。 That is, before the process ends 1UL UL data held in the register or the like, by calculating the coordinate position at the beginning of the next UL, the amount of opening of the next UL processing determined in advance. 在图53中,针对先前失真校正处理(图示的坐标l〜n的处理)的第2像素求出下一UL的开头坐标,在较早阶段取得缓冲器开放量。 In Figure 53, the distortion correction processing for the previous (coordinate l~n process shown) of the second pixel coordinates determined at the beginning of the next UL acquires the amount of buffer open at an earlier stage. 这样,可使缓冲器开放提前,如图54所示,与图52相比可大量减少流水线空闲。 Thus, in advance of the damper is open, as shown in FIG. 52 can greatly reduce the pipeline 54 compared to idle. 然而,具有电路复杂、控制也困难的问题。 However, a complicated circuit, control is difficult. 因此,图55对图53进行了进一步的改善。 Therefore, 55 in FIG. 53 has been further improved. 在图55中,在当前处理中的UL结束时开放的量(开放量l)在上一UL处理中是已知的。 In Figure 55, the open end of the current process in UL amount (amount of opening l) are known in a UL process. 在算出坐标n后,计算2个后续UL开头坐标,预先求出在下一UL处理结束时可开放的量(开放量2)。 After the calculated coordinates n, UL at the beginning of the subsequent calculation of the coordinates 2, an amount (amount of opening 2) at the open end of the next available UL processing determined in advance. 这样,没有必要如图53的情况那样在UL处理中进行例外的坐标生成(坐标2的生成)。 A coordinate, FIG. 53 there is no need for exceptional cases such as UL in the generation process (coordinate generation. 2). 这样,如果求出下下2个后续UL的开头坐标而不是下一UL幵头坐标,则不进行图53所示的例外处理,因而电路简单。 Thus, if the beginning of two subsequent UL determined at the next UL Jian coordinates instead of coordinates of the probe, the exception process is not shown in FIG. 53, and thus the circuit is simple. 如以上所述,根据本发明的第3实施方式,不使总线的传送量和存储器的容量大幅增大,即可实现失真校正。 As described above, according to the third embodiment of the present invention, without the transfer amount and the memory capacity of the bus is increased significantly, the distortion correction can be realized. 即,可使用小的缓冲容量进行比较大的失真校正处理。 That is, using a small buffer capacity for a relatively large distortion correction processing. 产业上的可利用性本发明可广泛利用于在数字照相机等的电子摄像装置中使用的图像处理装置、图像处理方法以及失真校正方法,不使电路规模和数据传送量增大,即可实现失真校正功能,并且在进行失真校正处理的情况下, 可计算空间位置,可进行有效利用作为源数据的摄影图像数据的失真校正处理,可针对枕形失真、桶形失真以及曲弦失真等的基于畸变像差的失真进行有效的失真校正处理。 APPLICABILITY The present invention can be widely used in industrial image processing apparatus used in the electronic imaging device of a digital camera, an image processing method and a distortion correction method of circuit size and data transfer amount is increased, distortion can be realized the case where the correction function, and performing distortion correction processing, the spatial position may be calculated, can be efficiently performed using the distortion correction image data as the imaging process source data, can be for the pincushion distortion, barrel distortion and curvature based on string distortion distortion aberration distortion effective distortion correction processing. 37 37

Claims (11)

1.一种图像处理装置,具有失真校正单元,其特征在于, 该图像处理装置还具有:失真校正范围计算部,其计算由前述失真校正单元进行失真校正处理的输入图像范围, 前述失真校正范围计算部具有: 坐标生成部,其生成内插坐标; 失真校正坐标变换部,其输出把规定的失真校正式应用于前述所生成的内插坐标而变换的坐标;以及校正范围检测部,其根据前述所变换的坐标位置计算前述输入图像范围。 An image processing apparatus includes a distortion correcting unit, wherein the image processing apparatus further comprising: a distortion correction range calculation unit that calculates a distortion correction range of the input image processed by the distortion correction means, the distortion correction range calculation unit comprises: coordinate generation unit, which generates the interpolation coordinates; a distortion correcting coordinate-transformation unit, outputs the correction formula is applied a predetermined distortion coordinate transformation of the coordinates of the interpolation generated; and a correction range detecting unit, based on converting the coordinate position of the range of the input image is calculated.
2. 根据权利要求l所述的图像处理装置,其特征在于,前述坐标生成部仅使用与失真校正后的输出图像范围内的边缘部的各像素对应的坐标来生成内插坐标。 The image processing apparatus according to claim l, wherein the generation unit using only the coordinates corresponding to the coordinates of each pixel of the output image edge portion within the range of the distortion correction coordinate generating interpolation.
3. 根据权利要求2所述的图像处理装置,其特征在于,前述输出图像范围是矩形,前述坐标生成部仅使用与该输出图像范围的4边的各像素对应的坐标来生成内插坐标。 The image processing apparatus according to claim 2, wherein the range of the output image is a rectangle, the coordinates generating unit to generate only the coordinates corresponding to four sides of each pixel of the output image area the interpolation coordinates.
4. 根据权利要求3所述的图像处理装置,其特征在于,前述失真校正范围计算部针对通过坐标变换所生成的坐标,根据与前述输出图像范围的4边的各边对应的像素的坐标的最大值和最小值以及与前述输出图像范围的4顶点对应的坐标中的至少任意一项,来计算前述输入图像范围。 The image processing apparatus according to claim 3, wherein the distortion correction range calculation unit for coordinate generated by the coordinate transformation from the coordinate with each side of four sides of the output image area corresponding to the pixels 4 corresponding to the coordinates of the vertices and a maximum and minimum range of the output image in at least any one of, the range of the input image is calculated.
5. 根据权利要求l所述的图像处理装置,其特征在于,前述失真校正坐标变换部对前述规定的失真校正式内所包含的运算进行时序处理。 The image processing apparatus according to claim l, wherein the distortion correction calculating the distortion correction coordinate conversion formal portion of the predetermined timing included in the process.
6. 根据权利要求l所述的图像处理装置,其特征在于,前述坐标生成部对前述失真校正处理用的内插坐标进行规定的疏化,来求出内插坐标。 The image processing apparatus according to claim l, wherein the coordinate generation unit for processing the distortion correction coordinates using interpolation predetermined thinning, interpolation to determine the coordinates.
7. 根据权利要求l所述的图像处理装置,其特征在于,前述失真校正范围计算部针对进行失真校正处理的多个输入信号依次重复范围计算,来计算前述输入图像范围。 The image processing apparatus according to claim l, wherein the distortion correction range calculation unit sequentially repeated for a plurality of input signals distortion correction range calculation processing to calculate the range of the input image.
8. 根据权利要求7所述的图像处理装置,其特征在于,通过重复进行前述范围计算,决定失真校正后的图像范围相对于输入图像范围在规定范围内的校正倍率M。 The image processing apparatus according to claim 7, characterized in that, calculated by repeating the aforementioned range, the scope of the decided distortion correction image magnification is corrected in a predetermined range with respect to the input image range M.
9. 根据权利要求l所述的图像处理装置,其特征在于,前述失真校正范围计算部在前述失真校正单元执行的失真校正中,计算进行下一失真校正的输入图像范围。 The image processing apparatus according to claim l, wherein the distortion correction range calculation unit in the distortion correction unit performs the distortion correction, distortion correction calculation of the next input image range.
10. —种进行失真校正处理的图像处理方法,在进行前述失真校正处理时,计算进行失真校正处理的输入图像范围,其特征在于,包括以下步骤:生成内插坐标;输出把规定的失真校正式应用于所生成的内插坐标而变换的坐标;以及根据所变换的坐标位置计算前述输入图像范围。 10. - The image processing method of the kind of distortion correction process, when the distortion correction process is performed, the range of the input image calculated distortion correction processing, the method comprising the steps of: generating the interpolation coordinates; the predetermined distortion correction output formally applied to coordinate transformation of the coordinates of interpolation generated; and calculating the range of the input image according to the converted coordinate position.
11. 根据权利要求10所述的图像处理方法,其特征在于,在前述失真校正处理的执行中,计算下一次进行失真校正的输入图像范围。 The image processing method according to claim 10, wherein, in the distortion correction process is performed, the next time is calculated distortion correction range of the input image.
CN 200480021876 2003-07-28 2004-07-27 Image processing apparatus, image processing method, and distortion correcting method CN100552712C (en)

Priority Applications (4)

Application Number Priority Date Filing Date Title
JP202664/2003 2003-07-28
JP202493/2003 2003-07-28
JP202663/2003 2003-07-28
JP2003202493A JP4772281B2 (en) 2003-07-28 2003-07-28 Image processing apparatus and image processing method

Publications (2)

Publication Number Publication Date
CN1830002A CN1830002A (en) 2006-09-06
CN100552712C true CN100552712C (en) 2009-10-21

Family

ID=34262194

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 200480021876 CN100552712C (en) 2003-07-28 2004-07-27 Image processing apparatus, image processing method, and distortion correcting method

Country Status (2)

Country Link
JP (1) JP4772281B2 (en)
CN (1) CN100552712C (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102316262A (en) * 2010-06-01 2012-01-11 精工爱普生株式会社 Display control device

Families Citing this family (27)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10176559B2 (en) 2001-12-22 2019-01-08 Lenovo (Beijing) Co., Ltd. Image processing method applied to an electronic device with an image acquiring unit and electronic device
JP4832878B2 (en) * 2005-12-09 2011-12-07 倉敷紡績株式会社 Image synthesis processing apparatus and method
WO2009013845A1 (en) 2007-07-20 2009-01-29 Techwell Japan K.K. Image processing device and camera system
JP4940050B2 (en) * 2007-08-09 2012-05-30 キヤノン株式会社 Image processing method for performing a distortion correction on the image data, programs, and recording medium
KR101428054B1 (en) 2007-12-13 2014-08-07 주식회사 엘지씨엔에스 Apparatus and method for media image detection, and system with the same
JP5195288B2 (en) * 2008-10-30 2013-05-08 パナソニック株式会社 The camera body, the camera system
CN101771796B (en) 2008-12-29 2013-01-30 比亚迪股份有限公司 Method for correcting barrel distortion image and device thereof
JP5218071B2 (en) 2009-01-07 2013-06-26 ソニー株式会社 Image processing apparatus, image processing method, and program
JP2010176547A (en) * 2009-01-30 2010-08-12 Dainippon Printing Co Ltd Controller included in image processor, control method and control processing program
JP5443844B2 (en) 2009-06-17 2014-03-19 オリンパス株式会社 The image processing apparatus and an imaging apparatus
JP5489871B2 (en) 2009-06-24 2014-05-14 オリンパス株式会社 Image processing apparatus
JP5593060B2 (en) * 2009-11-26 2014-09-17 株式会社メガチップス Method of operating an image processing apparatus, and image processing apparatus
JP5316811B2 (en) 2010-01-18 2013-10-16 株式会社リコー The image processing method and apparatus, and an image capturing apparatus
US8116587B2 (en) * 2010-02-16 2012-02-14 Ricoh Co., Ltd. Method and apparatus for high-speed and low-complexity piecewise geometric transformation of signals
JP5829094B2 (en) 2011-10-20 2015-12-09 オリンパス株式会社 Image processing apparatus and image processing method
TWI489859B (en) * 2011-11-01 2015-06-21 Inst Information Industry Image warping method and computer program product thereof
CN102490496B (en) * 2011-12-07 2014-02-26 东莞市盛雄激光设备有限公司 Laser marking barrel pincushion distortion compensation value determining method and device as well as laser marking method and device
JP5893445B2 (en) * 2012-03-07 2016-03-23 株式会社メガチップス Method of operating an image processing apparatus, and image processing apparatus
JP6146819B2 (en) * 2012-03-14 2017-06-14 富士機械製造株式会社 Image correction processing apparatus
JP6124720B2 (en) * 2013-07-22 2017-05-10 オリンパス株式会社 Imaging device, image processing method and image processing program
JP2016025443A (en) * 2014-07-18 2016-02-08 キヤノン株式会社 Image processing apparatus, imaging apparatus, image processing method, and program
CN104270571B (en) * 2014-10-20 2018-06-01 联想(北京)有限公司 The image processing method and an electronic apparatus
EP3035664B1 (en) 2014-12-19 2017-06-21 Axis AB Method for processing a video stream
CN104616256B (en) * 2015-01-15 2018-01-09 桂林长海发展有限责任公司 Method and apparatus for correcting barrel distortion of the image
JP6143199B2 (en) * 2015-03-18 2017-06-07 カシオ計算機株式会社 Image correction apparatus, an image correction method, and a program
CN105141827B (en) * 2015-06-30 2017-04-26 广东欧珀移动通信有限公司 The method of correcting a deformed and a terminal
JP2017175498A (en) * 2016-03-25 2017-09-28 日立オートモティブシステムズ株式会社 Image processing apparatus and image processing method

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1308465A (en) 1999-12-18 2001-08-15 Lg电子株式会社 Image distortion correcting device and method and image display
CN1386361A (en) 2000-07-25 2002-12-18 松下电器产业株式会社 Image distortion correcting device and image distortion correcting method

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH06119442A (en) * 1992-10-07 1994-04-28 Meidensha Corp Distortion correcting method for graphic
JPH06153065A (en) * 1992-11-11 1994-05-31 Olympus Optical Co Ltd Image pickup device
JP3035416B2 (en) * 1992-11-18 2000-04-24 キヤノン株式会社 Imaging device and an image reproducing apparatus and a video system
JP3297511B2 (en) * 1993-10-15 2002-07-02 オリンパス光学工業株式会社 Video processing device
JP3686695B2 (en) * 1994-10-20 2005-08-24 オリンパス株式会社 Image processing apparatus
JPH0998340A (en) * 1995-09-29 1997-04-08 Sanyo Electric Co Ltd Image pickup device
JPH11243508A (en) * 1998-02-25 1999-09-07 Matsushita Electric Ind Co Ltd Image display device
JPH11250239A (en) * 1998-02-27 1999-09-17 Kyocera Corp Digital image pickup device for operating distortion correction by yuv data
JPH11275391A (en) * 1998-03-20 1999-10-08 Kyocera Corp Digital image pickup device for selecting distortion correction
JP2000132673A (en) * 1998-10-28 2000-05-12 Sharp Corp Image system
JP4179701B2 (en) * 1999-04-28 2008-11-12 オリンパス株式会社 Image processing apparatus
JP2001266176A (en) * 2000-01-14 2001-09-28 Sony Corp Picture processor, picture processing method and recording medium

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1308465A (en) 1999-12-18 2001-08-15 Lg电子株式会社 Image distortion correcting device and method and image display
CN1386361A (en) 2000-07-25 2002-12-18 松下电器产业株式会社 Image distortion correcting device and image distortion correcting method

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102316262A (en) * 2010-06-01 2012-01-11 精工爱普生株式会社 Display control device
CN102316262B (en) * 2010-06-01 2016-04-27 精工爱普生株式会社 Display control means

Also Published As

Publication number Publication date
JP4772281B2 (en) 2011-09-14
JP2005044098A (en) 2005-02-17
CN1830002A (en) 2006-09-06

Similar Documents

Publication Publication Date Title
JP3934151B2 (en) Image generating device and image generating method
US20120162475A1 (en) Method and apparatus for raster output of rotated interpolated pixels optimized for digital image stabilization
US7457478B2 (en) Image processing apparatus, image processing system, and image processing method
KR101350193B1 (en) Taken-image signal-distortion compensation method, taken-image signal-distortion compensation apparatus, image taking method and image-taking apparatus
US7418155B2 (en) Image processing method, image processing apparatus and image pickup apparatus and display apparatus suitable for the application of image processing method
US20020028028A1 (en) Image processing apparatus, image processing method, and storage medium
US7881563B2 (en) Distortion correction of images using hybrid interpolation technique
US6671418B2 (en) Image processing apparatus and method which determines a block size in accordance with a filter structure
Gribbon et al. A novel approach to real-time bilinear interpolation
CN100379246C (en) Apparatus and method for processing an image
EP1657675B1 (en) Image processing device
CN103098077B (en) Real-time video frame pre-processing hardware
US6597394B1 (en) Programmable image transform processor for digital image processing
US9262807B2 (en) Method and system for correcting a distorted input image
JP4185468B2 (en) Correcting method for an exposure apparatus and a photographing image
JPH11122626A (en) Image processing method, system and record medium recording image processing program
US7587524B2 (en) Camera interface and method using DMA unit to flip or rotate a digital image
CA1260841A (en) Image reading and/or recording apparatus with enlarging and reducing function
JP2001160138A (en) Method and device for image processing
JP4151793B2 (en) High resolution method of an image pickup apparatus and an image
JP2001160139A (en) Method and device for image processing
US7593569B2 (en) Pixel defect correction device
JP2006217641A (en) Method for correcting distance between sensors in image reading apparatus
US6961084B1 (en) Programmable image transform processor
US20040105015A1 (en) Image processing device and image processing program

Legal Events

Date Code Title Description
C06 Publication
C10 Request of examination as to substance
C14 Granted