Tutorial

Administrator

Description

zzzScientific Instrument Software Corporation Limited Micrometrics™ Imaging Guide This paper discusses the key points to operate a Micrometrics™ camera using Micrometrics™ SE Premium software. Camera Manipulation Criterions of a “Good” Image Overall Brightness Color Reproduction Noise Level Considering adjustable factors only, the overall brightness of an image is a product of the following: 1. Illumination Level 2. Exposure Time 3. Gain For purpose of efficiency (frame rate or time needed to capture an image), illumination should be adjusted first. Whereas a desirable overall brightness cannot be achieved by adjusting illumination alone, exposure would then be adjusted, starting from the maximum allowed by the constant frame rate prerequisite. Gain adjustment must be attempted as a last resort, because a gain value greater than unity amplifies noise as well as brightness of the image. To avoid flicker introduced by light source, select “60Hz” in US and “50Hz” in Europe. If the light source does not flicker, select “DC” for the finest tuning of exposure time. Exposure setting may be automated by “Auto Exposure” or “Area AE” (Area Based Auto Exposure). The former is based on the average brightness of the latest frame in video stream (before it is processed by SDK functions such as Gamma, Saturation or LUT). The latter is based on a neighborhood therein as indicated by subsequent mouse click. [e-mail] Scientific ...

Informations

• Publié par: Effo
• Nombre de lectures: 175
• Langue: English

Extrait

Scientific Instrument Software Corporation Limited

Micrometrics™ Imaging Guide

This paper discusses the key points to operate a Micrometrics™ camera using Micrometrics™ SE Premium software.

Camera Manipulation

Criterions of a “Good” Image

zOverall Brightness zColor Reproduction zNoise Level Considering adjustable factors only, the overall brightness of an image is a product of

the following:

1.Illumination Level 2.Exposure Time 3.Gain For purpose of efficiency (frame rate or time needed to capture an image), illumination should be adjusted first. Whereas a desirable overall brightness cannot be achieved by adjusting illumination alone, exposure would then be adjusted, starting from the maximum allowed by the constant frame rate prerequisite. Gain adjustment must be attempted as a last resort, because a gain value greater than unity amplifies noise as

well as brightness of the image. To avoid flicker introduced by light source, select “60Hz” in US and “50Hz” in Europe. If

the light source does not flicker, select “DC” for the finest tuning of exposure time. Exposure setting may be automated by “Auto Exposure” or “Area AE” (Area Based Auto Exposure). The former is based on the average brightness of the latest frame in video stream (before it is processed by SDK functions such as Gamma, Saturation or LUT). The latter is based on a neighborhood therein as indicated by subsequent mouse click.

[e-mail]

Scientific Instrument Software Corporation Limited

“Continuous Auto Exposure” is “Auto Exposure” applied by SE Premium in regular time

intervals. If the result is brighter than expected, you can set “Bias” to a negative value.

If the result is darker than expected, you can set “Bias” to a positive value. Valid “Bias” values are in the range [-30, 30]. Automated exposure adjustment is based on light source frequency settings. For example, if “50Hz” has been specified, the resulting exposure time from auto exposure is always a multiple of 10ms. There are two sets of gain settings: one is the global gain, which amplifies all three channels of a camera simultaneously by the same factor; the other is color gains, which works channel-wise, amplifying each channel individually by a possibly different factor. Note that the end effect is that both sets of gains contribute to the brightness of the

images. For example, by default, the color gains are set to 2 for X18 cameras. If you set

these gains to 3, you will be capturing images 50% brighter. The color gains are mainly useful in color reproduction. A camera attempts to match the

“real” colors of an object by adjusting the relative strength of its red, green and blue channel gains. In SE Premium, green channel gain is set as the basis and the other two channel gains will be modified by white balance operations. So if the green gain is set to a greater value, in general you will arrive at a brighter image. In all case, if the color gain setting as the result of white balance is within the allowed range, colors will be

reproduced correctly. The accuracy of color reproduction is determined by the step size of color gains. For X18

cameras, the color gain increment is 0.125 in the range. For X90 cameras, the color gain increment is 0.001. Apparently, the X90 cameras are capable of reproducing finer color shadings than the X18 cameras. White balance is the process to adjust the color gains so that the colors of the specimen are reproduced correctly. You may set those color gains directly although arriving at a correct setting in this way is not easy. SE Premium provides two automation tools: “White Balance” and “Area WB”, which means “Area Based White Balance”. “White Balance” performs a statistics of red, green and blue values of all the pixels in the most

recent frame from the camera, and inverts the distribution by setting the color gains appropriately. Note that overly dark and bright areas in the live image are excluded from the calculation to ensure accuracy. “Area WB” is based on a rectangular region centered over a designated pixel, as indicated by the subsequent mouse click

[e-mail]

Scientific Instrument Software Corporation Limited

(eyedropper). In both “White Balance” and “Area WB”, the temperature of light source will be taken into consideration. SE Premium is able to process temperature within the range 3000K-9600K. It represents typical light temperatures by color squares. To set an appropriate temperature, first look through eyepieces at a blank slide, and then pick up a color square that matches the background. There are three points that one needs to bear in mind in white balance operations: 1.Base the white balance operation on areas of moderate brightness. Dim pixels will introduce noise and saturated pixels may not reflect the true color information at all.

2.

3.

Automated white balance operations will not change green gain. If one wishes to

change the overall brightness of the image while performing white balance, he/she

has to set the green gain manually.

Color gains must lie within allowed range. If a color gain is at its lower or upper

limits the white balance operation most probably has failed. To remedy, modify the green gain accordingly and repeat the white balance operation. SE implements a “smart” mechanism to capture a high-quality (low-noise) image efficiently (high-speed preview) by exploiting the still imaging function of Micrometrics™ cameras. Each Micrometrics™ camera is capable of integrating incoming photons for a very long time (hundreds of seconds for X90 models) in addition to delivering live images at a constant frame rate. The still imaging (long exposure) mode is intended to capture high-quality image (little or no noise introduced by large gain) while video imaging (high frame rate) mode is for image preview. If “Extended Exposure” is “Enabled” and “Reset Gain” is checked, upon receiving “Capture” command, SE Premium will set gain to unity, switch the camera to still imaging mode, set a long exposure time according to the previous gain so that the desirable brightness of the

image is retained, grab a frame and finally reset the gain to its original state. None of

these processing steps needs user intervention and he/she may continue to search for the next field of view and adjust focus quickly as the high-speed preview is

automatically restored.

If the captured image is for viewing and measurement purpose only and will not be subject to further filtering where linearity of pixel values is a must, Gamma may also be

employed to achieve the desired brightness of the image. A Gamma value that is less

[e-mail]

Scientific Instrument Software Corporation Limited

than unity enhances dark features and suppresses bright features of an image. A Gamma value that is greater than unity suppresses dark features and enhances bright features of an image. Note that Gamma is a non-linear transform of pixel values, originally introduced into SE Premium to offset the Gamma of a traditional analogy

video camera and that of a display device. A unity Gamma helps preserve the portability

of an image on various display devices and simplifies advanced processing of the image

where pixel values carry physical meanings. Since the X90 cameras are of 12-bit dynamic range and Gamma is applied before an

image is transferred to SE Premium, a small Gamma value (such as 0.1) or a large Gamma value (such as 5) may not cause a coarse looking to the image as it sometimes does for an 8-bit camera.

Saturation setting of a camera is purely artificial. It modifies the saturation of an incoming image from the camera by a specified amount. The default value 0.5 will not change the original saturation of the live or captured image. The value 0 will deplete color information from the image, i.e. effectively make a color image a monochrome one.

The value 1 will cause the image to be fully saturated.

Camera Manipulation Advanced Techniques

zFlat-Field Correction (FFC) zLook-Up Table z2-Channel Mode Real-time FFC is supported on all Micrometrics™ cameras. FFC may be applied both to live images or captured images. For optimal results, FFC should be enabled in the imaging process as it corrects not only shading but also color aberration. Real-time FFC refers to the calibration on the fly for sensor’s spatially varying response

to incident light, uneven illumination and non-uniform transmission rate of the whole light path from specimen to camera sensor, which includes objective, prism, video adaptor lens group and possibly some intermediate lenses. With the live image window

being the active window, select Enhance -> Flatfield Correction to create a calibration

[e-mail]

Scientific Instrument Software Corporation Limited

base. FFC calibration should be done on a per objective basis. For optimal result, it is

recommended that before executing this command: 1.Shoot the camera at a slowly varying white background. The field of view should not be obstructed by any object other than the white background. 2.Adjust the illumination, lens aperture or camera exposure time so that the image is

3.

as bright as possible but not exceeding its dynamic range.

Perform white balance to remove the apparent color aberrations. The temperature

should be set to white.

After the calibration base has been created, it may be turned off by un-checking “Apply Flatfield Correction” on the “Camera Control” sub-pane of the Control Panel, or turned on again by re-checking that option. FFC calibration base is part of a camera preset. LUT is most useful to imaging with X90 cameras to exploit the large dynamic range of these cameras. The latitude, or dynamic range, of an X90 camera is 16 times that of an X18 or X16 camera, which is standard 8-bit camera. Regardless the dynamic range of

the camera, the computer screen can at best display 8-bit image. However, if we map

only the lowest 1/16 portion of an X90 camera’s dynamic range we will see very dark

features and will not lose any details which the computer screen is capable of displaying.

The LUT is designed exactly for this purpose. A LUT is defined by two controlling points.

The three connected line segments, as made up by the two controlling points and two

end points, specify how the dynamic range of the camera to that of the digital image.

Needless to say, LUT is able to do much more than showing dark features. In fact, a general purpose LUT, as is provided in SE Premium, is used to enhance pixels within a particular intensity range and this intensity range is not necessarily limited to 8-bit. A

schematic plot of LUT is given below.

[e-mail]

1

y2

y1

Scientific Instrument Software Corporation Limited

(0, 0) x1 x2 1 Original Dynamic Range Setting both (x1, y1) and (x2, y2) to either (0, 0) or (1, 1) effectively cancels the intensity windowing. It is required that x2-x1>0.01. LUT may be defined by entering the coordinates of the two controlling points. To do this, select File -> LUT -> Define and fill the popup dialog box. LUT may also be adjusted in a WYSIWYG manner in the “Histogram” window. Press down the left mouse button at a position to the left of the left controlling point and drag to adjust the left controlling point. Press down the left mouse button at a position to the right of the right controlling

point and drag to adjust the right controlling point. Press down the left mouse button at a position between the two controlling points and drag to adjust the two controlling points simultaneously. In fluorescent microscopy it is rarely necessary to image all the red, green and blue primary colors simultaneously. 2-channel mode is thus introduced to SE Premiium and Micrometrics™ cameras to speed up the frame rate while all other parameters remain the same. Select File -> Fluorescent Special -> Green Red to preview only green and red channels. Select File -> Fluorescent Special -> Green Blue to preview only green and blue channels. Still image capture, on other hand, always includes all the three channels. There will be no interlacing in the 2-channel mode.

[e-mail]

Scientific Instrument Software Corporation Limited

Image Capture

zSystem Micrometer and Calibration zPresets zDestination of Captured Images zFrame Averaging zPost-Capture Sub-sampling zTime Lapse Capture Whenever an image is captured, it will be automatically calibrated according to the current system micrometer. After this process, a private micrometer is created to be

stored along with the image’s pictorial data. In SE Premium, an image from an external

source cannot be calibrated and thus is not acceptable to measurement. In other words,

the user may only use SE Premium to measure images that have been captured with SE

Premium. Before capturing an image to measure, the user must calibrate the system

micrometer itself. This is done by specifying a known distance (length) of a micrometer slide. The user must also provide a conversion formula, if he/she wants to use a measurement unit other than micron. Objective Switching (through Adjust -> Magnification) is a quick way to adapt the

system micrometer to a new magnification without another calibration. Depending on the microscope used and the nature of the application, objective switching method may not provide enough accuracy. The most exact method of calibration must be done for

each combination of objective and measurement unit (micron or 1/00 inch, for example).

Such a calibration base should be stored in a preset for fast access. A preset consists of system micrometer and camera control parameters for a particular

state. A preset is stored in the preset list, which is automatically stored and loaded by

SE Premium. Upon the selection of a preset from the preset list, the system micrometer and camera working parameter are restored to those of the state when the preset was created. By this mechanism, one is able to switch to a particular magnification,

measurement unit and camera parameter set at a mouse click.

A captured image may be stored in Field Group (the internal buffer), to a disk file directly, or both simultaneously. A captured image may also be loaded into an image

[e-mail]

Scientific Instrument Software Corporation Limited

window automatically, by checking “Preview Captured Image” on the “Capture Options”

sub-pane of the Control Panel. Frame Averaging serves two purposes. For a motionless object, it will reduce the noise.

For a moving object, it will create tracks. Post-Capture Sub-sampling is activated only when “Capture Full Field of View” is checked. It is a software processing function that has the following three advantages: 1.Calibration for the reduction of resolution is done automatically. 2.The image will be low-pass filtered to avoid aliasing. 3.The noise will be further suppressed due to spatial averaging. Time Lapse Capture is simply a time series of a regular image capture. It is worthy of mentioning that time lapse capture will not block SE Premium, i.e. the user may continue working on the captured images, or capture separate image, or even modify

camera or imaging control parameters.

[e-mail]