We appreciate the numerous compliments that we received on our first edition. We could have reprinted that edition but much has happened since 2007 and we felt it was important to include this new material.
Separating arrays by application or functionality would make writing this book easier. In support of this separation is that each array or camera manufacturer focuses on a particular market segment. The market could be divided into consumer, industrial, scientific, and military areas. It could be divided into low and high performance. Unfortunately, all these overlap making division difficult. As a result, some information is spread over several chapters.
The largest market is the cell phone (mobile phone) with camcorders and digital cameras close behind. The goal is to make these cameras small, low powered, and inexpensive. The high volume manufacturers consider technological advances as company proprietary information. Hence, the published literature tends to cover only scientific array technology. Nevertheless, there is sufficient information in this book to make an intelligent choice when buying these products.
In the early 1990s, some thought that CMOS detectors would replace CCDs, This did not nor will it happen. Each has it advantages and disadvantages. CMOS manufacturing capability can produce smaller-sized detectors with 1-μm square detectors as a goal. These smaller detectors impact performance. The signal-to-noise ratio is relatively low and the dynamic range is small. While even smaller detectors may be possible, the optical blur diameter will ultimately limit spatial resolution.
Chapter 2 (Radiometry and photometry) has not changed. CCD fundamentals are presented in Chapter 3. The overflow drain functions are described in more detail. It can act as a variable shutter and extend dynamic range.
Chapter 4 is the CMOS equivalent of Chapter 3. It lists the differences and similarities between CCD and CMOS sensors. Emphasis is placed on the difficulty and limitations of the shrinking detector size. Pixel structures (3T, 4T, 5T, and 6T) and associated on-chip functionality are compared with a view towards manufacturability. The higher 'transistor' (4T, 5T, and 6T) designs are for scientific applications. High volume manufacturers use the 3T design.
Array parameters (well capacity, dark pixels, microelenses, and color filter arrays) apply both to CCD and CMOS arrays (Chapter 5). Array sizes have grown over the years (e.g., HDTV and digital cameras). The optical format concept has been clarified. This chapter provides new color filter array (CFA) concepts and vertically stacked detectors. While Chapter 6 describes quantum efficiency, responsivity, and noise sources, the most important parameter is the signal-to-noise ratio. The section on dynamic range has been expanded.
Chapters 7 and 8 have not changed much. The order of Chapters 9 and 10 was reversed. Chapter 9 (sampling) now includes the Nyquist frequency for CFAs. More pictures have been added to illustrate sampling artifacts. Sampling must be considered during the camera design phase. The mathematics associated with sampling theory is complex but the human visual system is very tolerant. Aliasing, which is always present, is just ignored by many. As an example, we love our under-sampled TV.
It might initially seem odd that so many pages are devoted to video standards, CRT and flat panel displays, and printed images. The display type drives array size and camera design. Nearly every camera provides a digital output. Since digital data cannot be seen, imaging systems rely on the display medium and human visual system to produce a perceived continuous image. The display medium creates an image by painting a series of light spots onto a screen or ink spots onto paper. Since each display medium has a different spot size and shape, the perceived image will be different on each display type. Viewing distance (Chapter 11) significantly affects perceived image quality. Printers, monitors, and televisions are designed for an anticipated viewing distance. Chapter 12 has not changed.
For the mechanics, there 39 additional pages of text, 24 more figures, and 80 more references. We clarified numerous sections, deleted old material, and added the latest technological advances. We fixed the typos but probably introduced new ones.
Doug Marks, Pinnacle Communication Services, provided numerous art work. We appreciated Doug's "instant" response to our requests for drawing modifications.
February 2011
Gerald C. Holst
Terrence S. Lomheim
