Still image producing method and still image capture system

[0002] The present invention relates generally to a still image producing method and a still image capture system. More specifically, the invention relates to a still image producing method and a still image capture system, which can capture a plurality of frames images, which are slightly shifted from each other, by utilizing unavoidable vibration, such as the movement of the hands, which usually occurs on ordinary capture conditions, to synthesize the images by a predetermined method to capture a still image of a frame having a high resolution and a low noise, when the still image is captured by means of a sensor, such as a CCD.

[0003] With the rapid development of information processing systems, such as personal computers and PDAs (personal digital assistant), and digital communication networks for connecting the information processing systems, the demands for improved techniques for capturing a high-definition image information have been increased. As systems capable of capturing a still image as digital data, digital cameras have come into wide use recently. Therefore, as an example of a technique for capturing still images, a digital camera will be described below.

[0004] FIG. 6 is a schematic block diagram of a conventional digital camera. An image to be captured is imaged on an image sensor by means of an optical system. As the optical system, a lens and a stop are used. In addition, as the image sensor, a CCD or a CMOS is used. The image sensor converts an inputted optical information to an electric information, which is outputted as a time series signal. After carrying out the sampling and gain control of the electric information outputted from the sensor, the electric information is converted into digital data by means of an A/D converter. Then, a signal processing for dividing the electric information into a luminance signal and a color signal is carried out by means of a signal processing circuit, and the results thereof are stored in a still image memory. The still image information thus stored is outputted to the outside via an output part, if necessary.

[0005] Currently, most of digital cameras use a CCD area sensor as the image sensor. The CCD area sensor comprises photodiodes arranged in the form of a two-dimensional matrix, and CCD analog shift registers. When the photodiodes receive light, the photodiodes produce electrons in accordance with the quantity of the received light to accumulate the electrons therein. The analog shift registers transfer the accumulated electrons to output the electrons as time series signals.

[0006] The digital camera is designed to capture a still image, and the still image information outputted from a conventional digital camera is an image information of a single frame. That is, assuming that the exposure time of a sensor is T, an electric information corresponding to the quantity of light received by the respective photodiodes of the sensor between time t=0 and t=T is outputted from the sensor only once in time series, and the above described signal processing of the outputted electric information is carried out, so that the digital camera outputs a still image information.

[0007] However, in the main current of digital camera, an optical system having a small diameter of ¼ inch type format is adopted in order to meet the demand that the weight and size of the digital camera be reduced. In this case, the diagonal length of the image formed on the image sensor is only about 4.5 mm. On the other hand, CCD sensors having pixels of about 500×800 are adopted to meet the demand of the resolution. That is, photodiodes of about 500×800 are arranged on a CCD sensor. Consequently, the size of one photodiode is as small as 4.5 μm×4.5 μm. Therefore, the number of electrons, which can be produced by the photodiode, is only about 6000 when the exposure time is {fraction (1/60)} seconds at an illuminance of 10 luxes, which is ordinary brightness. It is known that a random noise of n^½ occurs when n electrons are produced since the electron producing process is a stochastic process. That is, when n electrons are produced, the S/N ratio is n/n^½=n^½ so that the S/N ratio decreases as the number of electrons decreases.

[0008] With the spread of digital cameras, the enhancement of resolution of digital cameras has been required. In order to take a still image of a high resolution which is equal to a silver halide photograph, the number of pixels must be further increased. However, if the size of the photodiode is reduced in order to increase the number of pixels, the number of produced electrons decreases, so that the S/N ratio further decreases. On the other hand, in order to increase the number of pixels without reducing the size of the photodiode, it is required to enlarge the optical system, so that this does not meet the demand that the weight be reduced.

[0009] Thus, in conventional digital cameras, there is a trade-off relationship between the resolution and the S/N ratio due to the random noise, so that it is not possible to take a high definition still image having a high resolution and a low noise while meeting the demand that the weight be reduced.

[0010] On the other hand, there is a method using a so-called swing CCD, as a method for improving the resolution without increasing the number of pixels of the CCD. This vibrates the whole CCD at a smaller amplitude than the pixel pitch of the CCD by utilizing a vibrating mechanism, such as a piezoelectric element, to intentionally shift the pixels from each other to picks up an image in order to improves the resolution of the image.

[0011] However, it is important for the swing CCD to accurately synchronize the image capture timing with the vibration of the CCD. Therefore, the vibrating mechanism is complicated, and the assembly and adjustment thereof are not easy. Moreover, it is not easy to accurately reproduce ultra micro vibration which is a fraction of the pixel pitch, so that it is not easy to ensure the stability of the system over a long period of time.

[0012] It is therefore an object of the present invention to eliminate the aforementioned problems and to provide an image processing method and a still image capture system, which can take a high definition still image having a high resolution and a high S/N ratio by means of light and simple mechanisms.

[0013] According to the present invention, it is possible to take a high definition still image having a high resolution and a low noise by synthesizing frame images, which are slightly shifted from each other by the “incidental movement of an apparatus such as a camera” in taking a picture. It is conventionally conceived that such a “movement of a camera” is a bad and undesired factor to deteriorate images, so that the incidental movement of an apparatus such as a camera is conventionally actively removed. However, the inventor converted the conventional idea to make the present invention after analyzing, in detail, such a “movement of a camera”, which unavoidably occurs when taking a still image.

[0014] The present invention positively utilizes the “movement of a camera” unlike conventional systems, and is characterized by the improvement of the resolution and S/N ratio of a still image. Such a “movement of a camera” does not match with pixel pitch. Therefore, different points of an object are recorded on the respective pixels on each of frames sequentially shot. Then, if the respective frames are overlapped with each other so that the positions of the object on the frames are overlaid with each other, it is possible to obtain a still image having a greater number of pixels than that of each of the frames.

[0015] According to one aspect of the present invention, there is provided a still image producing method for producing a still image of an object by means of an image pickup element capable of capturing-an image as a plurality of pixels, the method comprising the steps of: taking pictures of the object while allowing the relative displacements between the object under substantially the same exposure time with each other and the image pickup element; producing a plurality of frame images being formed by a plurality of pixels so that different points of the object are recorded on the corresponding pixels of at least a predetermined number N (N is a natural number equal to or greater than 2) of frame images of the plurality of frame images; calculating relative displacement amounts between the positions of the object on the plurality of frame images; and overlapping the plurality of frame images after shifting the plurality of frame images from each other on the basis of the relative displacement amounts so that the positions of the object on the plurality of frame images are overlaid with each other, to produce a synthesized still image of the object as an image containing an amount of information, which is N times as large as that of each of the frame images.

[0016] According to another aspect of the present invention, a still image capture system comprises: an image pickup element for taking a picture of an object to capture an image as a plurality of pixels; and a signal processing system for calculating relative displacement amounts between positions of the object on a plurality of frame images, each of which is formed by a plurality of pixels shot by the image pickup element, under substantially the same exposure time with each other so that different points of the object are recorded on the corresponding pixels of a predetermined number N (N is a natural number equal to or greater than 2) of frame images of the plurality of frame images, the signal processing system overlapping the plurality of frame images after shifting the plurality of frame images from each other on the basis of the relative displacement amounts so that the positions of the object on the plurality of frame images are overlaid with each other, to produce a synthesized still image of the object as an image containing an amount of information, which is N times as large as that of each of the frame images.

[0017] With these constructions, the present invention has the following advantages.

[0018] First, according to the present invention, it is possible to capture a high definition still image having improved resolution and random noise while maintaining an optional system of a camera to be small and light.

[0019] According to the present invention, in order to obtain such a high definition still image, it is not required to use a so-called swing CCD having complicated mechanisms or an expensive CCD having a large number of pixels. Therefore, the costs can be reduced, and the mechanisms are simple and difficult to break down, so that it is possible to obtain superior reliability.

[0020] According to the present invention, it is possible to suitably adjust the balance between resolution and noise in accordance with the type of an object, the picture taking conditions, and the characteristics of an optical system and image sensor of a camera. In addition, the interpolation processing can be carried out any numbers of times as long as data are stored in a sequential still image memory. Therefore, the interpolation processing conditions can be modified using the feedback based on the image quality of the processed still image.

[0021] Moreover, according to the present invention, it is possible to capture a high definition still image by means of a simple hardware by positively utilizing the “movement of a camera”, which is conventionally conceived that it is bad and undesirable to take a still image. It is enough for this “movement of a camera” to correspond to an order of the picture pitch, so that it is not required to intentionally move the camera. For example, the vibration caused by the pushing operation of the shutter of the camera is enough. However, the present invention can similarly cope with the ordinary “movement of the hands”, which is far greater than the vibration caused by the pushing operation of the shutter. Therefore, according to the present invention, it is possible to capture a high definition still image without being careful in the “movement of the hands” in taking a picture. As a result, in comparison with conventional digital cameras, it is far easy to use a digital camera, and it is possible to obtain a higher image quality at low costs.

[0022] Thus, according to the present invention, it is possible to easily obtain a high definition still image, so that the invention has great industrial merits.

[0023] The present invention will be understood more fully from the detailed description given herebelow and from the accompanying drawings of the preferred embodiments of the invention. However, the drawings are not intended to imply limitation of the invention to a specific embodiment, but are for explanation and understanding only.

[0024] In the drawings:

[0025] FIG. 1A is a schematic perspective view of a digital camera according to the present invention;

[0026] FIG. 1B is a schematic sectional view of the digital camera of FIG. 1A;

[0027] FIG. 2 is a schematic block diagram showing the construction of a digital camera according to the present invention;

[0028] FIG. 3 is a schematic block diagram showing a processing algorithm in a sequential still image synthesizing/still image converting part according to the present invention;

[0029] FIG. 4A is a view showing an example of a two-dimensional distribution state of signals on an image obtained by combining frames;

[0030] FIG. 4B is a view showing an example of a desired distribution of signals as a still image data when being outputted from a digital camera to the outside;

[0031] FIGS. 5A through 5C are conceptualized views schematically showing an interpolation processing, wherein FIG. 5A is a conceptualized view showing the distribution of signals before an interpolation processing, FIG. 5B is a conceptualized view showing the state that the distribution of signals is uniform after the interpolation processing and that the resolution of a still image is doubled in comparison with the image of each of frames, and FIG. 5C is a conceptualized view showing the state of an improved S/N ratio after the interpolation processing; and

[0032] FIG. 6 is a schematic block diagram showing the construction of a conventional digital camera.

[0033] Referring now to the accompanying drawings, the preferred embodiments of the present invention will be described below.

[0034] FIG. 1A is a schematic perspective view of a digital camera as an example of a still image capture system according to the present invention, and FIG. 1B is a schematic sectional view showing the construction of a principal part of the digital camera of FIG. 1A. As shown in FIG. 1A, according to the present invention, a digital camera 10 comprises a body 12, a lens opening 14, a finder 16, a grip portion 18, a shutter switch 20 and an external output terminal 22. The lens opening 14 is provided with a lens 24. The finder 16 may be a simple optical finder. Alternatively, a display, such as a liquid crystal display, may be provided on the back face (not sown) of the camera so as to serve both as a finder and a monitor for output images.

[0035] As shown in FIG. 1B, an iris 28 is provided behind the lens 24 to adjust the quantity of light. The image of an object converged by the lens 24 and the stop 28 is formed on an image sensor 30 serving as an image pickup element. As the image sensor 30, a solid image pickup element, such as a CCD image sensor and a CMOS sensor, is often used. On a circuit board 32, there are provided a drive circuit for controlling the image sensor 30 and a circuit of a signal processing system for processing an output signal of the image sensor 30. To these electronic circuit parts and a power output part for the iris 28 and so forth, a driving power is supplied from a battery 34.

[0036] FIG. 2 is a schematic block diagram showing the construction of a digital camera according to the present invention. The digital camera of the present invention is designed to capture a still image similar to conventional cameras. However, when the digital camera according to the present invention captures a single still image, the image sensor 30 sequentially captures and outputs a plurality of sequential frames as if a movie camera is used. Similar to conventional digital cameras, the image data of the respective frames sequentially outputted from the image sensor 30 are stored in a memory 46 via a sampling/gain control part 40, an A/D converting part 42 and a signal processing part 44. However, in the case of conventional cameras, image data of one frame are stored in a memory corresponding to the memory 46. On the other hand, in the case of the camera of the present invention, image data of a plurality of frames are stored in the memory 46. Therefore, throughout the specification, data of a plurality of frames will be collectively referred to as a “sequential still image (dynamic image) information”, and the memory 46 for storing these data therein will be collectively referred to as a “sequential still image (dynamic image) memory”, in order to clearly distinguish the data and memory 46 of the present invention from those of conventional cameras.

[0037] As an example, assuming that the exposure time, i.e., the storage time, for each frame is 1/60 seconds and that the number of frames N is 100, the image capture process will be described. In this case, the image sensor 30 converts the optical information of each of frames, which are shot when time t is 0 second≦t≦1/60 seconds, 1/60 seconds≦t≦2/60 seconds, . . . , 99/60 seconds≦t≦100/60 seconds, respectively, to an electric information to sequentially output the resulting electric information. The total time required to shoot these 100 frames is 100/60 seconds, i.e., about 1.7 seconds. The exposure time, i.e., the storage time, of each of the frames may be different. However, if the storage times of the frames are substantially the same to maintain substantially the same luminance level, it is convenient for the synthesis of the frames, which will be carried out later.

[0038] When taking a picture, the camera is supported by the hands or fixed to a support, such as a tripod. In either case, the sequential still image information sequentially taken contain some “movement of a camera” as described above. It is conventionally conceived that such a “movement of a camera” is a bad and undesired factor, which deteriorates images, so that the “movement of a camera” is conventionally actively removed. However, one feature of the present invention is that the resolution and S/N ratio of a still image are improved by utilizing the “movement of a camera” unlike conventional cameras.

[0039] That is, as shown in FIG. 2, in the digital camera according to the present invention, the image data of a plurality of frames stored in the sequential still image memory 46 are synthesized to a single still image by means of a sequential still image synthesizing/still image converting part 48. If the image data of the respective frames containing the “movement of the camera” are thus a synthesized by means of the sequential still image synthesizing/still image 48 by a predetermined method, a high definition still image can be obtained. The synthesized still image data are stored in a still image memory 50, and outputted to the outside via the output part 22, if necessary.

[0040] FIG. 3 is a schematic block diagram showing an example of a processing algorithm in the sequential still image synthesizing/still image converting part 48 according to the present invention. The sequential still image information, i.e., the information of N frames, is divided into brightness information and color information to be stored in the sequential still image memory. Y signals are stored as the brightness information, and R signals (red), G signals (green) and B signals (blue) are stored as the color information. Assuming that the number of photodiodes of the CCD sensor is N_x×N_y, the Y, R, G and B signals forming the sequential still image information of N frames can be expressed as the following function of the corresponding coordinates n_x, n_y, and frame number n_t:

[0041] wherein 1≦n_x≦N_x, 1≦n_y≦N_y, and 1≦n_t≦N.

[0042] According to the present invention, as shown in FIG. 3, the contour extraction is carried out on the basis of, e.g., the Y signals, for each frame. That is, the contour of a pattern transcribed on each frame is recognized on the basis of the luminance signals. The reasons why the contour extraction is carried out on the basis of the Y signals are that the contour of an image is often most clearly indicated by the luminance signals, and that there is a problem in that the resolution deteriorates if the contour is extracted every color since each of the photodiodes of the CCD is distributed to K, G and B so that the number of pixels of each color is one third of the total number of pixels. However, the present invention should not be limited to the counter sampling carried out using only the Y signals.

[0043] For example, when an object is formed by only red (R) and green (G) and when the luminance of R approximates to that of G, it is advantageous to carry out the contour extraction on the basis of the R and G signals instead of the Y signals. In the case of a so-called three-plate type CCD camera or the like, the resolution does not deteriorate even if the contour is extracted on basis of any color signal. Therefore, the kinds of signals, which are used to carry out the contour extraction, can be suitably selected in accordance with the distribution state of colors of the object and the kind of the camera. This is the same with respect to a characteristic pattern extracting step and a displacement amount calculating step, which will be described later.

[0044] Then, as shown in FIG. 3, a characteristic pattern is extracted from the extracted contour. The characteristic pattern means a pattern which has a sharp contour, the coordinates of which can be clearly identified. For example, the characteristic patterns include an intersection of lines, a corner of a line, and a luminescent spot.

[0045] After extracting the characteristic pattern, the relative displacement amounts are derived for each frame. That is, even if the pictures of the same object are sequentially taken, if the “movement of the camera” occurs when taking the pictures, the positions of the object in the respective frames are different from each other.

[0046] For example, when K characteristic patterns in common between frames are extracted, the coordinates of these characteristic patterns can be expressed as follows:

[0047] (X(k, n_t), Y(k, n_t))

[0048] wherein k is an integer in the range of 1≦k≦K, which indicates a number of the characteristic pattern, and n_tis a frame number. In order to simplify explanation, it is assumed that the “movement of a camera” in taking a picture is only a parallel movement and does not include rotation and distortion of the object itself. Then, the displacement amount of the n_tth frame with respect to the first frame can be indicated by only the X and Y components on plane coordinates. The X and Y components of the relative displacement amount vector Δ(k, n_t) of the n_tth frame, which are calculated by the k th characteristic pattern of the extracted K characteristic patterns, can be expressed as follows.

[0049] X component of Δ(k, n_t): (X (k, n_t)−X (k, 1))

[0050] Y component of Δ(k, n_t): (Y (k, n_t)−Y (k, 1))

[0051] In order to enhance the reliability of the calculated values of the relative displacement amounts of the frames, it is desired to calculate the displacement amounts of all the characteristic patterns for each frame to average the calculated displacement amounts. That is, the relative displacement amount Δ(n_t) of the n_tth frame can be expressed by the undermentioned formula as an average value of the displacement amounts calculated from all the K characteristic patterns:

[0052] Δ(n_t)=ΣΔ(k, n_t)/K

[0053] wherein the addition expressed by the sum symbol Σ is carried out in the range of from k=1 to k=K. The displacement amount Δ(n_t) thus obtained is often a value which is out of the pixel pitch of the CCD, since it is an average value calculated by the K characteristic patterns. That is, the displacement amount of each characteristic pattern is spatially quantified by the respective pixels of the DDC to be often integer times as large as the pixel pitch. However, if the displacement amounts are averaged in accordance with the above formula, it is possible to derive a more accurate displacement amount which is out of integer times as large as the pixel pitch. In addition, if the displacement amount being out of the pixel pitch can be derived, the respective frames can be synthesized after shifting the pixels, so that the resolution can be improved.

[0054] Then, as shown in FIG. 3, the N frames are combined on the basis of the calculated relative displacement amounts. That is, the respective frames are overlapped with each other while being put back by the respective relative displacement amounts Δ(n_t), to synthesize a single still image. The combination of the respective frames is carried out with respect to the brightness information and the color information, respectively. Although the respective frames are object information which are taken at different times, the differences between times are disregarded since it is an object of the present invention to capture a still image.

[0055] In order to simplify explanation, it will be hereinafter supposed that all of the displacement amounts of the N frames are different and not zero. That is, it will be supposed that when the N frames are combined, the luminance signals Y of the N frames are not overlapped with each other and are not expressed by a multivalued function.

[0056] The distribution of the luminance signals Y (x, y) on the synthesized still image, which is thus obtained by combining the frames, can be expressed by the following formula:

[0057] Y(x, y)=Y(n_x, n_y, n_t)

[0058] (when x=X_pn_x−Δx(n_t) and y=Y_pn_y−Δy(n_t))

[0059] wherein X_pand Y_pare pixel pitches in X and Y directions on an image, respectively, the pixel pitches corresponding to the pitches of the photodiodes of the CCD, and n_xand n_yare numbers of pixels in x and y directions on the frame, respectively, and Δx(n_t) and Δy(n_t) being x and y components of the relative displacement amount Δ(n_t) of the n_tth frame, respectively.

[0060] As can be seen from the above formula, the luminance signals on the synthesized still image after combination only have brightness information at positions shifted from the respective pixels of the first frame by Δx(n_t) in the x direction and by Δy(n_t) in the y direction on each frame. That is, the number of pixels substantially increases. Such combination is also carried out with respect to the R, G and B signals serving as color information.

[0061] FIG. 4A is a view showing an example of a distribution state of signals on an image obtained by combining frames. FIG. 4A shows the case where the number of frames N=3, for simple explanation. In addition, the respective signal points are numbered by frame numbers for the sake of convenience. The displacement amounts of the respective frames caused by the “movement of the hands” and physical vibration when taking pictures are generally uniform. Therefore, as shown in FIG. 4A, the distribution state of signals after combination is often spatially uniform. Also in such a state, the number of pixels increases, so that the resolution of the still image is improved. Therefore, when information on the distribution of pixels as well as luminance signal data and color signal data are supplied, the obtained still image can be utilized as a high definition still image.

[0062] On the other hand, FIG. 4B is a view showing an example of a desired distribution of signals as still image data outputted from a digital camera to the outside. In FIG. 4B, the resolution of each frame shown in FIG. 4A is doubled. As a practical still image information, it is desired that signals are two-dimensionally uniformly distributed as shown in FIG. 4B. Therefore, the intervals between signals, which are ununiformly distributed as shown in FIG. 4A, are interpolated to synthesize a uniform distribution of signals as shown in FIG. 4B. In order to carry out this, it is enough for the number of synthesized frames to be m²(m is a natural number equal to or greater than 2). According to the present invention, it is possible to obtain a still image, which has a resolution multiplied by m, by carrying out the same processing after synthesizing m²or more frames.

[0063] Such synthesis of a distribution of signals is often called interpolation processing. The interpolation processing methods include a method using the Fourier series expansion and a method using the polynomial approximation or the linear approximation. The interpolation method using the Fourier series expansion will be described below.

[0064] First, the following function Y′ (x, y), which is capable of being integrated, is defined using the Dirac's δ function:

[0065] Y′(x, y)=Y(x, y)δ(X_pn_x−Δx(n_t)−x)·δ(Y_pn_y−Δy(n_t)−y)

[0066] wherein Y′ (x, y) is a function, which has an integration average equal to Y (x, y) near an optional point, at which Y (x, y) is defined, and which has integrated values of zero near the other points.

[0067] Then, the function Y (x, y) is expanded into a two-dimensional Fourier series (a half interval). Assuming that the function after interpolation is Y″ (x, y), the following formulae can be obtained. $Y^{″}\left(x,y\right)=\sum _{m,n}{a}_{m,n}\sin \left(\frac{m\pi x}{X_p{N}_x}\right)\sin \left(\frac{n\pi y}{Y_p{N}_y}\right)$ $\begin{array}{c}{a}_{m,n}=\frac{4}{N_x{N}_{y}N}·\int \int Y^{\prime }\left(x,y\right)\sin \left(\frac{m\pi x}{X_p{N}_x}\right)\sin \left(\frac{n\pi y}{Y_p{N}_y}\right)dxdy\\ =\frac{4}{\mathrm{NxNyN}}\sum _{n_x,n_y,n_t}Y\left(n_x,n_y,n_t\right)\sin \frac{m\pi \left(X_p·n_x-{\Delta }_x\left(\mathrm{nt}\right)\right)}{X_p{N}_x}\\ \sin \frac{n\pi \left(Y_p·n_y-{\Delta }_y\left(nt\right)\right)}{Y_p{N}_y}\end{array}$

[0068] This Fourier series has the amount of information N times as large as the amount of information of one frame for the original sequential still image.

[0069] In order to improve the resolution of the still image, it is assumed that the number of items of the Fourier series is N×N_x×N_y(1≦m≦N_x×N^½ and 1≦n≦N_y×N^½). If the luminance distribution Y (n_x, n_y) (1≦n_x≦X_n×N^½ and 1≦n_y≦Y_n×N^½) of the still image is calculated by the undermentioned formula, the number of terms of the series is equal to the amount of information, so that a still image having a resolution N^½ times as large as that of the original frame image can be obtained.

[0070] Y′″(n_x, n_y)=Y″(n_x/N^½, n_y/N^½)

[0071] FIGS. 5A through 5C are conceptualized views schematically showing such an interpolation processing. That is, FIG. 5A shows the distribution of signals before an interpolation processing. For simple explanation, FIG. 5A shows the distribution of signals of an image obtained by combining 2²=4 frames. When the above described interpolation processing of the image having such a distribution is carried out, the distribution of signals is uniform and the resolution of a still image is doubled in comparison with the image of each frame as shown in FIG. 5B.

[0072] On the other hand, in a case where the noise of a still image is reduced, it is assumed that the number of terms of the series is X_n×Y_n(1≦n_x≦X_nand 1≦n_y≦Y_n). If the luminance distribution Y (n_x, n_y) (1≦n_x≦X_nand 1≦n_y≦Y_n) of the still image is calculated by the undermentioned formula, the number of terms of the series is the same as the amount of information of one frame for the original sequential still image, so that the resolution is not improved. However, if a smooth function is applied by a low degree Fourier series, it is possible to obtain a still image having a reduced random noise 1/N^½ times as large as that of the original sequential still image.

[0073] Y′″(n_x, n_y)=Y″(n_x, n_y)

[0074] FIG. 5C is a conceptualized view showing the distribution of signals of the still image obtained by the above described interpolation processing. That is, the interpolation processing of the distribution of signals shown in FIG. 5A is carried out to obtain the still image having the distribution shown in FIG. 5C. As shown in this figure, although the resolution of the obtained still image is the same as that of each frame, the amount of original signals at the respective points is increased to improve the S/N ratio.

[0075] On the other hand, if the above described two intermediate processings are carried out, it is possible to obtain a still image having both of improved resolution and random noise in comparison with those of each frame for the original sequential still image. That is, the balance between resolution and noise can be suitably adjusted in accordance with the type of the object, the picture taking conditions and the characteristics of the optical system and image sensor of the camera.

[0076] In addition, the interpolation processing can be carried out any number of times as long as data are stored in the sequential still image memory. Therefore, the interpolation processing conditions can be modified by the feedback based on the image quality of the processed still image.

[0077] According to the above described interpolation using the Fourier series expansion, there is an advantage in that the accuracy of the interpolation is higher than that of the linear approximation when the displacement amounts of the respective frames are uneven.

[0078] The peripheral portion of the still image thus synthesized has a small amount of information due to the movement of the camera. In addition, in the case of the interpolation utilizing the Fourier series, the accuracy of the peripheral portion is bad due to the property thereof. Therefore, such a peripheral portion is cut from the still image to obtain a final still image.

[0079] While the color camera has been described as an example, the present invention should not be limited thereto. The present invention may be also applied to a monochrome camera. In this case, only the Y signals are processed. A Moreover, while the present invention has been applied to the digital camera, the invention should not be limited thereto, but the invention may be applied to any systems for capturing a still image of an object as electric information by means of an image sensor, such as a CCD and a CMOS.

[0080] In addition, the color signal system should not be limited to the R, G, B color system. For example, the present invention may be applied to a signal system of complementary colors of Ye (Yellow), Cy (Cyan) and Mg (Magenta) or a signal system containing G (Green) added thereto:

[0081] While the above described digital camera has had the sequential still image memory, the output of the signal processing part may be directly outputted to the sequential still image synthesizing still image converting part without the need of the sequential still image memory. In addition, the converted still image may be directly outputted to the outside without providing the synthesized still image memory in the digital camera.

[0082] While the signal processing part and the sequential still image synthesizing/still image converting part (the signal processing system) have processed the digital data, the analog processing may be carried out without the need of the A/D converting part.

[0083] While the contour and characteristic pattern have been extracted from the respective frames forming the sequential still images to derive the displacement amounts in the above described embodiment, the displacement amounts of the respective frames may be derived by other methods. For example, the camera may be provided with an angular velocity/velocity sensor, and the displacement amounts of the respective frames may be derived by monitoring the output of the angular velocity/velocity sensor when taking a picture. Alternatively, the camera may be provided with a dedicated image sensor for deriving the displacement amounts.

[0084] Moreover, while the plurality of frame images sequentially shot have been used to synthesize a still image, a plurality of frame images, which are not sequentially shot, such as frame images obtained by electronic shutter operations, may be used.

[0085] While the present invention has been disclosed in terms of the preferred embodiment in order to facilitate better understanding thereof, it should be appreciated that the invention can be embodied in various ways without departing from the principle of the invention. Therefore, the invention should be understood to include all possible embodiments and modification to the shown embodiments which can be embodied without departing from the principle of the invention as set forth in the appended claims.