SubscribeWhy the delivery of ever greater pixel counts in digital consumer cameras does not address imaging requirements in the laboratory
Anyone thinking of buying a digital camera could easily be forgiven for continually delaying the purchase. After all, less than two years ago the typical purchase was 1.3 megapixels, last year it was 2-3 megapixels and there has been a recent flood of 4-, 5- and 6-megapixel cameras. Nor is the rapid scale up in pixels, processing power and memory confined to the traditional camera market. The same trend is already evident in the camera-phone arena.
“And a good thing too!”, did I hear you say? “If sensor manufacturers continue to develop CCD sensors with ever increasing pixel density, we’ll all be able to take better pictures, wont we”? Well, we’ll all be able to take more detailed pictures admittedly, but we probably won’t challenge David Bailey’s compositional strength.
On the face of it, though, the laboratory appears to be a different matter. Here, there is no place for the photographer’s eye. Authenticity, accuracy, clarity and precision are paramount. Is there, then, a case for taking advantage of the inexpensive consumer camera when selecting electronic imaging equipment for professional laboratory applications or is an even greater pixel count required?
Right answer, wrong question
It’s important to recognise that the push for pixel power comes because standard issue 35-mm film is so good at capturing image detail and colour intensity. Since a professional SLR camera is equivalent to a 22 megapixels digital camera, digital imaging still has a long way to go if it just a matter of raw pixel count. However, consumer SLRs and 35-mm compacts range between 9 and 12.5 megapixel equivalents and the lowly disposable camera is equivalent to 5 megapixels. What this demonstrates is that the limiting factor is the resolution of the lens used and not the film.
Like the very best camera lenses, a great deal of effort is invested in the quality and performance of microscope objectives: high resolution, good colour correction, wide field of view, and so on. Are they, then, likely to overwhelm the top-of-the-range consumer camera? And, if so, why are so many scientific cameras also 5 megapixels or less?
To answer this, we need to look at the resolving power of the modern microscope, shown in the table below.
ObjectiveNumerical ApertureRequired camera resolutionMegapixels
2.5x 0.122448 x 18434.51
5x 0.151530 x 11521.76
10x 0.251275 x 9601.22
20x 0.501275 x 9601.22
40x 0.75952 x 7170.68
63x 1.401139 x 8580.98
100x1.40714 x 5380.38
Although it may seem paradoxical, the need for high-resolution imaging is most apparent for those using relatively low magnification. Even here, a 5-megapixel scientific camera, such as the new AxioCam MRc5 (Figure 1), matches the resolving power of the most highly refined microscope optics. And the new crop of 5-megapixel consumer cameras uses similar sensor technology.
Figure 1
The consumer digital cameras are much less expensive than professional ones and they are of interest to users for that reason. The real question we should ask is, are they physically constructed to support, and does the other electronic componentry sustain, their use in microscope imaging?
Simple mounts, calibrated resolution
The first and most obvious difference between a consumer and scientific camera is that only the consumer camera has a lens. At the lower end of the market, consumer cameras have permanently integrated zoom lenses that cannot be removed. In contrast, professional digital cameras are fitted with an optical-mechanical C-mount interface that allows the use of simple adapters with defined imaging scales.
Because of their own lens systems, consumer cameras need long adapters with several re-projection lenses in order to be able to map the microscope image onto the camera's optical system. Rather than relying solely on the expensive, high quality microscope optics, it demands that a series of other lenses be placed in the lightpath and compromises ultimate image quality.
The great advantage of the simple C-mount, however, is the reproducibility of the optical adjustment. Because quantitative image analysis demands that the optical pathway remains constant after calibration, a zoom lens is a serious disadvantage. One brief accidental touch on the zoom button of a consumer camera and your previously set calibration is gone. In addition, this calibration must be reset every time the camera is switched on, whereas a C-mount camera needs only to be calibrated the very first time you set up the system.
WYSIWYG (What You See Is What You Get)
Although digital consumer cameras are incredibly powerful on the move, the lack of a direct, rapid, high-resolution live image on the PC monitor really makes its absence felt with stationary applications. Most consumer cameras offer neither a live image that can be displayed on the PC nor remote control of settings and image capture from your PC.
As a result, focusing and the selection of a relevant point on a sample can only be done via the cameras integrated LCD screen and, remember, the long extension tubes mean this may be high above the microscope and bench. Even where the camera provides a USB or FireWire output, this only shows as a small screen-in-screen on the monitor. Either arrangement is not conducive to rapid and accurate focusing on the sample whereas scientific cameras, like the AxioCam MRc5, deliver a live, full-screen image on the monitor.
A Bigger Window on the World
Although both consumer and scientific 5-megapixel cameras employ similar sensor technologies, the amount of information that they extract from an image is radically different. This is because the underlying signal processing electronics in a scientific camera collect more information about each of the five million individual pixels.
We touched on this earlier in relation to the power of film chemistry, which not only captures image detail but also records the different levels of light and colour intensity. A typical colour print shows a difference between bright and dark areas of around 1:60. In other words, it can show 60 different shades of brightness. A digital consumer camera offers a real improvement on that, storing image intensities for each pixel in ‘eight bits’ or 256 shades of grey. However, microscopy often throws up fields of view with a greater dynamic range. The camera can record the brightest 256 shades, in which case fainter signals are lost, or the darkest 255 shades, which leads to ‘bleaching’ of, and loss of information from, the brighter areas, or be set to capture somewhere in the middle and lose both brightest and faintest signals!
Most scientific cameras record light intensity in twelve or even fourteen bits. This offers a vastly superior dynamic range of 4096 resolvable differences in brightness in a 12-bit camera (16,000+ in a 14-bit camera). With a sixteen-fold greater dynamic range, the scientific camera is ideally suited to distinguishing subtle colour changes. Histopathologists regularly characterise normal and diseased tissues on their appearance and the advent of the digital camera with its superior dynamic range has been a vital contributor to increased diagnostic accuracy. The key parameters of a high quality scientific camera are shown below.
Overview of the technical features of the AxioCam MRc5
- 5-Megapixel (2584 x 1936) resolution
- 36-bit colour depth
- Dynamic range of 1:1300 for optimal brightness resolution
- Peltier cooling for minimised background noise
- Fast live image modes for easy orientation on the specimen
- Flexible read-out modes for optimal capture conditions
- Binning from 1 x 1 to 10 x 10 to increase camera sensitivity
- Read-out of regions of interest (ROI) on the sensor to define important image sections
- Integration times of 1 ms up to 60 s
- C-mount interface for easy mounting of camera on microscope
- 2/3" sensor for large field of view
- FireWire /IEEE 1394 interface for easy connection to PC
- Single-cable power supply
- Trigger In/Out signal for control of external components
- Intuitive imaging software with measuring functions for PC
Fluorescence imaging is also an area where faint signals often need to be recorded in the presence of brighter (Figure 2). Here, once again, the vastly superior dynamic range of the scientific camera is a valuable ally.
Figure 2
However, there is another aspect to their design that is equally significant – ‘On-Chip Cooling’.
Imaging faint signals requires long exposures, sometimes running into minutes. Although consumer cameras may be able to hold the shutter open for long periods, the faint fluorescence signal may be indistinguishable from the background noise that the CCD sensor emits. One way to markedly reduce the background noise in electronic circuits is to cool them and this is precisely what happens to the CCD sensors in high quality, scientific cameras.
Finally, the control and image capture capabilities of many scientific cameras like the AxioCam MRc5 are fully integrated into specialist image acquisition software. These packages save the information from the entire dynamic range, rather than cutting it down to 8 bits per colour and storing as a standard image format such as JPG, BMP, TIF, etc.
Advanced Sensor Modes
This is one area where consumer cameras really cannot keep up: no models have yet been produced that offer the special readout modes so commonly utilised by professional laboratories. These include sensor sub-region readout, which adjusts the image size to the relevant region of interest, and binning modes, in which individual pixels are aggregated to increase light sensitivity. The AxioCam MRc5 also has a special feature allowing a choice of Quality Mode (Interlaced Readout) for ultra high resolution or Fast Mode (Progressive Readout) for acquiring moving samples without necessitating a mechanical shutter.
The capability to fully integrate the camera control and image capture operations into sophisticated image analysis software, like AxioVision, offers even more advantages, especially in terms of image documentation. The software stores the objective used, filters, contrasting method, light level, exposure time, white balance, sample name/ number, notes, calibrated resolution in mm, and so on. The result is a reproducible, traceable lab operation that is a requirement of many standards.
In most laboratories, being able to transfer image data rapidly to a computer for analysis, documentation and storage is of prime importance. A camera with a fast data interface enables rapid throughput and the newer scientific cameras support fast FireWire transfers. This technical solution offers the further advantage of single-cable operation, since the camera receives its commands and draws its power from the PC along the same cable used to send back data.
Horses for Courses
Digital consumer cameras are testament to the rapid advances made in electronic imaging systems. Small, light, powerful, easy-to-use and inexpensive, they offer unsurpassed versatility to the mobile photographer. Perfectly designed for their particular niche, all be it a very large one, they are far from the optimum solution for the laboratory.
The specialist scientific camera might seem strangely unfinished by comparison, just a hole where the lens should be and a non-existent user interface. Married to the brilliant, pin-sharp images created by a modern microscope, controlled through advanced imaging software and underpinned by its own sophisticated electronics, it makes perfect sense as part of an integrated imaging solution.