Image analysis – Image transformation or preprocessing – Combining image portions
Reexamination Certificate
1999-11-24
2003-07-01
Patel, Jayanti K. (Department: 2625)
Image analysis
Image transformation or preprocessing
Combining image portions
C382S294000, C382S291000, C382S282000, C348S139000, C348S047000
Reexamination Certificate
active
06587597
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an image input method which obtains an image having a broad range by compositing a plurality of images, to an image input apparatus, and to a recording medium which stores programs for realizing the functions of this image input apparatus and which is computer-readable.
2. Background Art
Image input machines which are commonly employed, such as video cameras or the like, have a limited field of view, so that they are thought to produce images of a small part of the information of the outside world surrounding a user. When image input apparatuses are employed which are capable of obtaining images having a wide field of view in which this limited field of view is expanded, for example in the observation of the interior or exterior of a store, conventionally, a watchman was only able to view images in his immediate vicinity, whereas with these images having a wider field of view, the state of the surroundings may also be assessed quickly, so that the ability to confirm visually the state of the surroundings is dramatically improved. Furthermore, when image input devices having a wide field of view are employed, the range which may be observed by a single device is increased, and this is expected to lead to a general decrease in costs.
The main conventional image acquisition methods involve the scanning of film affixed to the side surface of a cylinder using light projected from slits moving over the opposite side surface (the so-called panorama camera) or the direct acquisition of wide-viewfield images using a fish-eye lens or a wide angle lens. The wide-viewfield images could be used as panorama pictures, or, in the case of fish-eye lenses or the like, could be passed through an optical system conducting a reverse conversion and then projected onto a screen; in order to employ the wide-viewfield images in a computer, the film or prints had to be inputted using a scanner, and this involved time and labor.
As it became possible to directly input images into a computer as a result of advances in hardware, by applying certain conversions or compositing processing to the images inputted into a computer, it became possible to obtain wide-viewfield images. Methods for the acquisition of wide-viewfield images which were precursors to such processing in computers were broadly divided into image compositing types, in which partial images taken using a standard camera were composited to produce a wide-viewfield image, and optical types, in which a wide-viewfield image was directly acquired using a wide angle lens or a mirror.
The image compositing types could be further classified into methods in which a plurality of partial images are acquired using a tripod or universal head which is adjusted so that a camera rotates thereon, and these are then composited to produce a wide-viewfield image, and methods in which partial images obtained using a plurality of cameras are composited. In the former, because a plurality of images acquired using standard cameras are composited, the resulting wide-viewfield image has a high degree of resolution; however, as disclosed in Japanese Patent Application, First Publication No. HEI 9-18750, “Electronic camera and electronic camera system employing this, photographic apparatus, signal processing apparatus”, or Japanese Patent Application, First Publication No. HEI 8-116490, “Image processing apparatus”, in order to avoid the problem of different points of view which is caused between images as a result of movements in the point of view in the conventional method, a mechanism is necessary for rotating the camera about the principal point of the lens. Furthermore, there were problems with such a method in that the operation by which the partial images were acquired was complex, the camera was heavy and large-sized, and as a result of the characteristics of the signal and power source cables, the camera could not be rotated continuously and at high speed, so that it was difficult to continuously acquire a wide-viewfield image.
In the latter methods, in which partial images obtained using a plurality of cameras were composited, cameras were arranged in a fixed manner, so that once the correct setting had been achieved, it was possible to acquire images in real time with little delay, but as a large number of cameras had to be deployed, the size of the apparatus was large, and furthermore, because the cameras were fixed, it was difficult to alter the degree of resolution or the image acquisition range. Furthermore, when the number of cameras was increased, the mechanism by which the image of each camera was inputted into the computer became elaborate.
In the optical types, a wide-viewfield image was directly acquired using a wide angle lens or mirror, and then the image was reconstituted using certain conversions. Methods have been developed which employ a fish-eye lens and which employ a mirror. In either method, a lens or mirror which is capable of acquiring the wide-viewfield image in a single image acquisition is fixed, and real-time acquisition is possible; however, there is a problem in that the degree of resolution of the wide-viewfield image obtained is low. In particular, in methods which employ fish-eye lenses, large distortions are present at the edges of images, and extremely rough images are obtained even if certain conversions are employed. There are also methods in which wide-viewfield image input apparatuses employing mirrors are combined with cameras which require close observation point images; however, although the close observation point images have a high degree of resolution, this does not solve the problem of the insufficient degree of resolution of the wide-viewfield images themselves. A method which is capable of solving the problems in the conventional wide-viewfield image input methods, and which is capable of acquiring, continuously and at high speed, wide-viewfield images having a high degree of resolution, is a method in which a plurality of partial images acquired by rotating a mirror in front of a camera are composited. If changes in the actual vertical or left-right correspondence with the image, and geometrical distortions, are excluded, then the rotation of the mirror about the optical axis of the camera corresponds to a pan operation, and the rotation of the mirror about an axis perpendicular to the optical axis of the camera corresponds to a tilt operation of the camera, so that using the rotation of this mirror, a plurality of partial images are acquired, and the partial images obtained are subjected to certain conversions, are composited, and a wide-viewfield image is acquired.
However, as pointed out in the specification of Japanese Patent Application, First Publication No. HEI 9-18750 described above, the following problems are present in image input apparatuses which input wide-viewfield images by compositing partial images obtained by causing a conventional mirror to rotate:
1. The mirror which serves to switch the photographic field is large, and
2. Differences in viewpoint are produced when the photographic field is changed by the mirror which is disposed in front of the lens.
In particular, the latter problem is one which is also caused when the camera is rotated at a point other than the principal point of the lens, and this causes problems in that the continuity of the composite image is negatively affected, and the image quality of the wide-viewfield image worsens dramatically.
SUMMARY OF THE INVENTION
The present invention was created in light of the above circumstances; it has as a first object thereof to provide an image input method which, in the case in which a camera is rotated about a point other than the principal point of the lens, and during the compositing of a plurality of partial images obtained by rotating a mirror before the camera, makes it possible to reduce the effects of parallax contained in the partial images and to input a wide-viewfield image having high image quality, as well as to provide an image inpu
Kashitani Atsushi
Nakao Toshiyasu
Kassa Yosef
NEC Corporation
Patel Jayanti K.
LandOfFree
Image input method, image input apparatus, and recording medium does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image input method, image input apparatus, and recording medium, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image input method, image input apparatus, and recording medium will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3077436