GB2273223A - Correcting image blur due to camera movement - Google Patents

Abstract

In a hand-held video camera lens barrel 1 is pivotable with respect to supporting body 3. Image blur induced by shaking the camera is corrected by driving the lens barrel 1 to compensate for the measured angular displacement and velocity of the lens barrel. Video signals from imager 2 (eg CCD) are fed to motion vector detecting circuit 12; from this motion detected in the image is used to determine the angular velocity of the lens barrel by angular velocity calculator 13 taking zoom into account. The angular displacement of the lens barrel is determined by sensor 9, and together with the angular velocity is used to drive 15 actuator 5. By moving only the lens barrel relative to supporting body 3, the weight of the image blur correcting means is reduced and made suitable for use in a hand-held video camera. <IMAGE>

Description

Image Blur Correcting Apparatus The present invention relates to an image blur correcting apparatus which corrects a screen blur in an imaging apparatus such as a miniature camera, more particularly a portable video camera.

A blur of an image in a screen such as a television screen is caused by movement of a subject while it is imaged and by parallel movement (pan) of a camera in general. In addition, movement of the image caused by vibration of the camera or the like according to condition of shooting also sometimes causes the blur. For example, in case of telescopic shooting, since an angle of view is small, only small vibration causes large blur in the image, which causes visual hindrance. In addition, when a user shoots scenes with a video camera in his hand while he is moving or he is in a helicopter or the like, vibration of the camera also causes the blur of the screen.

As measures against the above, a blur proofing apparatus using a gyro sensor is provided. This apparatus moves a position or an optical axis in the reverse direction by an amount of displacement detected by the gyro sensor to control and correct it.

Meanwhile, it is proposed that the blur is corrected by electrically processing the image information.

Figure 6 is a block diagram showing a general structure of a conventional image blur correcting apparatus in which the image information is electrically processed. In figure 6, reference numeral 601 designates an analog/digi;tal converter (referred to as an A/D converter hereinafter), reference numeral 602 designates a memory which writes and reads an input signal, reference numeral 12 designates a moving vector detecting circuit which detects a direction and size of blur of the input signal, reference numeral 603 designates a memory reading control circuit which controls a readout address of the memory 602 in accordance with the movement from the moving vector detecting circuit 12, reference numeral 604 designates an interpolation circuit which expands and interpolates a part of a video signal read from the memory 602 and reference numeral 605 designates a digital/analog converter (referred to as a D/A converter hereinafter).

Next, blur correcting operation in this structure will be described hereinafter.

The input video signal is sampled at a constant sampling frequency and converted to a digital signal by the A/D converter 601 and then it is written in the memory 602.

In addition, the moving vector detecting circuit 12 detects movement (speed) in each frame from the correlation between continuous frames. The memory reading control circuit 603 reads a part of signals written in the memory 602 by moving a reference point (readout address) in accordance with an amount of movement obtained by the moving vector detecting circuit 12. The read signal is expanded and interpolated to an original screen size by the interpolation circuit 604 and then converted to a video signal by the D/A converter 605 and then it is output.

Figure 7 is a view in which a signal is read from the memory 602 by moving the reference point. Reference character W designates a region in which the input signal is written in the memory, which is always constant. Reference character Rg is a region in which a signal is read from the memory when an amount of movement is 0 and reference character Rxy is a region in which a signal is read from the memory when the amount of movement is x in the horizontal direction and y in the vertical direction. The amount of movement can be obtained by integrating the movement (speed) in each frame which is found during the moving vector detection and accumulation of errors caused by the integration can be prevented by bringing the readout region to the Rg in stable time.

Meanwhile, in this image blur correcting apparatus, it is ideal to calculate how far and which direction each pixel in the image moved to detect the amount of blur of the image between frames and this method shows the best precision in vector detection. However, it needs a large- scale hardware and takes a lot of time, so that it is difficult to realize the above method. Thus, there is a general method for determining the moving vector in the whole screen by finding an amount of movement of several pixels (referred to as representative points hereinafter) in the screen.

Figure 16 is a block diagram showing a general representative point calculating circuit. Figure 17 is a view showing the relation between a block in the image and a representative point in a conventional example. The image of one field is divided into the prescribed number of blocks 115 and one representative point Rij 114 is provided in the center of each block. The level difference between the representative point in the previous frame and each of the pixels 8i+x, j+y 116 in the block is calculated for each block.

Referring to figure 16, an input video signal (a) is A/D converted by the A/D converter 101 and then a prescribed pixel to become the representative point 114 in the block 115 is written in the representative point memory 103 through a latch circuit 102. Data stored in the representative point memory 103 is read one frame later and applied to an absolute value circuit 106 through a latch circuit 104. On the other hand, the other A/D converted data of the video signal is applied to the absolute value circuit 106 through a latch circuit 105. The absolute value of the difference between the representative point signal (b) of the previous frame output from the latch circuit 104 and the pixel signal (c) of the present frame is calculated by the absolute value circuit 106.This calculation is made in each block and an output signal (d) from the absolute value circuit 106 is added one after another to tables corresponding to the same address of the pixel in each block of the accumulated addition table 108. The result of addition in the tables is input to a table value comparator 109 and finally it is determined how far and which direction the image position moved for one frame, that is, a moving vector value (e) is determined from an address in which the result of addition is the minimum.More specifically, the absolute value of the difference between the representative point Rij and a signal Si+x, j+y positioned in a horizontal direction x and a vertical direction y is found and then it is added to x and y of the same position to each representative point and then it is set as Dxy At this time, the Dxy is represented as follows; Dxy = z | Rij - Si+x, j+y I Then, the minimum x and y in the Dxy are set as moving vectors in the horizontal direction and vertical direction, respectively.

At this time, one block in which the representative point is positioned at the center thereof shows a maximum size of the vector, that is, a range in which the vector can correct the image. In addition, in view of the accumulated addition, the more the number of additions is, the more precision of detecting the vector is improved. Thus, the moving vector is obtained by adding the representative points in the whole screen, that is, averaging the movement of the whole screen.

Although the conventional imaging apparatus has the above structure in which the gyro sensor is used for detecting blur, the sensor is expensive and it is necessary to determine a position of the sensor and also to secure a space therefor.

In addition, in order to prevent the image from being degraded, an A/D converter and a D/A converter with high precision, a memory with large capacity and a large-scale circuit for memory control or expansion and interpolation are necessary, with the result that the system costs a lot and also current consumption becomes large. In addition, since the corrected signal is a part of the continuously input signals, an amount of information is reduced and resolution of that signal is low. If a range of correction is increased in order to correct blur, the amount of information is further reduced and the resolution is also lowered. Therefore, the range of correction can not be increased so much.

In addition, according to the above structure of the moving vector detecting circuit, since the moving vector is found in a plane manner (two dimensionally), the accumulated addition tables whose number corresponds to the number of whole pixels in the block are necessary. For example, if the number of the pixels are 32 in the horizontal direction and 16 in the vertical direction, 512 (= 32 X 16) accumulated addition tables are necessary, with the result that the scale of the circuit is increased. In addition, when the moving vector is calculated, whole data in the accumulated addition tables are compared, with the result that the number of times of comparison is very large and also it takes a lot of time.

In addition, according to the moving vector detecting apparatus in the above conventional image blur correcting apparatus, movement over the whole screen is averaged, so that arrangement of the blocks for detecting the moving vector is not always symmetrical about the center of the screen. Therefore, although the moving vector can be well detected when the image only moves in parallel, the moving vector often can not be detected when the image is zoomed in or out. In addition, when the vector is wrongly detected, it is more conspicuous at the time of zooming operation than that at the time of parallel movement, which looks strange for the user.

It is an object of the present invention to provide an inexpensive, small and light image blur correction apparatus.

The invention employs a pivotable lens barrel with an actuator which is controlled in response to a motion vector detected in the sensed image.

It is a further object of the present invention to provide an image blur correcting apparatus having a moving vector detecting circuit capable of detecting a moving vector which does not look strange for a user when zooming operation which is often used at the time of shooting by a video camera is performed.

Other object and advantages of the present invention will become apparent from the detailed description given hereinafter; it should be understood, however, that the detailed description and specific embodiment are given by way of illustration only, since various changes and modifications within the spirit and scope of the invention will become apparent to those skilled in the art from this detailed description.

According toonenbodbnt of the present invention, an image blur correcting apparatus comprises a lens-barrel part having a plurality of lenses and an imaging device, image signal processing means for forming an image signal from an electrical signal obtained at the imaging device, a supporting body for supporting the lens-barrel part in a freely pivoting manner around a rotating axis crossing a beam axis incident to the lens-barrel part at the right or almost right angles, actuator means attached between the lens-barrel part and the supporting body for rotating the lens-barrel part, relative angle detecting means for detecting a relative angle between the lens-barrel part and the supporting body, means for detecting a moving vector showing an amount of blur of the image from the correlation between two frames of image information from the image signal processing means which are continuous in view of time, angular velocity operating means for calculating rotation angular velocity using zoom multiplying factor from a detected result of the moving vector, and calculation controlling means for outputting a control signal of an actuator in accordance with the rotation angular velocity information and relative angle information.

In embodiments of the present invention, in an image blur correcting apparatus, moving vectors of the whole screen nay be detected from an amount of movement of a pixel in each block on the basis of representative--points set in a plurality of blocks in an image in order to detect an amount of movement between frames of the image.This image blur correcting apparatus may comprise means for operating the absolute value of the difference between the pixel of the representative point for each block of the frame direct before the frame whose movement vector is to be detected and the pixels on a horizontal straight line and a vertical straight line passing through the representative point, respectively, of the frame whose movement is to be detected, horizontal and vertical accumulated addition tables for accumulatedly adding the horizontal and vertical absolute values, means for detecting two intersecting onedimensional moving vectors from the horizontal and vertical accumulated added values, and means for calculating a twodimensional moving vector from the two intersecting onedimensional moving vectors.

In further embodiments of the present invention, an image blur correcting apparatus comprises means for finding the absolute value of difference between two continuous frames every block after one screen is divided into a plurality of blocks, means for calculating tables by adding the absolute value of the difference in each block over an area after one screen is divided into a plurality of areas comprising the plurality of blocks, means for finding a movingovector in each area from the above table in each area, means for determining whether the whole screen moves in parallel or zooms in or out from the obtained moving vector in each area, and means for outputting the moving vector having a value obtained by averaging the moving vectors in the whole screen when the parallel movement is determined or the moving vector having a value obtained by multiplying the average value of the whole screen by a prescribed attenuation constant k (0 ( k < 1) when the zooming operation is determined.

According to the above emb nts of the present invention, the moving vector of the image is detected directly from the video signal, rotation angular velocity is calculated by zoom multiplying factor or the like, the blur is corrected by driving the actuator to form a closed loop, a relative angle between the lens-barrel part and the supporting body is detected and the actuator is driven so that the above angle may be zero in a low frequency region.

Therefore, the blur of the image can be corrected independent on the zoom multiplying factor and the lensbarrel part can be controlled so as to return to the reference position of the supporting body in normal time.

In addition, according to the motion vector detecting apparatus having the above structure, since only two intersecting one-dimensional vectors are found from the accumulated addition tables, the number of the accumulated addition tables can be the sum of the number of pixels in the horizontal direction and the number of pixels in the vertical direction. For example, if the number of pixels in the horizontal direction is 32 and that in the vertical direction is 16 in a block, the number of the accumulated addition tables is only 48 (= 32 + 16), so that the accumulated addition table can be implemented by circuits whose scale is one tenth or less as large as the conventional one. In addition, time required for calculating the moving vector can be reduced when the number of the accumulated addition tables is reduced.

Furthermore, according to certain embodiments the movement in the whole screen is found not by averaging it but by detecting moving vector every area obtained by dividing the screen and determining whether the whole screen moves in parallel or it is in course of the zooming operation from a fact that the moving vector in each area is directed to the same direction or a fact that it is directed to the opposite direction. As a result, even when the zooming operation is determined, the user can detect the moving vector which does not look strange for him.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a view showing an image blur correcting apparatus in accordance with an embodiment of the present invention; Figure 2 is a view showing a specific structure of an actuator shown in figure 1; Figure 3 is a view showing a specific structure of a relative angle detecting circuit shown in figure 1; Figure 4 is a block diagram showing a structure of a moving vector detecting circuit shown in figure 1; Figure 5 shows the relation of positions between an screen and a representative point during calculating operation shown in figure 4; Figure 6 is a block diagram showing a structure of a conventional image blur correcting apparatus which performs electrical processing; Figure 7 is a view showing regions in which a signal is written and read, into and from a memory, respectively in the apparatus;; Figure 8 is a block diagram showing a representative point calculating circuit as used in a second embodiment of the present invention; Figure 9 shows the relation between a block and a representative point in an image during the representative point calculation shown in figure 8; Figures 10 and 11 are schematic views showing accumulated addition tables; Figure 12 is a view showing a horizontal accumulated addition table; Figure 13 is a block diagram showing a moving vector detecting apparatus as used in 'a third embodiment of the present invention; Figure 14 is a view showing an example of the relation between a block and an area; Figure 15 are views for describing operation of parallel movement/zooming operation determining process; Figure 16 is a block diagram showing a general representative point calculating circuit; and Figure 17 shows the relation between a block and a representative point in an image in a conventional example.

DESCRIPTION OF THE PREFERRED EMBODIMENT An embodiment of the present invention will be described in detail in reference to the drawings hereinafter. Figure 1 is a view showing a structure of an image blur correcting apparatus in accordance with a first embodiment of the present invention. In figure 1, a plurality of lenses (not shown) and an imaging device 2 (for example a CCD plate or an image pickup tube) are attached to a lens-barrel part 1, in which reflected light from a subject is converged to form an image at the imaging device 2, which is converted to electric charge signals (electric signal). Then, a video signal processing circuit 10 reads the electric charge signals, one after another and forms a video signal.

An actuator 5 is arranged between the lens-barrel part 1 and a supporting body 3, which rotates the lens-barrel part 1 around a rotation axis 6 in the yaw direction, the lensbarrel part 1 being freely pivotable on an almost horizontal surface when in use.

The rotation axis 6 of the actuator 5 passes through the centre of gravity G of the lens-barrel part 1 and is rotatably supported by the supporting body 3. In addition, a grip 4 of the imaging apparatus, held by the user, provided on the supporting body 3.

Reference numeral 11 designates a relative angle detecting circuit, which detects a relative angle between the lens-barrel part 1 and the supporting body 3 from an output of a Hall element 9; reference numeral 12 designates a moving vector detecting circuit, which detects a moving vector from the video signal formed by the video signal processing circuit 10; reference numeral 13 designates an angular velocity calculating means for calculating rotational angular velocity of the lens-barrel part 1 by a zoom multiplying factor or the like from the output of the moving vector detecting circuit 12 and reference numeral 14 designates calculation controlling means, to which the outputs of the relative angle detecting circuit 11 and the angular velocity calculating means 13 are input, for controlling the actuator 5 through the driving circuit 15.

Figure 2(a), (b) and (c) are views each showing a structure of the actuator 5 in detail. In figure 2, a back yoke 201, which is made of ferromagnetic material, of a magnet 202 is attached to the lens-barrel part 1 and rotates together with the rotation axis 6. The magnet 202 is magnetized so as to have four magnetic poles and generates a field magnetic flux. Coils 204a and 204b and a Hall element (magnetic sensor) 9 are fixed to the coil yoke 203 to which a bearing 207 of the rotation axis 6 is attached.

According to the embodiment of the present invention, the magnet 202 is attached to the lens-barrel part 1 and the coil yoke 203 is attached to the supporting body 3. The coil 204a is connected to the coil 204b in series and a rotating torque is generated by a current flowing from a terminal 205 to a terminal 206 and the magnetic flux of the magnet 205. In addition, the Hall element 9 is arranged in a part opposite to a part in which the magnetic pole of the magnet 202 is switched and generates an output signal corresponding to the relative difference of angles between the magnet 202 (an angle of the lens-barrel) and the coil yoke 203 (an angle of the supporting body). The output of the Hall element 9 which detects the magnetic field of the magnet 202 of the actuator 5 is input to a relative angle detecting circuit 11.

Figure 3 is a view showing a structure of the relative angle detecting circuit 11 in detail. DC signals obtained at two output terminals of the Hall element 9 are differentially amplified by prescribed number times by a differential amplifying circuit comprising an operational amplifier 301 and resistors 302, 303, 304 and 305, whereby an output signal (c) is provided. Voltages of +VH and -VH are appropriate voltages, which apply appropriate bias to the Hall element 9 through resistors 306 and 307.

Figure 4 is a block diagram showing an example of the moving vector detecting circuit 12. In figure 4, an input video signal (a) is generated by the video signal processing circuit 10 and A/D converted by an A/D converter 401. Then, one of the A/D converted signals is input to a representative point memory 403 through a latch circuit 402.

A prescribed pixel in a block written in this memory is set as a representative point. Therefore, the output of the representative memory 403 is the representative point written one frame before, which is read every block and becomes a signal (e) through the latch circuit 404. The other video signal input to this block is A/D converted and then becomes a signal (d) through the latch circuit 405.

Then, the absolute value of the difference between the signal (d) and the representative point signal (e) is calculated by the absolute value circuit 406. This calculationsis performed every block whose number is the same as that of the representative points and its output signals are added one after another to tables corresponding to the same address of each block of an accumulated addition table 207. The result of the tables is input to a table value comparator 408 and then it is found how far and in which direction the image position moved for one frame from an address in which the result of addition is the minimum at a final stage. More specifically, a vector value (b) is found.

Figure 5 shows the relation between a block of the image and its representative point during general representative point calculation described above, in which it is shown that calculation is made between a pixel 503 input to the block 502 and a representative point level in each block in view of the representative point 501 taken out one frame before.

More specifically, the absolute value of the difference between the representative point Rij and a signal Si+x j+y positioned apart in a horizontal direction by x and-in a vertical direction by y from that is added for each representative point and then it is set as Dxy At this time, the Dxy is represented as follows;

Then, the minimum x and y in the Dxy are set as moving vectors in the horizontal direction and vertical direction, respectively.

At this time, one block in which the representative point is positioned at the center thereof shows a size of the vector, that is, a range in which the image can be corrected by the vector. In addition, in view of the accumulated addition, the more the number of additions is, the more precision of detecting the vector is improved.

Thus, the moving vector is obtained by adding the representative points in the whole screen, that is, averaging the movement of the whole screen.

Next, a control loop will be described hereinafter.

When a hand of the user holding the grip 4 rotates at certain speed, the lens-barrel part 1 also rotates at the same speed. When the lens-barrel part 1 rotates, the video signal output from the image signal processing circuit 10 moves in the left direction in parallel and the image of the representative point in the previous frame moves in the left direction by n pixels in the next frame. Then, the absolute value of the output of an absolute value circuit 406 in the moving vector detecting circuit 12 is lowest for the pixel on the left by n pixels. Therefore, the accumulated addition table 407 in which the absolute value in each block is accumulated has its lowest entry corresponding to the address on the left side by n pixels. Then, it is found that the moving vector has a size of n pixels/frsame in the left direction.

The rotating angular velocity of the lens-barrel part 1 at this time is calculated by the angular velocity operating means from the thus obtained moving vector and a zoom ratio (when the zoom ratio is large (small), the angular velocity is calculated on a large (small) scale) and then the actuator 5 is driven so that this angular velocity may be zero.

However, oscillation or deviation is generated by way of only the angular velocity control loop. In addition, although the angle can be detected by integrating the detected angular velocity, a detection error is also accumulated at this time. Then, the actuator 5 is controlled so that the relative angle obtained from the relative angle detecting circuit 11 may be zero in a low frequency region. Thus, the blur of the image can be corrected by two loops independent of the zoom ratio and the lens-barrel part is controlled so as to return to the reference position of the supporting body in normal time.

Although a description was given of the embodiment in which the blur in a yaw direction is prevented, it is needless to say that the present invention can prevent blur in a pitch direction.

In addition, although the relative angle is detected by the Hall element attached to the actuator in the above embodiment, any sensor can be used if it can detect the angle between the lens-barrel part and the supporting body.

In addition, this image blur correcting apparatus can be used not only in the video camera but also in various other apparatuses. within a principle of the present invention.

Meanwhile, although the moving vector detecting circuit shown in figure 4 is used in the image blur correcting apparatus shown in figure 1, the circuit shown in figure 4 is large in scale and it takes time to calculate the moving vector therein.

Figure 8 is a block diagram showing a representative point calculating circuit in an image blur correcting apparatus in accordance with a second embodiment of the present invention, which can solve the above problems.

Figure 9 is a view showing the relation between the block of the image and the representative point in the embodiment of figure 8.

According to the apparatus shown in figure 8, the image of one field is divided into prescribed number of blocks 115 and one representative point R 114 is provided at the centre of each block. The level difference between the representative point in the previous frame and a pixel Si+x j+y 116 in the vertical direction of the representative point and a pixel Si+x j+y 117 in the horizontal direction of the representative point in the block is calculated every block.

In figure 8, an input video signal (a) is A/D converted by an A/D converter 121 and then a prescribed pixel in the block 115 to become the representative point 114 is written in a representative point memory 123 through a latch circuit 122. Data stored in the representative point memory 123 is read one frame after and then applied to a vertical absolute value circuit 127 and a horizontal absolute value circuit 128 through the latch circuit 124.On the other hand, the data of the A/D converted video signal is applied to the vertical absolute value circuit 127 through a vertical pixel latch circuit 125 which latches at a timing corresponding to the pixel in the vertical direction of the representative point and also applied to the horizontal absolute value circuit 128 through a horizontal pixel latch circuit 126 which latches at a timing corresponding to the pixel in the horizontal direction of the representative point.

A representative point signal (b) in the previous frame which is output from the latch circuit 124 and a pixel signal (c) in the present frame which is output from the vertical pixel latch circuit 125 are calculated at the vertical absolute value circuit 127 to find the absolute value of the difference between them. A representative point signal (b) in the previous frame which is output from the latch circuit 124 and the pixel signal (d) in the present frame which is output from the horizontal pixel latch circuit 126 are calculated at the horizontal absolute value circuit 128 to find the absolute value of the difference between them.These calculations are performed every block and an output signal (e) from the vertical absolute value circuit 127 is added one after another to tables corresponding to the same address of the pixel in each block in the vertical accumulated addition table 129 and an output signal (f) from the horizontal absolute value circuit 128 is added one after another to tables corresponding to the same address of the pixel in each block in the horizontal accumulated addition table 130. The result of the horizontal accumulated addition table 130 and the result of the vertical accumulated addition table 129 are input to one-dimensional vector detecting means 131 and then a vertical moving vector (g) and a horizontal moving vector (h) are detected therein.The vertical moving vector (g) and the horizontal moving vector (h) which are two intersecting one-dimensional vectors are input to twodimensional vector calculating means 133 and then it is found how far and in which direction the image position moved for one frame at the final stage, that is, a twodimensional moving vector value (i) is determined.

At this time, one block in which the representative point is positioned at the centre thereof shows a size of the vector, that is, a range in which the vector can correct the image. In addition, in view of the accumulated addition, the more the number of additions is, the more precision of detecting the vector is improved. Thus, the moving vector is obtained by adding the representative points in the whole screen, that is, averaging the movement of the whole screen.

Next, the one-dimensional vector detecting means 131 will be described. Figures 10 and 11 are schematic views showing the accumulated addition table for describing the one-dimensional vector detection. Figures 10 (a) and ll(a) show the accumulated addition table corresponding to the whole number of pixels in the block in which a horizontal direction (x), a vertical direction (y) and a value of the accumulated addition table (z) are represented by three dimensions with the representative point as the origin.In addition, Figures 10(b) and ll(b) show a vertical accumulated addition table Dy obtained by finding the absolute value of the difference between the representative point Rij and a signal Si, j+y positioned in the vertical direction y thereof and adding the same over those at the same position y relation with respect to each representative point, in which the vertical direction (y) and the value of the accumulated addition table (z) are represented by two dimensions with the representative point as the origin. Figures 10(c) and ll(c) show a horizontal accumulated addition table Dx obtained by finding the absolute value of the difference between the representative point Rij and a signal Si+x, j positioned apart from that in the horizontal direction by x and adding the same over those at the same position x relation with respect to each representative point, in which the horizontal direction (x) and the value of the accumulated addition table (z) are represented by two dimensions with the representative point as the origin.

At this time, they are represented by the following equations, that is;

Then, a horizontal moving vector (h) and a vertical moving vector (g) are detected by using the Dx and Dy.

First, a description will be given of a first detecting method in which the minimum value x in the horizontal accumulated addition table Dx is detected as the horizontal moving vector (h) and the minimum value y in the vertical accumulated addition table Dy is detected as the vertical moving vector (g).

If the video signal in which correlation in both horizontal and vertical directions is reduced in proportion to a distance is at a standstill, the accumulated addition table value is in the form of an inverted cone with the origin (0, 0, 0) as an apex as shown in figure 10(a). At this time, the value of the vertical accumulated addition table 9 shown in figure 10(b) is the minimum when y = 0 and the value of the horizontal accumulated addition table 10 is the minimum when x = 0. Therefore, it is found that the vertical moving vector and the horizontal moving vector are both 0 vector.

Then, if the video signal moves by c in the horizontal direction and by b in the vertical direction for one frame, the accumulated addition table is in the form of the inverted cone with (c, b, 0) as an apex shown in figure ll(a). At this time, the vertical accumulated addition table 9 is the value of a section of the cone as shown in figure ll(b) when a plane x - 0 and the horizontal accumulated addition table 10 is the value of a section of the cone as shown in figure ll(c) when a plane y = 0. Then, as can be seen from figure ll(b), the value of the vertical accumulated addition table 9 is the minimum when y = b.

Also, as can be seen from figure ll(c), the value of the horizontal accumulated addition table 10 is the minimum when x = c. Therefore, it is found that the vertical moving vector (g) is (O, b) and the horizontal moving vector (h) is (c, O).

Next, a description will be given of a second detecting method in which configuration of the accumulated addition table is previously detected and then the vertical moving vector (g) is found from the minimum value in the vertical accumulated addition table Dx and the horizontal moving vector (h) is found from the minimum value in the horizontal accumulated addition table Dy.

When the accumulated addition table having the form of inverted cone as shown in figure ll(a) moves by c in the horizontal direction and by b in the vertical direction, the accumulated addition table is represented by the following equation, that is; (x - c)2 + (y - b)2 = f(z))2 (z > 0) If it is assumed that it has a perfect cone shape for simplicity, it follows that f(z) = kz.Therefore, the accumulated addition table can be represented as follows, that is; (x - c)2 + (y - b)2 = (kz)2 (z > 0) Since the vertical accumulated addition table 9 is one having the section at the plane x = 0, it can be represented as follows, that is; c2 + (y - b)2 = (kz)2 When y = b, it follows that z0 = + c/k and then z is the minimum because of the following equation, that is; z = (c2 + (y - b)2)1/2/k Therefore, on the contrary, it is found that the horizontal moving vector (h) (c, 0) can be obtained when the minimum z0 of z is found.

More specifically, it is found that c = + k ' z0 and the horizontal moving vector is (k . zO 0) or (-k z0, 0).

Then, either one can be determined by determining whether c is positive or negative value in reference to the horizontal accumulated addition table.

In addition, since the horizontal accumulated addition table 10 is one having the section at the plane y = 0, it can be also represented as follows, that is; (x - C)2 + b2 = (kz)2 When x = c, it follows that zO = + b/k and then z is the minimum because of the following equation, that is; z = (x - c)2 + b21/2/k Therefore, on the contrary, it is found that the vertical moving vector (g) (0, b) is obtained when the minimum z0 of z is found.

More specifically, it is found that b = + k ' z0 and the moving vector is (0, k ' zo) zO) or (0, -k zO). Then, either one can be determined by determining whether c is positive or negative value in reference to the vertical accumulated addition table.

Although it is assumed that the accumulated addition table is in the form of perfect cone and a description was given of a case where f(z) = kz, the same can be found as the function of z in general. Although the function f(z) of z at this time varies with contents of the image at that time, since there is small variation in case of continuous video signals, it is always possible to find f(z) by detecting the configuration of the accumulated addition table every frame.

Figure 12 is a view showing an ideal horizontal accumulated addition table, which shows a steep hyperbolic curve when movement in a diagonal direction is small as shown in figure 12(a) and shows a gentle hyperbolic curve when movement in a diagonal direction is large (Figure 12(b)).

On the other hand, referring to figure 12, the first method is that xO when z is the minimum is detected and the horizontal moving vector is found from this xO and the second method is that the minimum value z0 of z is detected and then the vertical moving vector is found from this zO.

In case of an actual horizontal accumulated addition table, an error of az in the direction of z exists because of a quantization error, an round-off error, a difference in horizontal correlation level of the image or the like.

Therefore, an error of ax is also generated in x at this time. In case of the gentle hyperbolic curve as shown in figure 12(b), since lx is large, detecting precision of x is low but since z0 is large, detecting precision of z is improved.

Since the same is also applied in the vertical accumulated addition table, the second detecting method is more effective than the first detecting method when movement in the diagonal direction is large as shown in figure 12(b).

Although the first and second detecting methods of one-dimensional vector detecting means was described, a final result can be obtained by adding the detected results of the first and second embodiments at a prescribed ratio or either one can be selected in accordance with the configuration of the accumulated addition table.

Figure 13 is a block diagram showing a representative point calculating circuit in the image blur correcting apparatus in accordance with a third embodiment of the present invention, which solves the problems of the moving vector detecting circuit in figure 4 in which moving vector can not be successfully detected when the image is one at the time of zooming operation.

In figure 13, an input video signal (a) which is a video signal formed by the video signal processing circuit 10 is A/D converted by an A/D converter 141 first, and then one of the A/D converted signals is input to a representative memory 143 through a latch circuit 142. A prescribed pixel in a block written in this memory becomes a representative point. Therefore, an output from the representative memory 143 is the representative point written one frame before, which is read every block and becomes a signal b through a latch circuit 144. The absolute value of the difference between the signal (c) which is the other A/D converted video signal input to this block and passed through a latch circuit 145 and the representative point signal (b) is calculated by an absolute value circuit 406.Reference numeral 147 designates a switch through which the result of the absolute value circuit 146 is output to one of accumulated addition tables 148a to 148i of each area, reference numeral 150 designates parallel movement/zooming operation determining means for determining parallel movement or zooming operation from an output of each of table value comparators 149a to 149i at each area, reference numeral 151 designates moving vector averaging means for finding an average value of the outputs of the table value comparators 149a to 149i, reference numeral 152 designates a multiplying circuit which multiplies k (O ~ k < 1) and reference numeral 153 designates a switch.

Figure 14 is a view showing the relation between the block and the area in which one screen is divided into 81 (= 9 X 9) blocks and one area is divided into 9 (= 3 X 3) blocks, that is, one screen is divided into 9 (= 3 X 3) a to i areas. Figure 15 are views for describing the parallel movement/zooming operation determining process.

Next, operation thereof will be described hereinafter. The input video signal (a) is A/D converted by the A/D converter 141 and a prescribed pixel to become the representative point 14 in the block 15 is written in the representative memory 143 through the latch circuit 142 and the absolute value of the difference between the representative signal (b) in the previous frame output from the latch circuit 144 and the video signal (c) in the present frame output from the latch circuit 145 is calculated by the absolute value circuit 146. The above is the same as before. These calculations are made every block and the output signal (d) of the absolute value circuit 146 is switched by the switch 147 and then added one after another to a table corresponding to each pixel of 9 accumulated addition tables 148a to 148i each corresponding to each area comprising the blocks.

The result of addition at the table of each area is input to each of the table value comparators 149a to 149i in each area. Finally, it is found that how far and which direction the image position moved for one frame by a block address in which the addition result is the minimum in each area, that is, 9 moving vector values ea , ei are found at each of 9 areas.

These 9 moving vector values are averaged by moving vector averaging means 151 and one average value is input to the switch 153 as it is and the other passes through the multiplying circuit 152 multiplying k (0 ( k < 1) to be attenuated and then input to the switch 153.

On the other hand, these 9 moving vector values are input to the parallel movement/zooming operation determining means 150 and when parallel movement is determined, the switch 153 is switched so that the output of the moving vector averaging means 11 may be the moving vector as it is but when zooming operation is determined, the switch 153 is switched so that the output of the multiplying circuit 152 may be the moving vector.

Next, operation of the parallel movement/zooming operation determining means 150 will be described in reference to figure 15. Referring to figure 15(a), although the moving vectors of a area to i area have a little variation, their x and y directions are both positive values and close to the average value of the vector, so that the parallel movement is determined. Therefore, in this case, the switch 13 is switched so that the output (2, 2) from the moving vector averaging means 151 may be the moving vector as it is.

On the other hand, referring to figure 15(b), since the moving vectors of a area and i area, c area and g area, b area and h area, and d area and f area are almost oppositely directed, respectively, it is determined that they are under zooming operation. Therefore, in this case the switch 153 is switched so that the output (k 5/9, 0) from the multiplying circuit 152 may be the moving vector.

Although movement of the screen is detected between frames in the above embodiment, it can be detected between fields.

As described above, according to the image blur correcting apparatus of the present invention, the blur of the image is detected by electrically processing image information and then an optical axis is corrected by rotating the lens-barrel part in accordance with the above detected output. Therefore, when the blur of the image is detected, an expensive gyro sensor is not necessary and also an expensive A/D converter, D/A converter, memory or the like is not necessary, with the result that the image blur correcting apparatus which is inexpensive and has low power consumption can be provided.

In addition, according to the embodiments pixels to be compared with a representative point in a block are only positioned on the upper, lower, right and left sides of the representative point and then two one-dimensional vectors of a vertical moving vector and horizontal moving vector are individually calculated from the minimum tables of the vertical accumulated addition table obtained from the upper and lower parts and the horizontal accumulated addition table obtained from the right and left parts and then a two-dimensional moving vector is calculated from the above two intersecting one-dimensional vectors of the vertical moving vector and the horizontal moving vector. As a result, the moving vector can be detected with a few accumulated addition tables in short calculatina time.

Furthermore, according to the above embodiments,"the moving vector is detected not by averaging the movement in the whole screen but by detecting moving vector every area obtained by dividing the screen and determining whether the whole screen moves in parallel or it is in course of the zooming operation from a fact that the moving vector in each area is directed to the same direction or a fact that it is directed to the opposite direction to each other and then when the zooming operation is determined, the value obtained by multiplying the average of the moving vectors of the whole screen by a prescribed attenuation constant becomes the moving vector.As a result, there is provided the vector detecting apparatus capable of detecting a moving vector which does not look strange for the user even when zooming operation which is often used during shooting is performed.

Although the present invention has been described and illustrated in detail, it is clearly understood that the same is by way of illustration and example only and is not to be taken by way of limitation and example only and is not to be taken by way of limitation; the spirit and scope of the present invention being limited only by the terms of the appended claims.

The present disclosure also provides the features and combination of features set forth in the following numbered clauses: 1. An image blur correcting apparatus comprising: a lens-barrel part having a lens and an imaging device; image signal processing means for forming a video signal from an electrical signal obtained at said imaging device; a supporting body for supporting said lens-barrel part in a freely pivoting manner around a rotating axis crossing a beam axis incident to said lens-barrel part at the right or almost right angles; actuator means attached between said lens-barrel part and said supporting body for rotating said lens-barrel part; relative angle detecting means for detecting a relative angle between said lens-barrel part and said supporting body; ; means for detecting a moving vector indicating an amount of blur of the image from the correlation between two frames of image information from said image signal processing means which are continuous in view of time; angular velocity operating means for calculating rotation angular velocity using zoom multiplying factor or the like from a detected result of the moving vector; and calculating controlling means for outputting a control signal of an actuator in accordance with the outputs of said relative angle detecting means and said angular velocity detecting means.

2. An image blur correcting apparatus in which moving vectors of the whole screen are detected from an amount of movement of a pixel in each block on the basis of representative points set in a plurality of blocks in an image in order to detect an amount of movement between frames of the image, said image blur correcting apparatus comprising:: means for operating the absolute value of the difference between the pixel of the representative point for each block of the frame direct before the frame whose movement vector is to be detected and the pixels on a horizontal straight line and a vertical straight line passing through the representative point, respectively, of the frame whose movement is to be detected; horizontal and vertical accumulated addition tables for accumulatedly adding the horizontal and vertical absolute values; means for detecting two intersecting one-dimensional moving vectors from the horizontal and vertical accumulated added values; and means for calculating a two-dimensional moving vector from said two intersecting one-dimensional moving vectors.

3. An image blur correcting apparatus as defined in clause 2 wherein said one-dimensional moving vector detecting means detects moving vectors in the horizontal and vertical direction on the basis of the coordinates of the minimum accumulated addition values of said horizontal and vertical accumulated addition tables.

4. -An image blue correcting apparatus as defined in clause 2 wherein said one-dimensional moving vector detecting means obtains a curved surface representing the accumulated added values from the accumulated addition tables, detects a point closest to the XY plane on the curved surface, and obtains moving vectors in the horizontal and vertical direction on the basis of the coordinates of the closest point.

5. An image blur correcting apparatus comprising: means for finding the absolute value of difference between continuous two screens every block after one screen is divided into a plurality of blocks; means for calculating a table by adding the absolute value of the difference at each block over an area after one screen is divided into a plurality of areas comprising the plurality of blocks; means for finding a moving vector in each area from said table accumulated every area; parallel movement/zooming operation determining means for determining whether the whole screen moves in parallel or it is in course of zooming operation from said obtained moving vector at each area; and means for outputting the moving vector having a value obtained by averaging the moving vectors in the whole screen when said parallel movement/zooming operation determining means determines the parallel movement and the moving vector having a value obtained by multiplying the average value of the whole screen by a prescribed attenuation constant k (0 < k < 1) when said means determines the zooming operation.

6. An image blur correcting apparatus as defined in clause 5 wherein one screen is divided into nine areas.

7. An image blur correcting apparatus as defined in clause 6 wherein said parallel movement/zooming judging means judges parallel movement when the moving vector of each area is toward the same direction.

8. An image blur correcting apparatus as defined in clause 6 wherein said parallel movement/zooming judging means judges zooming operation when the moving vector of each area confronting to each other are in reverse directions.

Claims (14)

1. An image blur correcting apparatus comprising: a lens barrel part mounted pivotably within a supporting body of the apparatus, the lens barrel part having a lens and an imaging device for generating electrical image signals representing a moving image; actuator means for rotating said lens barrel part relative to the supporting body; means for detecting from the electrical image signals a motion vector indicating motion of the image between successive image frames; and means for controlling the actuator in response to the detected motion vector so as to compensate for shaking of the supporting body.

2. An apparatus according to claim 1 wherein the controlling means includes means for calculating an angular velocity from the detected motion vector.

3. An apparatus according to claim 2 wherein the angular velocity is determined in accordance with a variable zoom factor.

4. An apparatus according to claim 1, 2 or 3 further comprising parallel movement/zooming operation determining means for determining whether the whole image is moving in parallel or is in the course of the zooming operation and wherein the controlling means operates in a manner dependent on the result of said determination.

5. An apparatus according to claim 4 wherein a plurality of motion vectors are detected for respective areas of the image and processed by said parallel movement/zooming operation determining means.

6. An apparatus according to claim 5 wherein when parallel movement is determined the actuator is controlled in accordance with an average of the area motion vectors, and when zooming operation is determined the actuator is controlled in accordance with an attentuated average motion vector.

7. An image blur correcting apparatus according to claim 5 or 6 wherein each image is divided into nine areas.

8. An image blur correcting apparatus according to claims 5, 6, or 7 wherein said parallel movement/zooming operation determining means determines parallel movement when the motion vectors of all the areas have substantially the same direction.

9. An image blur correcting apparatus according to any one of claims 5, 6, 7 or 8 wherein the parallel movement/zooming operation determining means determines zooming operation when the motion vectors of opposing image areas are in substantially opposite directions.

10. An apparatus according to any preceding claim wherein the controlling means is further responsive to an output of means for detecting the relative angular displacement of the lens barrel and the supporting body.

11. An apparatus according to claim 10 wherein said actuator is controlled to maintain a zero frequency component of the relative angular displacement at a fixed reference value.

12. An image blur correcting apparatus substantially as described herein with reference to any of Figures 1 to 5 and 8 to 15 of the accompanying drawings.

13. A video camera including an image blur correcting apparatus according to any one of claims 1 to 12.

14. A video camera according to claim 13 suitable for hand-held operation.