JP2003079620A - Ultrasonic panorama image forming system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To form a panorama image of an ultrasonic diagnosing image by a comparatively small computation amount. SOLUTION: An image frame 100 is divided into a plurality of divided regions 104. Then, detection regions 110 are determined by removing a moving margin 106 or the like from respective regions 104, for which a probe movement between frames is taken under consideration. Then, for respective detection regions 110, a picture element wherein a difference in a picture element value with the picture elements on the periphery is largest is detected as candidates points 120a to 120d. The dispersion of a matrix region for an N×N picture element in the vicinity of the points is obtained for respective candidates points 120a to 120d. At the same time, distances between the points are obtained, and two candidates points wherein the dispersion is as large as possible, and the distance between them becomes as large as possible are selected as reference points 130a and 130b. Then, areas which match the images of the vicinity regions of respective reference points 130a and 130b are respectively detected from the following image frame. Then, an image is synthesized by positioning both frames in such a manner that two reference points 130a and 130b of the image frame 100 and corresponding areas of the following frame may coincide.

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a technique for generating a panoramic image from an image frame of a diagnostic image when an ultrasonic probe is moved.

[0002]

2. Description of the Related Art An electronic scanning ultrasonic probe has a limited beam scanning range.
The user moves the probe to cover the diagnostic range. Also, various techniques have been proposed for synthesizing moving image frames of ultrasonic diagnostic images obtained when a probe is moved to generate a panoramic image.

For example, it has long been known to attach an electronic scanning ultrasonic probe to a mechanical scanning mechanism equipped with a position sensor and synthesize an image frame at each time point based on the position information obtained by the position sensor. Has been. However,
This type has a problem in that the device configuration is large-scaled and it is difficult to accurately move the ultrasonic probe along the curved surface shape of the subject surface.

Therefore, in recent years, a technique for panoramic composition of image frames in a correct positional relationship by image processing has been studied.

For example, in Japanese Patent Laid-Open No. 2000-217815, the moving amount and the rotating amount between frames of a single specific area of a moving image frame are obtained, and image frames are moved to each other based on the moving amount and the rotating amount. It is aligned and synthesized. In this technique, the amount of movement of the image is obtained from the correlation between the frames of the projection distribution of the specific area. Further, the rotation amount is obtained by dividing the specific region into a plurality of partial regions in one axis direction, obtaining the movement amount from the inter-frame correlation of the projection distribution for each partial region, and obtaining from the difference in the movement amount between the partial regions. ing.

As a similar prior art, US Pat. No. 5,575,286 entitled "Method and apparatus for gen"
erating large compound ultrasound image ”.
In this conventional technique, an ultrasonic image frame is divided into a plurality of regions, the movement between the frames of each region is obtained by pattern matching, and the movement vector between the frames of the entire image is obtained from the movement of each region. Then, the image frames are combined with each other in accordance with the movement vector of the entire image. Here, the movement for each area is calculated by obtaining the corresponding area corresponding to the image of each area from the next image frame by pattern matching. Here, the movement vector of each area is given an evaluation value of quality and reliability based on the degree of pattern matching between frames and the consistency with the past movement history. The overall movement vector is obtained by weighted averaging the movement vectors of the regions.

[0007]

The conventional techniques for realizing panorama synthesis by obtaining image movement between frames (positional relationship between image frames) by using image processing have a large amount of calculation. However, there is a problem that a large amount of computer power and calculation time are required.

The present invention has been made in view of the above problems, and an object thereof is to provide an apparatus capable of realizing panoramic synthesis of ultrasonic diagnostic images with a relatively small amount of calculation.

[0009]

In order to achieve the above object, the present invention is an apparatus for synthesizing a plurality of image frames obtained by an ultrasonic diagnostic apparatus to form a panoramic image. Of the one image frame and the second image frame, a plurality of candidate point detection areas are set in the first image frame, and from each of these candidate point detection areas, the pixel having the largest pixel value difference with the surrounding pixels is selected as a candidate. From the candidate point detecting means for detecting as a point and each of the candidate points,
Reference point selecting means for selecting two reference points according to a predetermined rule considering the dispersion of pixel values in a predetermined neighborhood and mutual distances, and for each selected reference point, an image pattern in a predetermined range near that reference point The corresponding point searching means for searching the second image frame for a point having an image pattern most closely matching with as a corresponding point corresponding to the reference point, and the reference point and the corresponding point corresponding to each other coincide with each other. And an image synthesizing unit for synthesizing the second image frame by aligning the second image frame with the first image frame.

With this configuration, candidate points are detected from a plurality of candidate point detection areas set in the first image frame, and two reference points serving as a reference for frame alignment are selected from the candidate points. It Candidate points are obtained based on a relatively simple process of inspecting the difference in pixel value from surrounding pixels, and with respect to these candidate points, the variance of pixel values in a predetermined neighborhood range and the distance between these candidate points are determined. For example, a more detailed inspection regarding suitability as a reference point is performed. Thus, the corresponding points corresponding to the two selected reference points are obtained from the second image frame by pattern matching, and the first and second image frames are aligned so that the reference points and the corresponding points match, Both of these image frames can be combined in the correct positional relationship.

In this configuration, the predetermined rule used by the reference point selecting means is, for example, a rule that a candidate point having a larger variance of pixel values in the predetermined vicinity range is more likely to be selected as the reference point, or a mutual distance. A rule such that a pair of candidate points having a larger value is more likely to be selected as a reference point may be considered.

Further, in a preferred aspect of the present invention, the candidate point detection means includes a plurality of the candidate point detection areas in a range excluding a predetermined range assumed to be an area representing a skin layer in the first image frame. To set. According to this aspect,
The processing amount of the candidate point detection processing can be reduced by omitting the range of the skin layer having few features as an image in advance.

In still another preferred aspect, the candidate point detecting means is configured to detect the first image frame based on an estimated maximum movement amount of an imaging range between the first image frame and the second image frame. A plurality of candidate point detection areas are set within a range that does not deviate from the second image frame.
By limiting the candidate point detection area in this way, it can be guaranteed that the corresponding point of the reference point always exists in the second image frame.

[0014]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention (hereinafter referred to as embodiments) will be described below with reference to the drawings.

FIG. 1 is a diagram showing an outline of an ultrasonic diagnostic system including an ultrasonic panoramic image forming apparatus according to the present invention.

In the configuration of FIG. 1, the ultrasonic diagnostic apparatus main body portion 10 is provided with an ultrasonic probe 12, and this probe 1
2 is an apparatus for forming an ultrasonic diagnostic image such as a B-mode tomographic image or a Doppler tomographic image by transmitting and receiving ultrasonic waves to and from a subject by the device 2. As the ultrasonic diagnostic apparatus main body 10, a conventional general ultrasonic diagnostic apparatus can be used. The probe 12 includes an electronic scanning array transducer. By slowly moving the probe 12 in the array direction along the body surface of the subject, the electronic scanning range of the array transducer gradually moves. In response to this, a signal of a real-time image (moving image) corresponding to the current electronic scanning range of the probe 12 is output from the ultrasonic diagnostic apparatus body 10.

The panoramic image forming section 20 forms a panoramic image from the real-time image supplied from the ultrasonic diagnostic apparatus main body section 10.

The panorama image forming section 20 is provided with two frame memories 22 and 24. The frame memories 22 and 24 can hold one frame of image data of the real-time image. The frame memory 24 stores an image one frame before the frame memory 22. In the present embodiment, the positional relationship between these two images is obtained from the images of these two frames, and image composition is performed according to the positional relationship. The basic idea of this processing is to select two characteristic parts from the image of one frame, search for a part that matches each of the characteristic parts from the other image, and make sure that these matching parts match each other. Then, the two frames are aligned and combined.

For this image synthesizing process, first, the reference point determining unit 26 determines the above-mentioned 2 from the image of the nth frame (n is an integer) which is the previous frame of the two frames.
Two reference points corresponding to one feature part are determined.

FIG. 2 is a flow chart showing the processing procedure of the reference point determining section 26, and FIG. 3 is a diagram showing the manner until the reference point is determined according to this procedure.

Explaining with reference to these figures, first, the reference point determining section 26 removes a region of a predetermined width on the ultrasonic probe side as a skin cut area 102 from the image of the n-th frame and removes the image frame 100 from the image frame 100. The remaining area is vertically and horizontally divided into two, respectively, and divided into a total of four divided areas 104 (S10; see FIG. 3A).

Here, the skin part cut area 102 is a region corresponding to the skin part of the subject. Since the skin portion has a high tissue homogeneity and a high ultrasonic reflection intensity, the image becomes a homogeneous image as a whole, which is unsuitable as a region for finding a characteristic portion for matching. For this reason,
In the present embodiment, the area from the surface of the subject to a predetermined depth is set as the skin cut area 102, and this area is omitted from the detection range of the reference point in advance.

Next, each of these four divided areas 104
A portion excluding the movement margin 106 and the approach prevention margin 108 is set as the detection area 110 (S12. FIG. 3).
(See (b)). The detection area 110 is a target area for searching for candidate points that are candidates for reference points.

Here, the movement margin 106 is a margin in consideration of the movement of the electronic scanning range between two frames, and is provided for the side along the outer circumference of the image frame among the outer sides of each divided area 104. Moving margin 10
The width of 6 is set according to an assumed upper limit value of the movement amount between frames in the scanning range (this is called the maximum movement amount). This maximum movement amount is determined by the user manually operating the probe 12
It is set in advance according to the upper limit of the speed when moving the. The upper limit value of the moving speed may be arbitrarily determined and may be protected by the user. In addition, the upper limit of the probe moving speed at which a significant diagnostic image can be obtained is naturally determined in consideration of the frame rate and the like, and this may be used as a reference. By defining the detection area 110 in consideration of the movement margin 106 corresponding to such a maximum movement amount, it is ensured that the corresponding point of the point selected from the detection area 110 of one frame exists in the other frame. can do. The maximum movement amount is individually set in each of the horizontal direction (that is, the direction in which the user manually moves the probe 12) on the display screen of the diagnostic device and the vertical direction (the direction corresponding to the undulation of the surface of the subject). It is also preferable to be able to do so.

The approach prevention margin 108 is defined by each detection area 1
This is a margin for preventing candidate points selected from 10 from coming too close to each other, and is set for the side of the boundary with another detection area 110 in the outer periphery of each detection area 110.
That is, in this embodiment, 2 selected from one frame
From the angle formed by the straight line connecting the two reference points and the line connecting the corresponding points of the respective reference points obtained from the other frame,
The rotation amount of the image between the two frames is calculated. The closer the candidate points that are the candidates for the reference points are, the higher the possibility that the distances between the reference points will be closer. As a result, 1 The pixel error causes a large error in the rotation amount.
The approach prevention margin 108 is provided to avoid such a problem.

When the four detection areas 110 are set in this way, one candidate point 120a to 120d is detected from each detection area 110 (S14, FIG. 3).
(See (c)). In this process, for each pixel in the detection area 110, the pixel and its neighboring pixels (for example, 4 pixels adjacent to the top, bottom, left, and right, or 8 pixels adjacent to each other including diagonally adjacent pixels)
The pixel value difference between and is calculated, and the pixel having the maximum difference in the detection area 110 is selected as a candidate point. For example, in the case of a B-mode image, brightness is used as the pixel value, and the point where the difference in brightness from the surroundings is the maximum is obtained. Further, in the case of a difference in pixel value between a pixel and its surroundings, for example, the difference between the average of the pixel values of surrounding pixels (adjacent pixels) and the pixel value of the pixel is used, or the pixel and each adjacent pixel are used. It is conceivable to use the minimum value of the pixel value differences between and. The candidate point thus obtained is the point where the pixel value changes most rapidly in the detection area, and has strong image characteristics, so pattern matching is easy to perform.

Candidate points 12 from each detection region 110
When 0a to 120d are obtained, next, two points are selected as the reference points 130a and 130b from the candidate points (S16, see FIG. 3D). Selection is done according to the two policies of selecting points that are as characteristic as possible and two points that are as far apart as possible. As a procedure of the actual selection process, various methods can be considered depending on how to combine these two policies.

As an example, in the first stage, the three candidate points remaining in the second stage are excluded from the four candidate points except that the pixel value (luminance in the case of B mode) in the neighboring predetermined area has the smallest variance. Among the four detection areas 110, candidate points of two areas having a diagonal line (diagonal) relationship of the image frame are selected as reference points (in the example of FIG. 3D, the candidate points of the detection areas 1 and 4 are selected). (Selected as the reference points 130a and 130b). Of these, the first stage follows the policy of "picking the most characteristic points". The larger the variance of the predetermined neighborhood area is, the more characteristic portions are the points for pattern matching in the neighborhood area, and thus the higher the suitability as a reference point is. As the predetermined area near the candidate point, for example, a matrix area of N × N pixels centering on the candidate point can be used. If the matrix area is too small, the reliability becomes low, and if it is too large, the amount of calculation of the variance becomes enormous. Therefore, the size of the matrix area is determined in consideration of these trade-offs. For example, the size of the image frame 460
An example is conceivable in which 32 × 32 pixels are used as a matrix area for x400 pixels. In addition, the second step of the above-mentioned procedure is based on the policy of “selecting two points that are as far apart as possible”. That is, when the detection area 110 is set in the 2 × 2 arrangement illustrated in FIG. 3, it is better to select candidate points of two areas in the diagonal direction than to select candidate points of two areas that are vertically or horizontally adjacent to each other. The distance between the candidate points is likely to be large. A method of obtaining a distance for each combination of individual candidate points and obtaining the maximum distance among them can be considered, but the calculation amount is smaller if the diagonal direction is automatically selected.

In this method, the two candidate points in the diagonal direction have the smallest variance among the four matrix areas in the vicinity, but the variance is too small to be suitable for the pattern matching process. . Therefore, it is also preferable that a lower limit value is set for the variance of the matrix region, and if the variance of the candidate region in the matrix region is less than the lower limit value, the candidate point is not used. As a processing flow, for example, first, two points in the diagonal direction are selected from three candidate points excluding the candidate point having the smallest variance, and if the variances of the respective candidate points are both equal to or more than the lower limit value, the two A point is selected as the reference point, and if not, the remaining two candidate points excluding the candidate points corresponding to the variance equal to or lower than the lower limit value are selected as the reference points.

It can be said that the variance of three or more of the four candidate points may fall below this lower limit value, but in such a case, the same applies to candidate points below the lower limit value. It is only necessary to select another point from the detection area as a candidate point (for example, a point whose pixel value difference from the surrounding is the second place) and perform the above processing.

The reference point determination processing by the reference point determination unit 26 has been described above. When two reference points are selected from the nth frame in this way, the correlation calculation unit 28
Detects the corresponding point of each of these reference points from the (n + 1) th frame. In this case, the correlation calculation unit 28 cuts out an image in a predetermined range around the reference point 130 of the nth frame as the template 135, as shown in FIG.
Position 140 of the reference point 130 on the (n + 1) th frame
A search area 142 for the corresponding point is set around. The corresponding point should theoretically be in the range of the maximum movement amount 144 between frames from the position 140, so that range is set in the search area 142. This maximum movement amount may be the same as that taken into consideration when setting the detection area 110 described above. Then, while moving the center of the template 135 within the range of the search area 142, the correlation between the template 135 at each position and the image of the (n + 1) th frame is obtained. Then, the center of the template 135 when the correlation is highest is determined as the corresponding point corresponding to the reference point 130.

At this time, as the template 135,
The same area as the matrix area used in step S14 and having N × N pixels centered on the reference point may be used. In this case, the correlation R between the template 135 and the image in the search area 142 can be obtained by the following equation, for example.

[0033]

[Equation 1] This equation is for a B-mode image (pixel value is luminance), and Gray () indicates the luminance value. Further, i and j are indexes of columns (i) and rows (j) in the N × N matrix template 135, and n is a serial number of the image frame. Therefore, Gray (n, i, j) is the template 135
The luminance value of the pixel at the coordinate (i, j) in the (nth frame) is shown, and Gray (n + 1, i, j) indicates the luminance value at the same coordinate in the (n + 1) th frame.

The pattern matching based on the correlation calculation formula is merely an example, and other known pattern matching methods may be used.

When the correlation calculating unit 28 finds the corresponding points of the two reference points in this way, the moving amount / rotation amount calculating unit 30 then calculates the corresponding points based on the information of the reference points and the corresponding points.
The translation amount and rotation amount of the image between the nth frame and the (n + 1) th frame are calculated. As shown in FIG. 5, a corresponding point A ′ corresponds to the reference point A and a corresponding point B ′ corresponds to the reference point B.
, Respectively, first, the reference point A (xn,
yn) and the corresponding point A ′ (xn + 1, yn + 1) (xn + 1
-Xn, yn + 1-yn) is the n + th frame for the nth frame.
The amount of movement is one frame. Further, the angle formed by the line segment A′B ′ with respect to the line segment AB is the n + 1th frame for the nth frame.
Calculated as the amount of rotation of the frame. The rotation amount is represented by a sign, for example, with the counterclockwise direction as a positive direction. Therefore, the n + 1th frame is moved in parallel by the movement amount to match the corresponding point A ′ with the reference point A, and further rotated by the rotation amount to match the line segment A′B ′ with the line segment AB. The image of the (n + 1) th frame can be aligned with the image of the nth frame. In this example, the movement amount is obtained with respect to the reference point A, but even if the movement amount is obtained with respect to the reference point B, substantially the same result can be obtained.

When the movement amount and the rotation amount are obtained in this way, the image synthesizing unit 34, based on this information,
The frame image is combined with the existing panoramic image in the panoramic image memory 36. That is, the panorama image memory 3
A composite image obtained by panoramic composition of image frames obtained from the start of composition is stored in 6, and the image of the (n + 1) th frame is composed with the composite image. Here, the movement amount / rotation amount calculation unit 30 understands the relative movement amount and rotation amount of the (n + 1) th frame with respect to the nth frame, not the absolute position or inclination on the panoramic image. Therefore, in the present embodiment, the reference position information holding unit 32 stores the absolute position and inclination of the image of the nth frame on the panoramic image. By adding the movement amount and the rotation amount obtained by the movement amount / rotation amount calculation unit 30 to the stored information, the absolute position and the inclination of the (n + 1) th frame on the panoramic image can be obtained, and accordingly, the absolute position and the inclination can be obtained. The n + 1 frames may be positioned and combined. The information in the reference position information holding unit 32 may be updated by sequentially adding the movement amount and the rotation amount obtained for each frame.

In the synthesis of the image of the (n + 1) th frame with the existing image in the panoramic image memory 36, after performing the above-mentioned alignment, for each pixel in the overlapping portion, for example, the average of the pixel values of both images is taken. The processing may be performed by.

When the synthesis of the (n + 1) th frame image in the frame memory 22 is completed in this way, the image is transferred to the frame memory 24, and the ultrasonic diagnostic apparatus main body 10
The image of the next frame supplied from the frame memory 2
It is stored in 2, and the above processing is repeated.

In this manner, the panoramic image 200 is formed in the panoramic image memory 36 by sequentially synthesizing the image frames 100 of the real-time diagnostic image. And
The panoramic image 200 is displayed on the display device 40.

As described above, according to the present embodiment, the positional relationship between two reference points can be specified by performing pattern matching between frames for a relatively small area around the reference points. The panorama composition can be performed by simpler calculation than the conventional technology.
Here, as the reference points, points suitable for pattern matching are selected from the pixel value distribution of a predetermined area in the vicinity and mutual distances, so that image registration based on these reference points can also be expected to have considerably high accuracy. Also, in selecting the reference point, first, the candidate points are obtained by a simple operation of checking the pixel values with the surroundings,
Since a two-step procedure of selecting a reference point from those candidate points is taken, calculation of the pixel value variance of a predetermined area in the vicinity to see the suitability as a reference point need only be performed for a small number of candidate points. , The total amount of calculation is small.

Further, in the present embodiment, since the candidate points are selected from the range excluding the skin portion having few features as an image, the amount of calculation for detecting the candidate points can be reduced.

In the system configuration of this embodiment, the panoramic image forming section 20 is typically a reference point determining section 2.
6, the correlation calculation unit 28, the movement amount / rotation amount calculation unit 30, the reference position information holding unit 32, the image synthesizing unit 34, and the like are realized as a program and executed on a computer system. Of course, it is also possible to implement some or all of these functions as hardware. The panoramic image forming unit 20 can be built in the ultrasonic diagnostic apparatus or can be configured separately.

It is needless to say that the device configuration shown in FIG. 1 is conceptual only, and various implementations are possible within the range of this conceptual configuration. For example, the frame memories 22 and 24 can be physically realized on one memory, and further, the panoramic image memory 36.
Can also be implemented on the same physical memory. Also,
In the above example, it is described that when the composition of the image in the frame memory 22 is finished, the image is moved to the frame memory 24, but in an actual device, two memory areas for one frame are prepared,
The area for the nth frame and the n + th frame
If the area for one frame is sequentially switched and used, it is not necessary to move the image data.

In addition, in the example described above, the comparison
Although the reference point was obtained from the earlier frame of the frames and the corresponding point was obtained from the later frame, it is also possible to reverse this.

The method of narrowing down the reference points from the candidate point group in the procedure described above is merely an example. In addition to this, for example, an evaluation formula that comprehensively evaluates the variance of pixel values in a predetermined area in the vicinity of each candidate point and the distance between each candidate point is prepared, and two candidate points with high evaluation by this are prepared. It is also possible to select it as a reference point.

[0046]

As described above, according to the present invention,
Since the positional relationship between image frames can be obtained by a relatively simple calculation, a panoramic image can be formed with a small amount of calculation.

[Brief description of drawings]

FIG. 1 is a diagram showing an outline of an ultrasonic diagnostic system including an ultrasonic panoramic image forming apparatus according to the present invention.

FIG. 2 is a diagram illustrating an example of a processing procedure of a reference point determination unit.

FIG. 3 is a diagram for explaining a procedure for determining a reference point.

FIG. 4 is a diagram for explaining a search area for a corresponding point with respect to a reference point.

FIG. 5 is a diagram for explaining a movement amount and a rotation amount between frames.

[Explanation of symbols]

10 ultrasonic diagnostic apparatus main body section, 20 panoramic image forming section, 22, 24 frame memory, 26 reference point determining section, 28 correlation calculating section, 30 movement amount / rotation amount calculating section,
32 reference position information holding unit, 34 image compositing unit, 36
Panorama image memory, 40 displays.

Continued front page    F-term (reference) 4C301 EE08 JB21 JB29 JC14                 4C601 EE05 JB34 JB45 JC15 JC20                       JC21                 5B057 AA07 BA05 BA24 CA02 CA08                       CA12 CA16 CB02 CB12 CB16                       CC03 CE10 DA07 DB02 DB05                       DB09 DC05 DC22 DC33 DC36

Claims (7)

[Claims] 1. An apparatus for forming a panoramic image by synthesizing a plurality of image frames obtained by an ultrasonic diagnostic apparatus, wherein the first image frame of the first image frame and the second image frame to be synthesized is a first image frame. A plurality of candidate point detection areas are set in each of the candidate point detection areas, and a candidate point detection unit that detects a pixel having the largest difference in pixel value with surrounding pixels as a candidate point from each of the candidate point detection areas, and A reference point selecting unit that selects two reference points from the inside according to a predetermined rule that takes into consideration the variance of pixel values in a predetermined neighborhood range and the mutual distance, and a predetermined range near each of the selected reference points. Corresponding point searching means for searching from the second image frame a point having an image pattern that best matches the image pattern in the vicinity as a corresponding point corresponding to the reference point; Ultrasonic panoramic image forming apparatus including an image combining means for the reference point and the corresponding point is synthesized by aligning the second image frame with respect to the first image frame to match the. 2. The predetermined rule of the reference point selection means is such that a candidate point having a larger variance of pixel values in the predetermined neighborhood range is more likely to be selected as a reference point. Ultrasonic panoramic image forming device. 3. The ultrasonic panorama according to claim 1, wherein the predetermined rule of the reference point selection means makes a pair of candidate points having a large mutual distance easier to be selected as a reference point. Image forming apparatus. 4. The reference point selecting means selects, as reference points, two candidate points having the largest mutual distance among the candidate points having a variance of pixel values in the neighboring predetermined range larger than a predetermined lower limit value. The ultrasonic panoramic image forming apparatus according to claim 1, wherein the ultrasonic panoramic image forming apparatus is provided. 5. The candidate point detection means sets a plurality of the candidate point detection areas in a range excluding a predetermined range assumed to be an area representing a skin layer in the first image frame. The ultrasonic panoramic image forming apparatus according to claim 1. 6. The candidate point detecting means, based on an assumed maximum movement amount of an imaging range between the first image frame and the second image frame, the second image of the first image frame. The plurality of candidate point detection areas are set within a range that does not deviate from a frame.
The ultrasonic panoramic image forming apparatus described. 7. Each time the image composition is performed by the image composition means, a second image frame up to that point is newly added to the first image frame.
The superframe according to claim 1, wherein an image frame to be combined next is made a new second image frame so that the image frames sequentially supplied can be combined into one image. Sonic panoramic image forming apparatus.