WO2007043023A2

WO2007043023A2 - A novel system for capturing images

Info

Publication number: WO2007043023A2
Application number: PCT/IB2006/053760
Authority: WO
Inventors: Udayan Kanade; Balaji Ganapathy
Original assignee: Udayan Kanade; Balaji Ganapathy
Priority date: 2005-10-14
Filing date: 2006-10-13
Publication date: 2007-04-19
Also published as: WO2007043023A3

Abstract

A system and method for capturing an image through a light channel is disclosed. Light from an object to be imaged falls on one end of a reflective channel. The light exiting the other end is operated upon to recover an image of the object. Such operations are performed using optical or computational means. Applications of this invention include medical, industrial and surveillance applications.

Description

Description A NOVEL SYSTEM FOR CAPTURING IMAGES

[1] The present application claims the benefit of and priority to Indian Provisional

Patent Application No. 1294/MUM/2005 entitled "A Novel System for Capturing Images" and filed on 14^th October 2005.

Technical Field

[2] The present invention relates to a system for capturing and viewing images through a reflective channel. More particularly the present invention relates to a system and method for viewing and capturing images of distal objects through channels with a very small diameter. The system uses optical and optionally computational elements to achieve this objective. Applications of this invention include medical, industrial and surveillance applications.

Background Art

[3] A system used for viewing and capturing images through narrow channels is an endoscope. Rigid endoscopes are made using an optical system similar to that of a periscope. Flexible endoscopes can be made by employing optical fibers in addition to an optical system.

[4] Endoscopes are used for various medical and industrial purposes. Medical purposes include viewing and operating upon internal body parts. In such applications, the endoscope is inserted into the body either from a body orifice, or by making a small incision. Accordingly the size of the endoscope is kept as small as possible, to minimize the injury caused in the case of insertion through an incision/orifice, and to make more areas of the body accessible to the endoscope. In available art, endoscopes which are 5 mm in diameter can be built offering a resolution of 200 pixels in both the horizontal and vertical direction. This small diameter enables many medical applications of the endoscope.

[5] In the available art, an endoscope, after being inserted into the body, is not usually allowed to pass through other bodily barriers. For viewing beyond an internal bodily barrier, an incision has to be made inside the body. If such an incision is made during surgery, it has to be sutured up, as the diameter of the incision is too large for effective natural healing. Such suturing is a very complex procedure as it has to be performed with the help of the endoscope and other laparoscopic tools.

[6] Industrial uses of the endoscope include quality inspection and imaging for research. An endoscope used in an industrial application is called a borescope. A borescope is used to capture images through small openings such as engine valves. Alternatively, a bore may have to be created for the insertion of the endoscope. Such a bore might be created by drilling, and such a drill may be a part of the endoscope system.

Disclosure of Invention

Technical Problem

[7] The present invention is directed to a system and method for viewing and capturing images through a narrow constriction.

[8] An object of the present invention is to transfer images through a very narrow constriction.

[9] Another object of the present invention is to transfer images at a high resolution.

Technical Solution

[10] Thus according to an aspect of the present invention there is provided a system for capturing and displaying images through a reflective channel comprising:

[11] (i) a narrow channel for light being made out of reflective material such that when light falls on the distal end of the channel it is transmitted to the viewing end;

[12] (ii) means for reconstructing an image of the object near the object end of the channel by the utilization of the light emanating out of the viewing end of the channel, wherein such means is selected from an optical arrangement and computation means.

[13] The present invention provides a system and method for the capturing of images through a light guide. A light guide, as in many forms of prior art, is a rigid channel which is reflective from inside. Such light guides have been used in the prior art to transmit images, where one image pixel is transmitted per light guide. In a novel system and method provided for in the present invention, a single light guide is used to transmit an entire image. Light from objects to be viewed enters one end of light guide, such light comprising the image to be viewed, and exits the other end of the light guide after traveling the length of the light guide, to be recorded by an electronic or other light sensor. The relation between the pixels of the input image to said light guide, and the output light captured from the light guide is known at the time of fabrication of said light guide, or may be experimentally evaluated. Said relation is, or may be approximated to a high extent as a linear relation. Since the relation of the output light from the light guide to the input image pixels is linear, well known linear system solution methods may be used to estimate the input image from the output obtained from the light guide.

[14] The relationship of the output from the light guide to the input image of the light guide may be evaluated by one of many methods. One method is to computationally simulate the working of the light guide for various input pixels. Another method is to use a system which experimentally evaluates the response to each input pixel by providing various single pixel inputs. Though both these methods would be quite evident to one sufficiently skilled in the art, they are evident only after the inventive step of using a single light guide to transmit a whole image, and the inventive step of reconstructing the original image by computational means is known. Thus, these methods of evaluating the input to output relationship are also novel. Furthermore, the method of experimentally evaluating the input to output relationship may be used for each light guide, or may be used to recalibrate the computational image reconstruction means after the light guide characteristics change due to age and use.

[15] Though said light guide may have any shape, the present invention provides for certain preferred light guide shapes, having useful properties. One of these preferred shapes provides for a light guide in the form of a cylinder, having a circular profile, which is a well known light guide in the prior art. A novel method of evaluating the relation between the input and output of a cylindrical light guide is based on the principle that a single light ray going into the cylindrical light guide will primarily get distributed at the output in a conical form, the light rays forming the cone making the same angle with the axis of the cylinder as the input ray did. Furthermore, rotational symmetry of the apparatus implies that the input-to-output relationship for a cylindrical light guide may be estimated as a convolution operation per angle of incoming light with respect to the axis of the cylinder. The convolution impulse response may be evaluated by computational means or by experimental means, as described above. Thus, a novel method of reconstructing the original signal after it has passed through a cylindrical light guide is to deconvolve it using known deconvolution methods.

[16] Another preferred shape for a light guide is a rectangular parallelepiped with a rectangular profile (henceforth referred to as rectangular light guide). A novel method of evaluating the relation between the input and output of a rectangular light guide is based on the principle that a single light ray going into the rectangular light guide will primarily get distributed at the output into four outgoing rays, these output directions characterized by the rule that each outgoing ray bears the same or opposite angle with each wall of the rectangular light guide as the input ray does with the same wall. The pattern of distribution in these four directions may be computationally or experimentally evaluated, as described above. Thus, a novel method of reconstructing the original signal after it has passed through a rectangular light guide is to solve a linear system for each quadruple of output ray directions, such output ray directions being produced by four input ray directions in the manner described.

[17] One of the basic advantages of the above preferred shapes and other shapes is the computational ease of solving the linear input-output system to reconstruct the original image. Unfortunately, because of interference, diffraction, and light guide imperfections, the above mentioned reconstruction methods, in some cases, may only act as approximations. For more precise reconstruction, well known iterative inversion methods, where a part of the system is easy to invert, and another part of the matrix is is easy to apply may be used. Though these inversion methods are well known, their use, with the part of the system that is easy to invert being the approximations mentioned above for respective preferred light channel shapes, for the present reconstruction problem of the present invention is novel.

[18] Also provided for in the present invention are many optical image reconstruction means which simplify or make redundant the job of the computational image reconstruction means. A simple optical image reconstruction system is an arrangement of focusing apparatuses which focuses the output light from the light guide onto the image sensor.

[19] A particular optical image reconstruction means relates to the rectangular light guide. A novel system of mirrors and half-mirrors is provided to optically reconstruct the input image of the light guide. This system merges all four output light rays relating to a particular input light ray into a single light ray. The other light rays also relating to the same four output light rays are either blocked at the input, or blocked at the output, or cancelled by destructive interference.

Advantageous Effects

[20] A particular advantage of the present invention is that images may be captured through very narrow channels. These channels may be small enough to be able to be pushed through tissue of a living object without causing substantive harm.

[21] Another advantage is that images are captured of a relatively high resolution.

Brief Description of Drawings

[22] Figure 1 illustrates the schematic diagram of system of the present invention for viewing and capturing images.

[23] Figure 2A illustrates a rectangular light channel in accordance with a preferred embodiment of the invention.

[24] Figure 2B illustrates a cylindrical light channel in accordance with a preferred embodiment of the invention.

[25] Figure 3 illustrates the ray diagram of the path taken by a single light ray inside a hypothetical two dimensional channel for the purpose of illustration of a principle on which present invention is based.

[26] Figure 4 illustrates the ray diagram of the path taken by a plane wave inside a hypothetical two dimensional channel for the purpose of illustration of a principle on which present invention is based.

[27] Figure 5 illustrates an apparatus to experimentally evaluate the input output relation of light channel in accordance with an embodiment of present invention.

[28] Figure 6 illustrates an optical arrangement for the purpose of merging light rays emanating in two different directions into a single light ray. [29] Figure 7 illustrates an apparatus for merging light rays emanating from a rectangular light channel, the apparatus being an optical image reconstruction system.

[30] Figure 8 illustrates an arrangement for illuminating a distal object using the same light channel of the present invention.

[31] Figure 9 illustrates a light channel with an aperture having diameter less than the diameter of the light channel.

[32] Figure 10 illustrates an apparatus used to augment a flexible fiber endoscope with a very narrow channel at the distal end according to a particular embodiment and use of the present invention.

Detailed Description

[33] Fig. 1 is a schematic diagram of the invention in accordance with a preferred embodiment of the invention. A distal object 102 is to be viewed. Light from the distal object 104 enters the distal end 105 of light channel 106. Light 108 exits the viewing end 107 light channel 106 and goes through an optional optical image reconstruction system 110. Light 112 emanating from optical image reconstruction system 110 then falls on an image sensor 114. In an embodiment of the present invention, the image sensor 114 is the human eye. In another embodiment of the present invention, the image sensor 114 is an electronic image sensor. In case that the image sensor 114 is an electronic image sensor, the data from the image sensor is fed to a computational image reconstruction system 116 and the image reconstructed by this system is displayed on a monitor 118. It would be obvious to a person sufficiently skilled in the art that the monitor 118 may be replaced by other image display mechanisms such as a projection system or a printer.

[34] For the invention to meet its objective of imaging through a narrow channel, the light channel 106 is a channel with a very small diameter. Channels of diameters approximately 0.5 mm or even smaller may be used. Various ways of constructing the light channel 106 are depicted in Fig. 2 A and Fig. 2B. Fig. 2 A shows a light channel having a rectangular cross section 204, and thus forming a rectangular parallelepiped 202 and Fig. 2B shows a light channel having a circular cross section 210 and thus forming a cylinder 208.

[35] The inner wall 206 and 212 of a light conduit is made in such a manner that it will reflect light. According to one manifestation of the preferred embodiments, such reflection may be achieved by using polished metallic surfaces. Alternatively, it may also be achieved by using the principle of total internal reflection at the junction of two transparent media.. It would be obvious to a person skilled in the art that the present invention can be used with light channels of shapes other than the ones depicted in Fig. 2A and Fig. 2B.

[36] Fig. 3 depicts a ray of light 306 entering a hypothetical 2 dimensional light channel 300 at a particular angle. It is obvious that the light ray will exit in one of two angles as depicted in cases 302 and 304. The outgoing ray will make the same or opposite angle with the axis of the light channel.

[37] Fig. 4 depicts a plane wave illuminating the distal aperture of the hypothetical 2 dimensional channel 300. The energy of the plane wave is split into two outgoing beams of light 404 and 406. This splitting of the energy need not be into equal parts.

[38] In a similar fashion to a ray of light coming out in one of two directions depending upon the direction of the ray of light going in for a hypothetical 2 dimensional light channel 300, a ray of light going into a light conduit with rectangular cross section 202 will come out in one of four directions, these four directions depending only upon the direction of the ray of light going in. These four directions may be characterized by the fact that the ray of light coming out will make either the same or the opposite angle with each of the walls of the light channel 202 as the angle that it made with the same respective walls while going into the light channel 202.

[39] In a manner similar to Fig. 4, if a light conduit with rectangular cross section 202 is illuminated with a plane wave, the energy of the beam will be split into four outgoing beams, such splitting not necessarily being in equal. The pattern of such energy splitting will be fixed, dependant only on the direction of the incoming ray of light.

[40] Similarly, a ray of light going into a cylindrical light conduit 208 will come out in one of numerous directions, these directions being characterized by the rule that the outgoing ray of light will bear the same angle and distance relationship with the axis of the cylinder as the ray of light going in. If a cylindrical light conduit 208 were illuminated by a plane wave, the rays of light coming out would approximately form a cone. The energy of the incoming ray will be distributed over this cone, such distribution being fixed for an input direction, but not necessarily uniform.

[41] For light channels of shapes other than rectangular parallelepiped or cylinder, an incoming plane wave (beam of light) will disperse while going out in a fixed pattern, such pattern being dependant on the direction of the incoming beam and the shape of the light channel.

[42] Fig. 5 depicts an apparatus to evaluate the input output relation of light channel

106 in accordance with an embodiment of present invention. A light generator of variable orientation 502 is used to shine light 104 onto the distal end 105 of light channel 106. Light 108 coming out of light channel 106 is focused using focusing apparatus 504 onto a light sensor 508. Data from the light sensor 508 is stored in storage 510 for various orientations of incoming light.

[43] It is possible that the input-output relation of a light channel cannot be perfectly determined at the time of designing the light channel, but varies with individual light channel and with time. If the input-output relation of a light channel changes with time, the apparatus of Fig. 5 can be provided to the user of the light channel to periodically calibrate the estimation performed by the image reconstruction systems. If the input- output relation of a light channel changes with the individual light channel, either the apparatus of Fig. 5 can be provided to the user of the light channel, so that the user may perform calibration per light channel used, or calibration could be performed using said apparatus by the manufacturer, and sent along with the individual light channel as data for the particular light channel. In the case that the data generated by Fig. 5 is going to be used to only calibrate the computational image reconstruction system 116 and not the optical image reconstruction system 110, it may be beneficial to include the optical image reconstruction system 110 in place of focusing apparatus 504, and the sensor 114 in place of any special sensor 508. Thus, the input-output relation directly from the distal end of the light channel to the sensor 114 can be estimated, which is the relation which has to be inverted by the computational image reconstruction system 116.

[44] Hereinafter is described the principle of working of the computational image reconstruction system 116. The computational image reconstruction system 116 is employed to generate an estimate of light 104 falling on the distal end 105 of light channel 106 from a description of the light 112 sensed by a light sensor 114. The computational image reconstruction system 116 uses mathematical computational means to achieve these objectives. These computational means may be implemented in a computer or in electronic or optical circuitry.

[45] The computational image reconstruction system 116 calculates the inverse relation of the input-to-sensor relation from the light 104 going into the distal end 105 of the light channel 106 to the light 112 going into the light sensor 114. Since this relationship is usually expected to be linear, many well known matrix inversion methods may be used to calculate and evaluate the inverse relationship.

[46] Specialties of the particular light channel shape may be incorporated in such inversion procedures. For example, as explained hereinbefore, the rectangular light conduit shape 202 will cause an incoming light beam to affect four different spots on the light sensor 114. (We are assuming here that the optical image reconstruction system 110 is absent or in the form of a simple light focusing arrangement.) Furthermore, there will be three other light beam directions that will affect the same four spots on the light sensor 114. The computational image reconstruction system 116 can thus look at the light received at only these four spots and estimate the values of the particular four incoming directions. It will do this for each four input spots related in the same, and thus reconstruct the whole image. As another example, as explained hereinbefore, the cylindrical light conduit shape 208 will cause an incoming light beam to affect a circular locus on the light sensor 114. Furthermore, there will be a whole cone of light beam directions that will affect the same circular locus on the light sensor 114. Furthermore, the symmetry inherent in the cylinder dictates that the relationship between the intensities of the incoming cone of light and the sensed values on the light sensor 114 are related by a mathematical operation called convolution. The computational image reconstruction system 116 can use many known deconvolution methods to obtain an estimate of intensities for the particular cone in question. It will do this for each incoming cone of light, each distinct cone causing a distinct circular locus on the image sensor 114. For certain combinations of channel shape and light wavelength, diffraction and interference effects cause a considerable deviation from the behavior predicted in this paragraph. In such cases, the methods described above may still be used as approximations, as initial estimates, or as estimation steps in converging iterative algorithms. As a specific enabling example of the latter, it is considered the input-output relationship as made up out of two parts, one which is easily invertible (by the methods specified above) and one which is not. Many methods are found in the literature for iteratively solving matrix problems where the matrix can be decomposed into parts, one of these parts being easily invertible.

[47] Hereinafter is described the principle of working of the optical image reconstruction system 110. The optical image reconstruction system 110 is employed to simplify the task of the computational image reconstruction system 116 by transforming light 108 emanating out of the light channel 106 into a form which is suitable or easier for the computational image reconstruction system 116 using optical means. Said optical means may include lenses, mirrors, half silvered mirrors or other optical means. For example, the optical image reconstruction system 110 may focus parallel rays of light onto particular points on the light sensor 114. In particular embodiments of the present invention it is possible to perform the image reconstruction entirely by optical means which obviates the need for the computational image reconstruction system 116. In such cases, it is also possible to view the image directly with the eye, without use of the sensor 112 and monitor 118.

[48] Fig. 6 depicts an optical arrangement for the purpose of merging light rays emanating in two different directions into a single light ray. Light rays 602 and 604 emanate in two directions from light channel 600. Mirrors 606 and 608, a half silvered mirror 610 and means for absorbing light 612 are employed to merge both the rays 602 and 604 into a single light ray 614. This optical arrangement will merge any two rays bearing the same relationship with each other that light rays 602 and 604 bear with each other. Namely, the said two light rays should together lie in a plane perpendicular to the half silvered mirror, and should subtend opposite angles at the plane of the half silvered mirror. Since this is exactly the relationship borne by the emanating light rays in the cases 302 and 304 for hypothetical two dimensional light channel 300, the ar- rangement of Fig. 6 may be used to merge the distal ray 306 back into a single ray 614. Rays of all orientations can be merged using this single apparatus. In Fig. 3, since there are two distinct input orientations which will give the same two output angles, it is important to block one of these rays at the distal end of the channel, which can be done with light absorbing material placed at such a position as to occlude one of each pair of incoming light rays related in this fashion. In a particular modification of the present part of the invention, a plane wave 402 coming into the distal end of the light channel at a particular orientation, the input-output relation will be known, i.e. the relative intensities of the split beam 602 and 604 will be known. Hence, the reflectance, transmittance and thickness of half silvered mirror 610 is chosen such that only one of the beams 614 will remain and the other beam 616 will get nullified due to destructive interference. For another plane wave coming into the distal end of the light channel and producing the same two rays 602 and 604, the relative intensities of the light rays 602 and 604 will be interchanged, causing the light beam 614 to nullify and the other beam 616 to have intensity. Different incoming orientations will have different relative intensities for the splitting of the light beam; thus different spots of the half silvered mirror 610 will be designed with different transmittance and reflectance properties.

[49] In the foregoing description, the hypothetical two-dimensional channel 300 is used for the purpose of illustration. The same principle of optical image reconstruction is used to reconstruct an image obtained from rectangular parallelepiped channel 202. For example, two optical arrangements of the kind shown in Fig. 6 may be used to merge the four outgoing ray directions related to a particular incoming ray direction.

[50] The apparatus shown in Fig. 7 is a single optical arrangement which will achieve the same results. 702 is the view of the apparatus looking into the light channel, whereas 704 is a side cross-sectional view of the same. Rectangular light channel 712 leads into a mirror cavity 706 with four mirrors, and two half silvered mirrors at right angles, 708 and 710. (It is possible that the rectangular parallelepiped 712 is itself used as the mirror cavity 706.) This arrangement will merge the various outgoing angles of an incoming ray of light into a single ray of light. To avoid interference from other related incoming rays of light, either these rays of light will be blocked at the distal end of the light channel, or the reflectance and transmittance of the half silvered mirrors 708 and 710 will be adjusted at various spots in such a manner that the interference is cancelled by the principle of destructive interference. It would be obvious to one skilled in the art that various optical means for partially transmitting and partially reflecting light may be used in place of half-silvered mirrors.

[51] It is thus possible to reduce or eliminate the responsibility of the computational image reconstruction system 116 by using optical methods. In cases where the optical image reconstruction system reconstructs a replica of the incoming image which is close enough to the optimal replica to be used for the purpose that the invention is employed for, the image may optionally be viewed directly by the eye, or projected onto a screen, or captured by a light sensor and directly displayed on a connected monitor.

[52] In many applications of viewing and capturing images through narrow channels, there is no natural illumination for the distal object whose image is being captured. In the known art, such illumination may be provided by a separate light guide that reaches the object. This known art increases the total size of the probe entering the narrow constriction, and also causes sideways illumination, forming shadows and an imperfect picture. Fig. 8 depicts an arrangement for illuminating a distal object using the same light channel 106 that is used to capture the image of the distal object. Lens 804 and partially silvered mirror 806 focus light from a light source 802 into the viewing end

107 of the light channel 106. This light emanates out of the distal end 105 of the light channel 106 and illuminates the distal object 102. Light reflected and dispersed from the distal object 102 enters the distal end 105 of the light channel 106 and light 108 leaves the viewing end 107 of light channel 106 after part of it gets reflected away by the partially silvered mirror 806. The light 108 leaving the light channel may be operated upon to reconstruct an image of the distal object as described hereinabove. Light absorption means 808 and 810 are used to avoid contaminating the image light

108 with unwanted light.

[53] A particular means to achieve color image reproduction related to the above is to alternately send red, green and blue light into the light channel. The images captured by the sensor 114 will correspond to the red, green and blue images which make up the color image to be reproduced on the monitor. This will reduce the cost of the image sensor by allowing a non-chromatic image sensor, and also make it possible to have higher resolution and lower spatial aliasing in the image captured.

[54] Light diffraction and interference can cause chromatic aberration. The narrower the range of frequencies of light being captured, the lesser the chromatic aberration. Light of more than three different spectra can be sent alternately to reduce chromatic aberration. Each light spectrum will have associated with it its own input-output relation, which will be used to estimate the image by the computational image reconstruction system 116. These images can eventually be merged into the red, green and blue images.

[55] Referring to Fig. 4, it is seen that for almost any incoming light direction, there will be multiple outgoing beams of light. This effect can be reduced by limiting the light entering the distal end of the light channel by using an aperture which is smaller than the opening of the light channel. This arrangement is shown in Fig. 9. Using the arrangement of Fig. 9 light of a large number of incoming directions will emanate from the viewing end of the light channel in only one particular direction. Similar effect can be achieved by augmenting the viewing end of the light channel with a smaller aperture. Since there will be lesser optical interference from various incoming directions, the optical reconstruction in this case is a simpler task. For example, very simple optical image reconstruction means might be used. An optical system which will separate light such that the input and output orientations are exactly the same from light such that the input and output orientations are inverted in one or both of the principal axes of the rectangular cross section of a rectangular parallelepiped light channel can be constructed. For example, a particular such system will be transparent in places where the former kind of light shines, and reflective in places where the latter kind of light shines. Using such an arrangement of patterned mirrors, the whole image may be correctly reconstructed. Alternatively, for a low cost and simpler solution, only the former kind of light could be construed to be a close approximation of the real image, and the latter kind of light would be blocked.

[56] Referring to Fig. 3, it is seen that the higher the angle of the incoming light ray, the more the number of reflections that it sustains before emanating out of the viewing end of the light channel. In practical cases where the mirror is not perfectly polished, each reflection will cause a small distortion in the angle of the ray. Such distortions will add up and might make it impossible to estimate the incoming light from the emanating light from the light channel. Furthermore, such distorted light might seep into areas of the image which it would otherwise be possible to estimate. Another effect that would be additive to this effect in hampering the estimation of light that it would otherwise be possible to estimate, is a fraction of the light suffering diffuse instead of directional reflection at each reflection. Such seepage of light should be avoided, and thus light undergoing a high number of reflections before emanating from the light channel should be cut. There are many optical systems that can achieve this effect. One such method is to augment the distal end of the light channel with means to optically block light of more than a particular angle. A lens may be added at the distal end so that a light field of wider angle may be compressed into a light field of narrower angle.

[57] Another means that achieve cutting of light undergoing a large number of reflections is a light channel such that the mirrored walls of the light channel are slightly absorptive. At each reflection, a small amount of energy will get absorbed, and the higher the number of reflections, more the energy that will get absorbed.

[58] It is well known in the art that for the computational reconstruction of a signal, it is beneficial to have a large number of measurements of the signal. Furthermore, it is beneficial if each of these measurements is a different function of the signal that is to be reconstructed. There are many well known methods for estimating a signal given more measurements of the signal than the number of data points (samples) that are to be reconstructed. For example, if the said functions are linear, then least squares estimation methods may be employed. Thus, it would be beneficial, for example, to not only know what are the intensities of light 112 on the surface of a two dimensional screen, but to also know what are the intensities of light throughout a three dimensional volume through which it travels. To perform such a measurement, we can logically divide the volume through which light 112 travels into many slices, and measure what is the intensity of light on the surface of each of these slices. An optical arrangement that will achieve this result is a stack of partially silvered mirrors, each partially silvered mirror reflecting the light onto a different sensor. Another optical arrangement is a stack of LCD panels, each LCD panel amenable to being made transparent or a means of scattering light. At any time, only one LCD panel will be made to scatter light, thus acting as an intermediate screen. The image formed on this screen will be captured by a light sensor and lens arrangement. The whole measurement data will be captured by making each of the LCD panels scatter light in quick succession, and storing the results for each LCD panel separately. Computational image reconstruction system then utilizes data captured thus to produce an estimate of the original image.

[59] The present invention can be used to augment a medical or industrial endoscope to have a very narrow channel at the distal end. The rest of the endoscope could be a fiber optic scope, and thus will be flexible. The advantage of a very narrow channel at the distal end would be added accessibility or the possibility of capturing images through extra incisions with very little harm to the body. One arrangement of achieving this is depicted in Fig. 10. Light 108 emanating from very narrow light channel 106 is made to fall on the distal end of a bunch of optical fibers 1002. Light emanating from the other end of these optic fibers is sensed and the original image will be reconstructed by computational means. Additional optical image reconstruction systems may be employed between light channel 106 and optical fibers 1002 as well as between optical fibers 1002 and the image sensor.

Industrial Applicability

[60] Endoscopes have many uses in medicine. Since the present invention can be used as an extremely thin endoscope, this will increase the usability and suitability of endoscopes for many purposes. For example, some incisions in a patient's body can be made in an outpatient setting. Such incisions, usually less than 0.5 mm in diameter, do not need any post-operative care as the natural human healing system heals the wound. Reducing the size of an endoscope to 0.5 mm will allow outpatient usage of the endoscope for medical purposes, including investigative and emergency purposes.

[61] In an alternative embodiment, disclosed above, the narrow channel of the present invention augments a flexible endoscope, thus providing an extremely narrow tip to a flexible endoscope. This narrow tip will allow the endoscope to pass through extra incisions after the primary incision or insertion with much ease. The very minor extra incision will get healed due to the body's natural healing processes, thus reducing the complexity of the surgical operation and post-operative care and risks. Very hard to reach places such as the alveoli or the capillaries may also be imaged this way.

[62] An borescope with a smaller diameter than a regular borescope will allow lesser destruction of the item being drilled and imaged. It will also allow insertion through smaller holes than is possible with current technology.

[63] Concealment of imaging devices has important applications in surveillance. Though very small imaging devices are available, for complete concealment systems such as one-way mirrors have to be used. An imaging device with a very small diameter would be very easily concealed. Furthermore, an imaging device with very small diameter can enable new surveillance methods such as pushing a surveillance device discretely through wood walls or other material.

[64] A method and system for capturing images through narrow channels are disclosed.

It is understood that the embodiments described herein are for the purpose of elucidation and should not be considered limiting the subject matter of the present patent. Various modifications, uses, substitutions, recombinations, improvements, methods of productions without departing from the scope or spirit of the present invention would be evident to a person skilled in the art.

Claims

[I] An apparatus comprising a channel comprising a transparent interior and reflective walls, and a system that reconstructs an image of an object near one end of said channel from light which comes out of other end of said channel.

[2] The apparatus as recited in claim 1 wherein said reflective walls are made of a material that reflect light.

[3] The apparatus as recited in claim 1 wherein said reflective walls are made of a material of refractive index lower than that of said transparent interior.

[4] The apparatus as recited in claim 1 wherein said channel is shaped in the form of a rectangular parallelepiped.

[5] The apparatus as recited in claim 1 wherein said channel is shaped in the form of a cylinder.

[6] The apparatus as recited in claim 1 wherein said system that reconstructs said image comprises an image sensor and computational means to reconstruct an image of said object from a record of the image captured by said image sensor.

[7] The apparatus as recited in claim 1 wherein said system that reconstructs said image comprises means to sense the intensity of light at various points in a three- dimensional space and computational means to reconstruct an image of said object from a record of said intensity of light.

[8] The apparatus as recited in claim 7 wherein said means to sense the intensity of light comprise at least one partially silvered mirror and a plurality of image sensors.

[9] The apparatus as recited in claim 7 wherein said means to sense the intensity of light comprise a plurality of panels each of which has a state in which the panel is transparent, and a state in which the panel disperses light.

[10] The apparatus as recited in claim 1 wherein said system that reconstructs said image comprises an optical image reconstruction system comprising optical means that transform light which comes of said other end of said channel.

[II] The apparatus as recited in claim 10 wherein said optical means comprises a light focusing apparatus.

[12] The apparatus as recited in claim 10 wherein said optical means comprises at least one partially silvered mirror that merges beams that a single beam entering the channel splits into.

[13] The apparatus as recited in claim 12 wherein said partially silvered mirror has reflectance, transmittance and thickness such that the optical effect of only one of the original beams that said beams have split into remains in one outgoing beam. [14] The apparatus as recited in claim 10 wherein said optical means comprises at least one sheet which is a mirror in certain places and transparent in other places. [15] The apparatus as recited in claim 10 wherein said optical means comprises at least one sheet which is transparent in certain places and absorbs light in other places. [16] The apparatus as recited in claim 1 wherein said apparatus comprises an apparatus that blocks some light rays before they enter said channel. [17] The apparatus as recited in claim 1 wherein said apparatus comprises an apparatus that blocks some light after they exit said channel. [18] The apparatus as recited in claim 1 wherein said apparatus comprises a lens situated between said object and channel. [19] The apparatus as recited in claim 1 wherein said apparatus further comprises means to illuminate said object at said one end of the channel by sending light through the said channel. [20] The apparatus as recited in claim 19 wherein said apparatus further comprises a partially reflecting mirror arranged so that illumination light may enter said channel and light exiting the channel may be used for image reconstruction purposes. [21] The apparatus as recited in claim 19 wherein said means illuminates said object successively by light of various spectral composition. [22] The apparatus as recited in claim 1 wherein said reflecting walls absorb a small fraction of the light incident on them. [23] The apparatus as recited in claim 1 wherein the apparatus further comprises a plurality of light conductors to carry the light which comes out of said other end of channel to said system that reconstructs an image of said object. [24] The apparatus as recited in claim 23 wherein said light conductors are optical fibers. [25] The apparatus as recited in claim 23 wherein situated between said other end of channel and said light conductors is an optical image reconstruction system. [26] An apparatus comprising a light source capable of illuminating at various orientations one end of a channel comprising a transparent interior and reflective walls, and a light sensing mechanism capable of sensing and recording light emanating at various orientations from the other end of the channel. [27] The apparatus as recited in claim 26 wherein an optical image reconstruction system is situated between said other end of said channel and said light sensing mechanism. [28] A method of reconstructing an image of an object near one end of an apparatus comprising a channel from image captured of light which comes out of other end of said apparatus, comprising calculating a function of said image, said function being an approximation of the inverse of the function relating image of said object to said captured image. [29] The method as recited in claim 28 wherein the step of calculating said function further comprises the step of estimating the function relating image of said object to said captured image. [30] The method as recited in claim 29 wherein the step of estimating said function comprises estimating said function by simulating said apparatus by computational means. [31] The method as recited in claim 29 wherein the step of estimating said function comprises estimating said function by illuminating said channel of said apparatus at various orientations, and recording image captured of light which comes out of other end of said apparatus. [32] The method as recited in claim 29 wherein the step of estimating said function comprises estimating said function as a linear function. [33] The method as recited in claim 28 wherein the step of calculating said function further comprises calculating said function as an inverse function of a linear function. [34] The method as recited in claim 28 wherein the step of calculating said function further comprises grouping all points in said image into multiple groups, and calculating a function of points in a particular group. [35] The method as recited in claim 34 wherein at least one said groups is a group of four points. [36] The method as recited in claim 34 wherein at least one said groups has a circular locus in the image. [37] The method as recited in claim 28 wherein the said captured image comprises record of light falling on a plurality of planes.