Chapter 2. Biological Eye  Designs

Chapter 3. Eye
 Design Illustrations

Chapter 4. Eye 
Reproduction

Chapter 5. Optical 
 Systems Design 

A. Introduction

B. Manufactured optics 
1. Astronomy and 
surveillance
2. Stable platform for 
optical systems
3. Robotic camera 
applications
4. Flying robotics
5. Microscope and 
endoscopic applications
6. New technologies to see building blocks of cells

C. Present vision system technology approaches toward artificial eye development 

D. Integration of mans technology with 
biological eyes 

Chapter 6. The Eye Designer

Related Links

Appendix A - Slide Show & Conference Speech by Curt Deckert

Appendix B - Conference Speech by Curt Deckert

Appendix C - Comments From Our Readers

Appendix D - Panicked Evolutionists: The Stephen Meyer Controversy
 

 
 
 
 
 
 
 

EYE DESIGN BOOK
Chapter 5
Sections A and B
- Prev Page            Go to Chapter Links -             Next Page - 

(Click on PICTURE IN TEXT to bring up LARGE PICTURE)

5. OPTICAL SYSTEM DESIGNS 
LEADING TO ARTIFICIAL EYES 
A. Introduction     Present machine vision systems are very limited in overall capability as compared to most biological eyes. Comparing eye-type optical systems in the 1990's with nature's eyes is much like the Wright brothers' aeronautical science being compared to spacecraft of the late 1990's. As we learn more about biological eyes, we understand they have specific cell designs, not random chance mixtures of elements.
In the following sections we will consider optical systems that may be foundational in developing technology to design and build special purpose “eyes” that approach the capabilities of those found in nature. Reproduction and continuous repair of nature’s eyes are far ahead of the complex optical systems man has designed. Since the living cell structure is dynamic, the repair function that goes on in the eyes' retinas is more advanced than any adaptive image sensor man has been able to design. 

B. Manufactured optics
     As recently as several centuries ago, the wide spread use of lowest commercial optics such as lenses and mirrors was limited. The earliest commercial uses of optics were, of course, eyeglasses, magnifiers, mirrors, telescopes, and microscopes. After hundreds of years of optical component design and production, there are now a wide variety of commercially available optical building blocks with many choices of materials, and technologies for the development of new optical systems. As optical industries developed, manufacturing technology and efficiency enabled lower cost optical products, thereby increasing their use. Optics in television and motion picture industries gave entertainment to the world's population. 
     Early attempts to duplicate the eye's function of capturing images were limited to film cameras, which have been around for more than 150 years. As we all know, film cameras can only sense a limited number of frames or pictures, which have to be processed or developed. But video cameras, which have only been around since the 1940's, can sense a large number of scenes for recording on tape, disk, hard disk or other computer memory. These scenes can be transferred directly to a computer for image processing. Processing of electronic image information in a computer requires sensors, interface electronics, and software for recognition for what we may define as useful vision. During this century, scientists and engineers have made increasingly determined efforts to understand eyes. The software and hardware for man's vision system did not just appear, but required intelligence to develop. 
     Have we really come very far in attempting to duplicate eyes? 
     We are just scratching the surface of total biological eye technology. There are many layers of information we have not even penetrated. In trying to equal the size and function of the eyes of nature, the size of video cameras, computers, and electronic components has been greatly reduced. Image processing software and hardware has come closer to simulating eye functions, but these innovations are only small preliminary steps toward significant artificial eye design and developments. During the last 20 years, high-speed micro-circuitry, sensor arrays, arrays of micro lenses, gradient index materials, special glasses, and plastic optics manufacturing techniques are starting to give us more suitable building blocks for artificial eyes. These components are still limited to "non-living", "non-reproducible" optical systems approaching the size of eyes we find in nature. 
     The big challenge of optical products is to be affordably priced to reach large international markets. Most current attempts at creating eyes are very simplified and less versatile when compared to biological eyes that are constantly reproduced from a variety of cell types from generation to generation. When living eyes are compared to man's ability to design optical systems that repair damaged building blocks, we learn that living eyes are far advanced. 

1. Astronomy and surveillance

Magnifiers were developed around the 10th century while  eyeglasses were developed around the thirteenth century. Telescopes, using simple optics to obtain magnification over a limited field of view, were still rare in the seventeenth century while eyeglasses were becoming more common. Telescopes were further developed during the next century for more serious astronomy and military surveillance. Optical development then accelerated duringthe Nineteenth Century. Scientific and military optics developed during the Twentieth Century have helped drive down costs. Before and during World War II, there were very few significant optical surveillance developments that were related to functional eye design. Now many individuals can afford eyeglasses, telescopes, binoculars, or microscopes. The following are some illustrative optical designs developed over seven hundred years. The first picture is from approximately 1252. (Fig. 5.1 from Pg. 20 and Fig 5.2 from Pg. 1, Looking Back - An Illustrated History of the American Ophthalmic Industry, by Joseph L. Bruneni, 1994 - Optical Laboratories Association)      Before the 1940's, large telescopes were reserved for large schools, industry, and the military, because of their high cost. The present cost of telescopic systems over five inches in diameter is such that many people can now afford them. The same can be said for early high-powered microscopes. As compared to instruments available during the age of Darwin, now many people can afford to study both the huge heavens as well as our tiny cellular building blocks because of new cost-effective optical system developments.      Infrared and ultra-violet surveillance systems have been developed to work beyond normal human visual capability. These systems were used for offensive and defensive night fighting during the 1980's and 1990's. The were also used for flame detection, finding losses of energy in power systems, finding people in burning buildings, optical surveillance, astronomy, and biological studies.       Night vision systems were used effectively in the Gulf and other military operations during the 1990's. Modern infrared and ultra violet research optical systems can see more than we do with natural human image systems. Man-made infrared optical systems need special materials for optical transmission and low-temperature cooling, for good sensor response. Low temperature cooling used for current IR sensors is not needed by animals having IR vision to see targets with very little heat radiation. Cooling requirements make the overall vision system size differences even greater, between small natural and larger man-made IR vision systems. Some of man's evolutionary designs lead to commercial designs while most are not produced. Now we can display this information directly on the retina with a very small display.      Design studies of insect eyes with UV capability, and snakes with IR capability, will undoubtedly provide further insight into new technologies for the development of more useful and compact optical systems. Ultra-violet systems also require special materials to transmit short wave length light. Different optical materials are required by insects to see UV.

Figure 5-1. Historic Optics Figure 5-2. Historic Optics Figure 5-3. Historic Optics -  Early Microscope (from the Journal of The  Microscopical Society of  Southern California) Figure 5-4. Modern Optics -  Typical Microscope Design (Nikon LaboPilot-2, Microscope  Brochure, 2CEMXL2)

2. Stable platforms for optical systems
     Many of the large camera systems built for advanced study of space and for strategic military surveillance have very narrow fields of view. Man-made space camera systems that record movement, scenes, or events, generally require precision scanning and tracking ability. Vision systems are specifically designed to function from ground, ships, aircraft, and even satellites. Suitable platforms require different mountings and interfacing to be useful. Biological tracking systems are often taken for granted. Eye control design requires significant programming which requires many parallel brain interactions. A simple stable platform for optical equipment is shown on Fig.5-7. (Optical Equipment  2000 Catalog)

Figure 5-5. Modern Optics Confocal Laser Scan Microscope (Zeiss  Instrument, about 1998)

Figure 5-6. Modern Optics- 1m Mirror telescope (Pg. 47, Zeiss Instrument, No. 2, 1992-1993, Instrument No. 12)

Some satellite systems can track other satellites. Typical surveillance satellite cameras use high-resolution electronic cameras to sense and record details from the surface of the earth, and then electronically relay these pictures to earth. It took considerable analysis and design effort to conceptually evolve these intricate vision systems and their supporting platforms from the early types that dropped film to aircraft, to those that were able to transmit complex images. This illustrates the requirement for intelligence in biological eyes. Typical camera optical designs are shown in the following optical diagrams. (By Curt Deckert)       Useful magnification by a telescope for a given aperture is limited by light diffraction. Information on diffraction limits can be found in many optical texts. The larger the aperture of a high-quality diffraction limited telescope, the more it can theoretically resolve. Space telescopes, operating on platforms beyond the dense atmosphere, have optical apertures large enough to enable us to see into space much farther than man has ever been able to see from the surface of the earth.

Figure 5-7. Stable Platform  for Optical Systems  (Newport Research) Figure 5-9. Typical Meniscus  Camera Lens Optical Design

Figure 5-8. Typical 3 Element  Camera Lens Optical Design

Figure 5-10. Typical Telephoto  Camera Lens Optical Design

     Thus, beyond the light-absorbing atmosphere, we are witnessing the development of new science to describe our universe. The Hubble telescope is a prime example of a large telescope camera in space that is providing facts for updating present scientific theories. See Figure 5-11 for an optical diagram of the telescope. Many additional optics are used for the instrumentation it supports. It works in visible, IR, and UV spectral regions and is not limited by atmospheric interference. necessary to form biological eye optical designs. (ZMAX Demo Program Illustration) 

Light comes from left side and bounces off of the large curved mirror before it comes to a focus on the right side. Scientific theories change with new optical developments, as we gain more insight into the world of cell functions including the eye vision platform. This broadens the scientific foundations of  artificial optical design.

Figure 5-11 Hubble telescope  Optical Diagram

     As we understand how to build more advanced optical systems, we learn to appreciate the wide variations in eye technologies related to biological optical systems. Here we can start to comprehend the complexity and variations of the cells

3. Robotic camera application
     Early remote robotic camera systems were developed to handle hazardous devices and substances. This became necessary during the time nuclear energy sources were developed in the 1940's. These remote handling systems were direct mechanical devices using mirrors and/or windows to see from isolated viewing areas. As motor drives and TV cameras with special optics were added to mechanical systems, technicians were able to work farther away from hazardous substances, while having as much control remotely as if they were very close.
     Machine vision system containing small video cameras, or other sensors, and computer controller systems has been used since the 1970's. Now machine vision technology is using low-cost, high-resolution multicolor video cameras, special lighting, high-speed computers, and special computer software, for complex recognition purposes. As a result of these developments, many workers who previously were exposed to chemical hazards have been replaced or supplemented by robotic systems. These are examples of conceptual evolution by intelligent selection. 

Typical visual systems for robotic application are much less versatile than the eyes of living creatures. Man-made robotic systems often need specially structured lighting, special optics, or color filtering in order to function for even very limited applications. Biological systems make use of variable, but optimized density pattern of sensors. For example see Fig. 5-12 for an illustration of how resolution fall off from the center of vision of a monkey, like many other eyes. This certainly indicates an intelligent design approach to optimizing the visual usefulness of the final output of the retina. It also takes into account the useful information coming from the optical system. (Pg. 188, Neuro-Vision Systems, Ed. by Madan M. Gupta, George K. Knopf, IEEE Press, 1994)

Figure 5-12 Diagram showing  concentration of information  near the center of the eye

Advanced remote robotic systems are now being used for surgical procedures using 3-D virtual reality and the proper robotic instrument interfaces. The intelligence required for design of these optical systems and to program a computer to recognize specific subjects in a variety of remote environments is similar to that required for natural vision systems which have been scaled down in size. Here is a diagram of an advanced robotic neural vision system developed at the University of Houston. (Pg. 338 SPIE Selected papers on Model Based Vision - Vol. MS72, by R. T. Chin & C. R. Dyer)

Figure 5-13. Neural Robotic  Vision System Diagram.

     Animal eyes and brains are small and compact, compared to typical man-manufactured robotic camera and image processing systems, but they have many of the same characteristics. One can equate overall size to the level of technology. Smaller components such as computer processor chips require higher technology. The higher the technology, the greater the need for intelligent design for engineering and integration with the environment. Present intelligent robotic systems may approach the optical capability of a honeybee's vision system, but not its small size, due to the bees' bio-molecular design. For example, many small bees can sense the direction of polarization of light and can sense UV light, which typical robotic camera systems do not sense. Bees' eyes also use much less volume, to process three-dimensional information, than present man-made systems.

The thousands of interconnections connecting nature's eyes to brains are more compact than present interconnections of computer-based robotic systems. For example, compare small two-dimensional color cameras and their sensor interfaces with eyes and brains of birds that see in three dimensions. Bio-molecular design in birds allows much smaller vision systems. (Vision - High Resolution CMOS Sensors, VLSI Vision Limited, UK and San Jose, May 1998)      One can say that higher intelligence was involved in bird eyes since they have smaller pixels than our present cameras and have vision systems that process in 3D. To do this nature's eyes use many very small parallel connections, with some processing occurring between the eye and the brain.

Figure 5-14 Diagram of Small  Camera Electronic Interface  for a Vision System.

     This parallel processing allows a large number of functions to take place simultaneously, as compared to typical serial (or one operation following another) computer processing of figure. Programming eye functions to work in parallel, or at the same time, shortens the overall processing time, allowing three-dimensional recognition to occur at the speed of observation. Limited serial processing is one reason why most robotic systems have a difficult time matching the typical short reaction times of insects. 
     In comparing typical computer communications and nature's "computer communications," just imagine thousands upon thousands of very small connections coming out of a video camera and going to a computer. Even though man has come up with very small integrated circuits, where connections on a computer chip can go down to as small as 0.15 micron, the interconnections between the computer chips are dramatically larger than the insect's tiny brain. Wires for attaching major computer subsystems are much larger in diameter than many insect vision system cables. Now imagine the size and cost of a cable made up of millions of wires going from a camera to a computer to process visual data. If we take wires the size of a human hair, or about 100 microns in diameter, then 1000 of these would be approximately 100mm wide. Such a cable would be larger in diameter than small cameras or computer plugs. This gives some perspective to the image processing complexity of nature's eyes. 
     More importantly, we have to remember an eye begins with a single cell and then grows by multiplying cells, while coordinating millions of different cells for the vision process. Imagine the intelligence to design, produce, integrate, and power all the different kinds of cells that go into vision systems. Just the powering of modern flying robots is still a very significant problem. Reproduction is even more of a problem to explain without intelligent design. 

4. Flying robotics      Considerable new work is being done in this area. There are a number of insect and bird size robots being considered. Researchers are finding natural insect flight it is not a simple task to duplicate. For example, the flight of insects and birds is so complex that the smallest details provide significant engineering challenge. We have already shown that vision is very complex, but the sense of smell in insects like bees, or moths and wasps is so sensitive it could be used to sense high explosives. So people building small robots with infrared detection capabilities could learn a lot from beetles, vipers, and other small creatures.      Naturally, utilizing existing creatures would be easier than creating new robotic bugs. In any case, there appears to be considerable design effort required to build creatures that can fly and see. Their power conversion system design is well beyond our present battery technology. An example of a robotic insect is shown in Figure 5-15 and biological insects adapted for a similar purpose are shown in Figure 5-16 and 5-17.

Figure 5-17 Insect Robot  showing controls attached to  insect (Researched by  U. of Japan, reported by  Orange County Register, Friday,  June 10, 1997, News pg. 21)

Figure 5-15 Robotic Insects (NASA)

Figure 5-16 Insect Robot showing remote  controls attached to insect (SPIE  Micromachining & Microfabrication 1999  Symposium Technical Program)

5. Microscope and endoscope applications
     Microscopes have a long history of specialized use for research. Now they are used more widely in businesses, schools, and homes. Optical microscopes are able to magnify cells over 1000 times to see them reproduce, to identify abnormal cells, and to examine the design of molecular building blocks. See the optical diagram of a typical microscope on figure 5-4.
     Some new camera vision systems are designed to have ultra violet (UV) imaging capability for high technology areas, such as semiconductor fabrication. It takes very high-resolution short-wave length UV image transfer capability and analysis of quality for the production of today's integrated circuits for computers. Considerable intelligent design and production effort also went into developing the process for building very- high-resolution UV systems for semi-conductor production. 
     Even fifty years ago, people did not imagine the feasibility of producing these advanced semi-conductor computers with sub-micron features. This technology contributes to how we can produce small artificial vision systems using human intelligence. 
     Some insects have UV capabilities, which human eyes do not have. Insects are not able to see shadows or forms with nearly as much resolution as humans, but they do an effective job of rapidly sensing image motion as they travel. Much of their rapid image processing seems to take place in the form of motion sensing over part of a scene. This is quite different than processing each complete scene in full color. One attempt to control a cockroach is shown in figure 5-17. Here the man-made control mechanism is quite crude compared to the basic insect.
     In comparing IR eyes of nature to typical video IR night sight vision systems, one can see great value in the technology of nature's small compact IR eyes to provide remote night-vision. This typically allows one to visualize distant temperature sources or the temperature distribution of hot objects. Improved compact IR vision systems may be developed as a result of new viper eye research about IR vision capability. 
     During the last 200 years, straight optical bore scopes were developed to see into body cavities. These developments led to new flexible optical endoscopes containing flexible cables of glass or plastic fiber optics to transmit images. Fiber optic endoscopes are much more versatile than straight bore scopes that use optics to relay an image within a small tube.

Flexible endoscopes are typically used to detect and help correct abnormal conditions within the body or to identify conditions within closed volumes not accessible to cameras or to direct viewing. A wide variety of straight and flexible specialized endoscopic devices are now used to enter all body cavities and some blood passages to view and correct conditions inside the body. Minimal size holes are also cut in the body to gain entrance to organs a blood passage not accessed from body cavities. See the following figures for illustrations of bore- scopes and endoscopes.

Figure 5-18 Diagram of Borescope  (Olympus Marketing Literature) Figure 5-19 Endoscope Diagram (Pg. 4,  Fiberoptic Endoscopy and the difficult airways,  Andranick Ovassapian Typrocott Room, 1996)

     Although there are large UV and IR optical devices available, there are few small microscopes, borescopes, or endoscopes with high-resolution UV or IR capabilities. Multi-spectral technology would be desirable for microscopes and endoscopes used for cancer research and detection and other applications. For example, detection of many types of cancer could be made much easier if we had small intelligent vision systems that could enter the body for direct analysis of tissue, instead of taking tissue samples for evaluation using current instruments. 

6. New technologies to see the building blocks of cells
     We have been able to see major components of cells for some time, but have not been able to see details of very small cell components. New technologies are starting to let us see more of the fine details of the building blocks of nature, such as the construction of chromosomes. Special man-made optical, confocal, electron-beam, and atomic-force microscope systems can now be used to study DNA and other key building blocks of life. The small DNA features are beyond optical system diffraction limits, but electron beam and atomic force microscopes allow scientists to form images of fine details that can approach enough resolving power to sense the presence of individual atoms that make up materials. These and other new techniques may be required to see genetic DNA code information. Because of the way sample eyes have to be prepared for analysis, it is very difficult to work with live cells using electron beam microscope procedures that may require the process of coating surfaces within a vacuum chamber.
     Optical sensor resolution is beginning to approach that required to see the interconnecting biological structure that makes up the very intricate nerve cells of our retinas. New multi-color fluorescence cell analysis is starting to provide detailed information on the selective functioning of the large numbers of interconnections between the nerve cells required for vision. Vision cells interconnections may include 10,000's of connections per cell. These new discoveries are giving us new insight into the complexity of our eyes' image processing system. 
     The size of these connections is small relative to integrated circuits with conductors approximately 0.3 microns wide. This compares to molecules as much smaller building blocks than the lines. We need to remember that it takes 10,000 angstroms to equal one micron or about 3,000 angstroms to equal a thin conductor width. As a comparison, features of DNA code in cells are on the order of 10 angstroms 

A new type of near-field optical microscope, without conventional lenses, can be used to see detail as small as about 50 nanometers (.05 microns), but 1nm resolution may be required for comprehensive DNA research within cells. Technologies are not yet adequate for direct viewing of the complete DNA structure. This could be similar to viewing a large segment of a CD ROM at one time. Atomic force microscopes have to be positioned so close to samples, that they risk cell damage. These instruments are somewhat cumbersome, difficult to use, and have limited fields of view, compared to simple optical or electronic microscopes. Sensitive high-resolution microscopes and related systems require stabilization to function, as compared to typical lower resolution microscopes, working without stabilization. For illustration of their small field of view a typical image is shown, in Figures 5-20 (Photonics Design and Applications Handbook, 1996) and 5-21 (Pg. 19 Laser Focus World, Jan. 2000. This is one reason why research in this area will take so much time, and does not give wide-field visibility needed for detailed cell studies.

Figure 5-20 Image From  Near-field or Atomic  Force Microscope of  an optical coating  Figure 5-21 Image of a Wasp  Parasite from an Electron- beam Microscope image --  Image of a parasite that  lives inside a wasp.

      Rapid parallel processing by groups of cells, with thousands of connections per cell, is one of the most amazing features of the vision process. Since it is so difficult to view these parallel inter- connections, it is very difficult to design a means to reproduce the thousands of interconnections between cells. Just imagine the processing taking place in the insect parasite shown on figure 5-21 and the interconnections between cells. The programming and connections necessary for time-efficient parallel processing by the brain has to be designed and in place to allow vision. 
     Considerable design effort was necessary to form stable images for moving man-made vision systems, in order to provide feedback to control where the eye looks. This is no less true of eye stability design in nature. We are still learning how to build better, stable, moving optical systems leading to better vision system platforms. As we learn more about nature's eyes, it can help designers provide more stable structural mountings for optical systems. 
     Many different scanning eyes are used in various small animals where the space and potential resolution is limited. This requires some special processing of the visual information, especially when used with additional Eyes that do not scan. The following table illustrates several applications of scanning eyes. (Reference: table 9.2, p. 197, Animal Eyes, Michael F. Land, Dan-Eric Nilsson, Oxford Animal Biology series, Oxford University Press, 2002- Please see their book for more details )

Figure 5.21a Table Showing Examples Of Scanning Eyes