DSC-H50/B

Cyber-shot Digital Still Camera
The Sony DSC-H50 features Smile Shutter technology captures smiles the moment they happen5 , as well as Face Detection technology to optimize flash, focus, exposure and white for up to eight faces. Capture detailed images with 9.1 megapixel resolution, and get in close to the action with the 15x optical zoom lens. 9.1 MP Super HAD CCD More megapixels give you more detail and definition when you make big prints or crop in tight on your subject. The advanced Sony Super HAD (Hole Accumulated Diode) CCD design allows more light to pass to each pixel, increasing sensitivity and reducing noise. Carl Zeiss 15X Optical Zoom Lens 15X optical zoom provides a superb ability to bring long-distance subjects up close a significant advantage for situations where you cant be near the action, such as sporting events or wildlife photography. 3.01 Tilt-up Clear Photo LCD Plus Display The large, bright tiltable 3.01 Clear Photo LCD Plus display (230K pixels) tilts up for comfortable low-angle shooting, and an anti-reflective coating provides for excellent visibility to help you compose, view, and share photos with superb clarity and color reproduction, even in bright sunlight. Face Detection Face Detection technology automatically controls flash, focus, exposure, and white balance to help reveal faces in shadows, make skin tones look more natural, reduce redeye, and eliminates harsh facial glare. Smile Shutter Technology5 Smile Shutter technology captures smiles the moment they happen. The mode can be set to capture when your subject laughs, smiles or even grins, and like Face Detection is able to differentiate children and adults to set priority. Sony Double Anti-Blur Solution The combination of Super SteadyShot optical image stabilization and high ISO sensitivity helps compensate for shaky hands, minimizes blur, and allow flash-free shooting to preserve the mood. Super SteadyShot Optical Image Stabilization Super SteadyShot Image Stabilization compensates for shaky hands and minimizes blur with a built-in gyro that detects camera movement and sends correcting signals to a stabilization lens -- so theres no need to crop in on your shot to reduce blur. High Sensitivity (ISO 3200) Mode With high sensitivity of ISO 3200, you can shoot effectively in low light without flash to preserve the mood, while the Sony Clear RAW Noise Reduction system helps suppress the color noise that can degrade low-light shots. Intelligent Scene Recognition Mode Intelligent Scene Recognition mode automatically detects 5 different types of scenes and selects the appropriate camera setting: Backlight, Backlight Portrait, Twilight, Twilight Portrait, and Twilight using a Tripod. Intelligent Scene Recognition has two different modes, advanced and auto. Advanced mode takes one with the standard automatic settings and the other with optimized settings. Auto mode takes a single shot, utilizing the camera's optimal settings. Dynamic Range Optimization Standard & Plus Powered by the exclusive Sony BIONZ high-speed processing engine, Dynamic Range Optimization preserves image data in bright highlights and reveals more detail in shadows or backlit areas -- for great results even in difficult lighting conditions. New DRO Plus mode enhances optimization, analyzing each region of an image and performing additional image processing. NightShot Infrared System Sonys NightShot Infrared System lets you shoot photos in near-dark or totally dark (0 lux) conditions without using a flash, so you can capture creative night scenes that lie beyond the power of conventional cameras. Variable Noise Reduction (NR Level Setting) To allow greater control over NR, three NR Level Settings are available -- High, Low and Standard allowing you to select the optimal setting depending on the scene and individual tastes. For example, applying high NR helps obtain smoothly textured images for portraits or dark scenes, while the low setting is helps reduce noise without sacrificing resolution, making it especially useful for landscapes, detailed subjects, and contoured surfaces. Advanced Sports Shooting mode Advance Sports Shooting mode is ideal for shooting fast-moving subjects. Just choose Advanced Sports Shooting mode with the mode dial. The camera will then automatically use Auto Focus to analyze subject motion and predict the next move while the shutter button is halfway pressed. A high shutter speed (up to 1/4000 sec.) helps ensure you capture the moment clearly even if the subject is moving at high speed. Bracket Shooting Mode In addition to standard exposure bracketing (0.3EV/0.7EV/1.0EV), Bracket Shooting mode now includes settings for white balance and color modes. This function records three images with one shot, each with a different white balance or color mode applied, making it easy to capture the ambience and feeling of the original scene. Convenient Photo Modes To adjust quickly for specific shooting situations, you can choose from several convenient photo modes, including Auto, Program Auto, and Scene Selections such as Twilight, Twilight Portrait, Landscape, Portrait, Snow, High Sensitivity, Smile Shutter, Advance Sports Shooting, Fireworks, and Beach environments. ADDITIONAL FEATURES MPEG Movie VX Fine Mode High Contrast and Low Chromatic Aberration Versatile long distance flash Automatic Macro Shooting 9-Point Auto Focus Smart Zoom Feature8 Burst Mode Color Filter Kit

Specifications

General
Megapixel: 9.1 MP Imaging Device: 1/2.3 Super HAD CCD Recording Media: 15MB7 internal Flash Memory, optional Memory Stick DUO Media, optional Memory Stick DUO PRO Media 35mm Equivalent: 31 - 465 mm Focus: 9 Area Multi-Point AF, Monitoring AF, Flexible Spot AF Aperture Range: Auto (F2.7-F8(W)) / Program auto (F2.7-F8(W)) / Manual (F2.7-F8(W)) Shutter Speed: "Auto(1/4 - 1/4,000) / Program Auto(1"" - 1/4,000) / Shutter Priority(30"" 1/4,000) / Aperture Priority(8"" - 1/2,000) / Manual (30"" - 1/4,000)" Exposure Compensation: Plus / Minus 2.0EV, 1 / 3EV step ISO: Auto / 80 / 100 / 200 / 400 / 800 / 1600 / 3200 Smart Zoom Technology: 8 5M:Approx.20x (Total), 3M:Approx.25x(Total), VGA:Approx.81x(Total), 16:9 (2M):Approx.27x(Total) Digital Zoom: 0-2.0X (Precision) Optical Zoom: 15X Macro Mode: Yes Total Zoom: 30X Face Detection: Yes

Convenience

White Balance: Automatic, Cloudy, Daylight, Fluorescent 1, Fluorescent 2, Fluorescent 3, Incandescent, Flash, One Push, One Push (Total 9 modes) Self Timer: Yes (10 seconds, 2 seconds, Off) Memory Stick PRO Media Compatibility: Tested to support up to 16GB Memory Stick DUO PROTM media capacity1 ; does not support Access Control security function. Still Image Mode(s): Burst, Normal, Bracket Exposure Red-Eye Reduction: Yes (On/Off all modes) Burst Mode: Up to 100 Shot at 1.6fps (all resolutions) Erase/Protect: Yes/Yes Date/Time Stamp: No/ No Media/Battery Indicator: Yes/Yes Color Mode(s): Black & White, Normal, Sepia, Vivid
Battery Type: Lithium-Ion NP-BG1 Battery Capacity: 3.6V, 960 mAh

Software

Supplied Software: Windows: Picture Motion Browser Vers 2.0 + USB Driver Operating System Compatibility: Microsoft 2000 Professional, XP Home and Professional; Macintosh OS 9.1/9.2/OS X (10.0-10.5)

Convenience Features

AF Illuminator Light: Yes PictBridge Compatible: Yes Multi-Pattern Measuring: Yes Scene Mode(s): Auto, Easy Shooting, Program Auto, Shutter Speed Priority, Aperature Priority, Manual Exposure, Movie, High Sensitivty, Twilight, Twilight Portrait, Portrait, Landscape, Beach, Snow, Fireworks, Advanced Sports Shooting, Smile Shutter Movie Mode(s): Optical Zoom during movie recording: Yes Power Save Mode: Yes (after approx. 3 min. of inactivity) In-Camera Editing: Red-eye correction, soft focus, partial color filter, fish-eye filter lens, , cross filter, retro, radial blur, unsharp masking and happy face Super SteadyShot optical image stabilization: Yes

Dimensions

Weight: Weight: 14.6 oz (415 g) Body; 1lb 3.3 oz. (547 g) including Battery and optional Memory Stick DUO Media Measurements: Dimensions: 4 9/16 x 3 3/16 x 3 3/8" (116.1mm x 81.4mm x 86.0mm)
Service and Warranty Information
Limited Warranty Term: 1 Year Parts & Labor Color: Black UPC Code: 027242728981

Inputs and Outputs

Accessory Terminal: Multi connector Audio/Video Output(s): Yes via Mutli-Use Connector USB Port(s): Yes (Supports USB 2.011 ) HD Output: Yes (1080i)

Hardware

LCD: 3.0" Flip-up Clear Photo LCD Plus Display Viewfinder: 0.2" 201k FLC(Ferroelectric Liquid Crystal) EVF Lens Construction: 13 elements in 8 groups (including 1ED glass element and 4aspheric elements) Microphone/Speaker: Yes/ Yes Lens Type: Carl Zeiss Vario-Tessar Docking Station: N/A

Operating Conditions

Flash Effective Range: ISO Auto: Approx.0.2Approx.9.1m(W) / Approx.1.2-Approx.5.5m (T); ISO3200: up to Approx.18m(W) / Approx.11m(T) Flash Mode(s): Auto, Forced On, Forced Off, Slow Synch
1. Viewable area, measured diagonally. 2. HD viewing requires a Sony HD connector cable or HD Cyber-shot Station cradle and HDTV. All sold separately. 3. Music tracks can be up to five minutes long. When using Music Transfer to download music, tracks longer than five minutes will be reduced to five minutes when uploading to the Cyber-shot. 4. 35mm equivalent. 5. Up to 6 shots while in Smile Shutter mode. 6. Extended High Sensitivity ISO 6400 available in 3 megapixel image quality setting only. 7. Available storage capacity may vary and a portion of the memory is used for data management functions. 8. Smart Zoom feature will not work at highest resolution setting. 9. Special TV settings are required. Further details are available in the BRAVIA instruction manual. 10.Battery life may very depending on usage patterns, product settings, battery and environmental conditions. 11.Not all product with USB connectors may communicate with each other due to chipset variations. 2008 Sony Electronics Inc. All rights reserved. Reproduction in whole or in part without written permission is prohibited. Sony, BIONZ , BRAVIA TV, Clear Photo LCD Plus, Cyber-shot, Memory Stick Duo, Memory Stick PRO Duo, NightShot, Smart Zoom, Stamina, Super HAD, and Super SteadyShot are trademarks of Sony. All other trademarks are trademarks of their respective owners.

Optics/Lens

Focal Length: 5.2 - 78 mm
Please visit the Dealer Network for more information at www.sony.com/dn Sony Electronics Inc. 16530 Via Esprillo San Diego, CA 92127 1.800.222.7669 www.sony.com Last Updated: 10/16/2008

MultiCraft

International Journal of Engineering, Science and Technology Vol. 2, No. 1, 2010, pp. 35-48
INTERNATIONAL JOURNAL OF ENGINEERING, SCIENCE AND TECHNOLOGY www.ijest-ng.com 2010 MultiCraft Limited. All rights reserved
Improvement in quality testing of Braille printer output with Euclidean distance measurement using camera calibration
Suman Karmakar*, Rudra Prasad Chatterjee* and Uma Dutta
Central Mechanical Engineering Research Institute,M.G. Avenue, Durgapur-713209, India E-mails:(r_chatterjee@cmeri.res.in (Rudra Prasad Chatterjee*), suman2007@gmail.com (Suman Karmakar*) *Corresponding authors); umadutta@cmeri.res.in (Uma Dutta)
Abstract This paper focuses on quality testing of Braille printed paper using calibrated camera by detecting dots and measuring the Euclidean distances between them with equal gap, vertically and horizontally. For higher accuracy, camera calibration is essential to observe a planar checker board pattern from different distances and orientations. In this process, the position of the camera is fixed and the pattern can be freely moved. Radial lens distortion is modeled. Machine simulation and experimental results have also been discussed. Quality improvement can be achieved by giving a feedback after finding the distorted edges from image processing of the paper. This approach thus definitely helps the blind reader to avoid disturbances in reading the printed documents by both, single sided and inter-point printer. Keywords: Image processing, camera calibration, quality testing of Braille printer 1. Introduction Braille, a touch-reading system for the visually impaired people was first introduced in 1825 by Louis Braille (Hentzshel, 1993). Braille is a primary medium of reading and writing for the blind people or having low vision. Many blind and visually impaired individuals find their ability to access information more quickly and perform tasks that involve reading or writing more efficiently using Braille than by listening to a personal reader, dictating to personal secretary or alternative technologies such as audio recordings, talking computers, or other electronic devices. Braille character appears in dot format on Braille paper that can be identified by blind people on touching the enclave. To identify the proper position of the dots in the printer output, we need to maintain a fixed gap between every dot, as well as between every letter and every line. A method of thorough image processing to measure the gap between dots can help in this case. Generally, images taken by the digital camera result into a distorted image due to the lens distortion i.e. the actual output is always not available. Geometric distortion is an error on an image, between the actual image coordinates and the ideal image coordinates. Among various such nonlinear distortions, the radial distortion, which is along radial direction from the center of distortions, is the most severe part and is also the most common. Straight lines in the undistorted subject bulge in the characteristic barrel fashion, in the image. Straight lines running through the image center remain straight and a circle concentric with the image center remains circle, although its radius is affected. The most typical cases of distortion, barrel and pin-cushion distortion are primarily radial in nature, with a relatively simple one or two parameter model accounting for most of the distortion. Thus Camera calibration is a necessary step in 3-D computer vision in order to extract metric information from 2D images. The term, Camera re-sectioning (often called, camera calibration) is the process of finding the true parameters of the camera that produces a given photograph (Bougnoux, 1998; Brown, 1971; Gennery, 1979; Hartley and Zisserman, 2003; Faig, 1975).
Karmakar et al. / International Journal of Engineering, Science and Technology, Vol. 2, No. 1, 2010, pp. 35-48
Related works A four-step camera calibration procedure with implicit image correction was carried out by Heikkila and Silven (1997). Planebased camera calibration by Sturm and Maybank (1999) was discussed. A novel camera calibration method based on genetic algorithm carried out by Chen et al. (2008) has been given. Simplified intrinsic camera calibration and hand-eye calibration for robot vision was carried out by Malm and Heyden in 2003. Perspective Geometry based single image camera calibration technique has been discussed by Avinash and Murali in 2008. In 2006, Yanqing and Zhiyan came out with a flexible camera calibration method for computer visual 3d reconstruction system. This paper gives a brief sketch on different camera calibration techniques in section 2, followed by section 3 with adaptive camera calibration technique. Section 4 describes tracking lens distortion. The undistortion process has been described in section 5. Section 6 suggests a procedure for quality measurement of processed images/outputs. Section 7 presents results and discussion while Section 8 is the conclusion. 2. Camera calibration techniques A. Photogrammetric calibration Photogrammetry is defined as a science of making measurements from photographs. Calibration is performed by observing an object whose geometry in 3-D space is known. The calibration object usually consists of two or three planes orthogonal to each other. Sometimes, a plane undergoing a precisely known translation is also used. These approaches require an expensive Calibration setup (Knyaz, 2006; Maas, 1997; Remondino and Borlin, 2004; Weckesser and Hetzel, 1994; Ritter et al., 2006). B. Self-calibration Calibration methods can be used when no Euclidean information is available, and, if properly designed, can also cope with varying intrinsic parameters. Knowledge of the camera motion and the intrinsic parameters allows for the Euclidean reconstruction of the scene. Images taken by the moving camera on a static image plane with fixed internal parameters are sufficient to recover both the internal and external parameters which allow reconstructing 3-D structure. This approach is very flexible. There are many parameters to estimate; sometime we cannot always obtain reliable results (Hartley, 1994; Luong and Faugeras, 1997; Hua et al., 2005; Maybank and Faugeras, 1992; Paulo et al., 1999). C. Zhengyou Zhang method This technique only requires the camera to observe a planar pattern shown at a few (at least two) different orientations. Zhang has developed a toolbox that allows others to implement calibrations that involve algorithms related to their personal research. This procedure is relatively accurate. In this technique several checkerboards pattern images (here 9x7 boxes) in various position with different distance are captured by a single digital camera (Zhang, 1998; Zhang, 1999). D. Mathematical Representation of 3-D Model Images

]T. A 3D point is denoted by M = [X, Y, Z]T. We use x to denote the augmented vector by ~ ~ T T adding 1 as the last element: m = [u, v,1] and M = [X, Y, Z,1]
A 2D point is denoted by m = u, v
A camera is modeled by the usual pinhole: the relationship between a 3D point M and its image projection m is given by

s m = A[R

u 0 v0 1
where s is an arbitrary scale factor; (R, t), called the extrinsic parameters, is the rotation and translation which relates the world coordinate system to the camera coordinate system; A is called the camera intrinsic matrix, and (u0, v0) are the coordinates of the principal point, and the scale factors in image u and v axes, and c the parameter describing the skewness of the two image axes. E. Homography between the model plane and its image: Without loss of generality, we assume the model plane is v on Z = 0 of the world coordinate system. Lets denote the column of the rotation matrix R by r i. From (1), We have,
X u X v = A[r1 r 2 r 3 t ] Y = A[r1 r 2 t ] Y s 1 1
By abuse of notation, we still use M to denote a point on the model plane, but since Z is always equal to 0. In turn, therefore, a model points M and its image ms are related by a homography H:

sm = HM

H = A[r1 r2 t ]
The 3x3 matrix H is defined up to a scale factor. 3. Solving camera calibration This section provides the details regarding how to solve the camera calibration problem, effectively. The straight lines in several orientations throughout these images are used to determine the pattern of radial lens distortion. Figure 2. Shows the calibration process have been adapted here.
Figure 1. Original image (640pix X 480pix) of a calibration checkerboard pattern, taken with a Sony Cyber Shot DSC-H50 camera A. Determining the radial distortion coefficients The first part of the calibration process is to determine an image coordinate remapping that causes images taken by the camera to be true perspective images, that is, straight lines in the world project as straight lines in the image. The procedure makes use of one or several images with many known straight lines in it. Architectural scenes are usually a rich source of straight lines, but for most of the work here we used pictures of the checkerboard pattern shown below (Figure 1) to determine the radial lens distortion. The checkerboard pattern is a natural choice since straight lines with easily localized endpoints and interior points can be found in several orientations (horizontal, vertical, and various diagonals) throughout the image plane. The checkerboard pattern also has the desirable property that its corners are localizable independent of the linearity of the image response. That is, applying a nonlinear monotonic function to the intensity values of the checkerboard image, such as gamma correction, does not affect corner

localization. As a counter example, this is not the case for the corners of a white square on a black background. If the image is blurred somewhat, changing the image gamma will cause the square to shrink or enlarge, which will affect corner localization. The pattern in Figure 1 was photographed on a Sony Cyber Shot DSC-H50camera. Since this lens, like most, changes its internal configuration depending on the distance it is focused at, it is possible that its pattern of radial distortion could be different depending on where it is focused. Thus, care was taken to focus the lens at infinity and to reduce the aperture until the image was adequately sharp. Start
Print Pattern & attaching to planner surface
Taking images of model plane

Detecting feature point

Calculation of five intrinsic Parameters
Coefficient of radial distortion measurement by least-squares
End Figure 2. Flow chart for showing the calibration process 4. Tracking lens distortion A. Sobel edge detector For more easily visually tracking lens distortion of the test images, simple Sobel edge detector process is run which produce the image shown in Figure 2. Mathematically, the operator uses two 33 kernels which are convolved with the original image to calculate approximations of the derivatives - one for horizontal changes, and one for vertical. If we define I as the source image, and there are two images which at each point contain the horizontal and vertical derivative approximations, the computations are as follows:
+ 1 + 2 + 1 Py = 0 *I 1

+Px = + +

* I 1
where * here denotes the 2-dimensional convolution operation. The x-coordinate is here defined as increasing in the "right"-direction, and the y-coordinate is defined as increasing in the "down"-direction. At each point in the image, the resulting gradient approximations can be combined to give the gradient magnitude. Using this information, we can also calculate the gradient's direction:

(Px 2 + Py2 )

Using this information, we can also calculate the gradient's direction:

P = arctan x P y

The pattern of distortion can now be made evident to a human observer by shrinking this edge image in either the horizontal or the vertical direction by an extreme amount. Table 1. A 9x9 Convolution filter that detects even corners of the checkerboard pattern 0 -1 -1 -1 --1 -1 -1 --1 -1 -1 --1 -1 -1 --1 -1 -1 --1 -1 -1 --1 -1 -1 --1 -1 -1 -Table 2. A 9x9 Convolution filter that detects odd corners of the checkerboard pattern -1 -1 -1 --1 -1 -1 --1 -1 -1 --1 -1 -1 --1 -1 -1 --1 -1 -1 --1 -1 -1 --1 -1 -1 -1

Figure 3. The edges of the checkerboard pattern Found by using Sobel edge detector of the Undistorted Pattern We observed that lines passing through the center of the image stay straight, as do the vertical lines at the extreme left and right of the image. Lines which lie at intermediate distances from the center of the image are bowed.
Since this filter Table-1&2 itself resembles a checkerboard pattern, it gives a strong response (positive or negative, depending on which type of corner) when centered over a checkerboard corner. Taking the absolute value of the filter output produces an image where the checkerboard corners appear as white dots, as in Figure 4 showing the odd and even position of the corners. Image points can be easily localized by first convolving the image with the filter in Tables 1&2. In abstract terms a convolution is defined as a product of functions and that are objects in the algebra of Schwartz function in. Convolution of two functions and over a finite range is given by
[ f * g ](t ) f ( )g (t )d
Where the symbol denotes convolution of and. Once the distortion parameters are solved for, it is possible to undistort any image taken with the same lens as the calibration images so that straight lines in the world image to straight lines on the image plane (Foster, 2006). 5. Undistortion image of Braille paper The undistortion process of the paper can be achieved by the method described in Zhang (1998). By comparing both Figure 4 and 5 we can observe a significant lens distortion. The dimension of Braille cell can be represented by maximum six dots and is shown in Figure 7.
Figure 3. The results of convolving the undistorted checkerboard image with the filter in Table 2 and 3, and taking the absolute value of the filter outputs
Figure 4. Braille original gray Image taken withSonyDSC-H50, showing Barrel distortion
Figure 5. Undistorted Braille image
Figure 6. Gray scale image of Braille Paper with selected 18 dots
Figure 7. Braille cell dimension
Figure 8. Threshold image by the thresholding value =.5451 6. Procedure for the quality measurement A. Process for finding distance between dots Thresholding is an image processing technique for converting a grayscale or color image to a binary image based upon a threshold value. If a pixel in the image has an intensity value less than the threshold value (Vth =0.5451), the corresponding pixel in the resultant image is set to black. Otherwise, if the pixel intensity value is greater than or equal to the threshold intensity, the resulting pixel is set to white. Image thresholding is very useful for keeping the significant part of an image and getting rid of the unimportant part or noise. This holds true under the assumption that a reasonable threshold value is chosen. Figure 8 shows the threshold image of printed Braille paper with the value of.5451. Removing small objects-removes from a binary image all connected components (objects) that have fewer than P pixels, producing another binary image, BW2. The default connectivity is 8 for two dimensions, 26 for three dimensions. Label connected components in a binary image returns a matrix L of the same size as BW. The centroid algorithm is based on the standard center of mass equation in discrete form (McDowell, 2004).

i. f ( j , i ) u x + j. f ( j , i ) u y j = j min i = i min j max

j = j min j = j min

f ( j, i)

jmax =maximum row index

Where: imin =minimum column index, imax =maximum column index, jmin =minimum row index,
7. Results and discussion The camera to be calibrated is Sony Cyber Shot DSC-H50 camera. The image resolution is 640 x 480. The model plane contains a pattern of 9x7 squares, so there are 252 corners. The size of the pattern is 29mm.Five images of the plane under different orientations were taken, as shown above. We can observe a significant lens distortion in the images. The corners were detected as the intersection of straight lines fitted to each square. Our experiments are composed of single camera calibration setup. For the single camera experiments, we have tried various numbers of images. On the other hand, we found that using more than 5 images did not increase the accuracy any more. We applied this calibration algorithm to all the 5 images. The results are shown in Table 8. Start

Reading Image

Conversion (RGB to gray)

Thresholding Unit Matrix

Removing small object

Convolution filter

Label Matrix

Image region properties

Centroid

No Object

Distance among center
Figure 9. Showing distance measurement process
In Table 7, we have shown the distortion in pixels against the number of images. We have taken the Corner points of the original (distorted) Image and the corner point of the undistorted images. Due to undistortion of the image the corner point will shift with respect to the original image. We can see that the distortion is varied with number of image shown in Figure 11. C1 R1 C2 C3 C4 C5 C6
R3 Figure 10. Detected Braille objects by centroid method Scanned from left to right Table 3. Distance (mm) between the centroids of R1 and R2 1R2(mm) Avg. Error(mm) 2.74 2.71 2.94 2.82 2.96 2.75 2.752.5=.25 10.35 10.24 11.11 10.65 11.18 10.39.94(pixels) Table 4. Distance (mm) between the centroids of C1 and C3 5.72 21.61 6.34 23.96 C1C3(mm) 5.92 5.11 22.37 19.31 5.72 21.61 6.55 24.75 Avg. 5.89 22.26 Error 5.896.64=.75 2.83(pixels)

2.48 9.37

Table 5. Distance (mm) between the centroids of C1 and C2 2.66 10.05 2.24 8.46 C1C2(mm) 2.45 2.24 9.25 8.46 2.65 10.01 2.04 7.71 Avg. 2.38 8.99 Error 2.382.3=.08.30(pixels)
Table 6. Distance (mm) between the centroids of R3 and R6 10.97 41.46 11.08 41.87 R3R6(mm) 13.49.81 37.79 9.24 34.92 7.9 29.85 Avg. 10.39 39.26 Error 10.3910.4=.01.03(pixels)
Table 7. Distortion results of single camera calibration No of images Distortion (per pixels) 1 1.1.1.1.1.62168
Karmakar et al. / International Journal of Engineering, Science and Technology, Vol. 2, No. 1, 2010, pp. 35-48 Table 8. Result of calibration processed by Matlab Principal Distortion(kc) Point(cc) K1 K2 K3 K4 312.0 222.2 0.00581 0.7948 0.14421 0.319.5 239.5 0.26423 -0.0000 0.22984 0.293.2 236.5 0.30255 0.6528 0.21249 0.319.5 239.5 0.13955 0.0000 0.18865 0.319.5 239.5 0.11923 0.0000 0.15842 0.00457 12

No Of Image 1

Focal Length(fc) 801.693.633.613.567.793.692.627.612.567.1 9535

K5 0.0.0.0.0.00 000

Pixel Error(err) 0.0.0.0.0.0.0.0.0.0.294 26
Figure 11. Average Pixel error of each corner point of the model plane 8. Conclusion An advantage of research work for camera calibration is that this method is much closer to accuracy than other processes used for object detection by centroid measurement and distance calculation. Zhengyou Zhang Method for camera calibration has been followed. This technique only requires the camera to observe a planar pattern shown at a few (at least two) different orientations. This research work has been carried out at Central Mechanical Engineering Research Institute, Durgapur, India and mainly focused on finding equal distance in the Braille printer output by image processing technique. The distorted images are compared with the modified undistorted one. Results, thereafter, finally establish this to be a very satisfactory quality testing procedure of a commercially fabricated Braille printer.
Acknowledgement The authors are grateful to Director, CMERI, Durgapur for extending his help regarding this project work. We thank Somajyoti Majumder scientist CMERI, Durgapur for his technical help, good ideas and valuable assistance during the experiment. References Avinash N. and Murali S., 2008. Perspective geometry based single image camera calibration. Journal of Mathematical Imaging and Vision, Springer Netherlands, Vol 30 , No. 3, pp 221 230. Bougnoux S., 1998. From projective to euclidean space under any practical situation, a criticism of self-calibration. 6th International Conference on Computer Vision, Bombay , India, pp. 790796. Brown D. C., 1971. Close-range camera calibration. Photogrammetric Engineering , pp. 855866.
Chen F, Zhao J. and Zhao H.W., 2008. A novel camera calibration method based on genetic algorithm. IEEE Transactions, Singapore, pp. 22222227. Faig W., 1975. Calibration of close-range photogrammetry systems: Mathematical formulation. Photogrammetric Engineering and Remote Sensing, Vol. 41, No. 12, pp. 14791486. Foster M. P., 2006.Normalized Convolution Techniques. MP Foster - clara.net. Gennery D., 1979. Stereo-camera calibration. In Proc. 10th Image Understanding Workshop, Stanford University Stanford, California 94305, pp. 101108. Hartley R. and Zisserman A., 2003. Multiple View Geometry in Computer Vision. Cambridge University Press, second edition, ISBN 8, pp. 155157. Hartley R.I., 1994.An algorithm for self calibration from several views. IEEE Conference on Computer Vision and Pattern Recognition, pp. 908912. Hentzshel T.W., 1993. An optical braille reading system. M.Sc. Thesis, University of Manchester Institute of Science and Technology. United Kingdom. Heikkila J. and Silven O. 1997. A four-step camera calibration procedure with implicit image correction. InfoTech Oulu and Department of Electrical Engineering University of Oulu FIN-90570 Oulu, Finland. Hua T.P., Sugiyama A. and Faucon G. 2005. A new self-calibration technique for adaptive microphone arrays, 2005 International Workshop on Acoustic Echo and Noise Control High Tech Campus, Eindhoven, The Netherlands, September 12 - 15, 2005, pp. 237-240. Knyaz V.A., 2006. Automated calibration technique for photogrammetric system based on a multi-media projector and a CCD camera, ISPRS Image Engineering and Vision Metrology, Dresden, pp. 25-27. Luong Q.T.and Faugeras O., 1997.Self-calibration of a Table IX technical data of the imaging system (Sony DSC-H50) moving camera from point correspondences and fundamental matrices. International Journal of Computer Vision, Vol. 22, No. 3, pp. 261-289. Maas H. G., 1997. Dynamic photogrammetric calibration of industrial robots. in videometrics V, SPIE Proceedings Series, Vol. 3174, SPIEs 42nd Annual Meeting, San Diego, 27.7.- 1.8. Malm H. and Heyden A., 2003. Simplified intrinsic camera calibration and hand-eye calibration for robot vision. IEEE Conference on Intelligent Robots and Systems, Las Vegas, Nevada. Maybank S.J. and Faugeras O.D., 1992.A theory of self calibration of a moving camera. International Journal of Computer Vision, Vol. 8, No. 2, pp.123152. McDowell M., 2004. An integrated centroid finding and particle overlap decomposition algorithm for stereo imaging velocimeter. NASA Technical Memorandum (NASA/TM2004-213365). Paulo R. S., Mendonc A. and Cipolla R., 1999. A simple technique for self-calibration. Conference on Computer Vision and Pattern Recognition (CVPR '99), 23-25 June 1999, Ft. Collins, CO, USA. IEEE Computer Society, ISBN 0-7695-0149-4, pp. 1063-6919. Remondino F. and Borlin N. 2004. Photogrammetric calibration of image sequences acquired with a rotating camera. Int. Archives of Photogrammetry and Remote Sensing, Vol.34 (5/W16). Ritter M., Hemmleb. M., Faber P.C., Lich B. and Hohenberg H. 2006. Sem/fib stage calibration with photogrammetric methods. ISPRS Commission V Symp. 2006 (Int. Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences), Vol. XXXVI Part 5. Salvi J., Armangu X. and Batlle J., 2002. A comparative review of camera calibrating methods with accuracy evaluation. Pattern Recognition, pp. 16171635. Sturm P. F and Maybank S. J., 1999. On plane-based camera calibration: a general algorithm, singularities, applications. Computational Vision Group, Department of Computer Science, The University of Reading Whiteknights, IEEE 1063-6919/99. Weckesser P. and Hetzel G., 1994. photogrammetric calibration method for an active stereo vision system. Intelligent Robotic Systems (IRS), pp. 430-436. Yanqing Z. and Zhiyan W., 2006. A flexible camera calibration method for computer visual 3D reconstruction system. IEEE Transactions 0-7803-9737-1/06, Vol 2, Guilin, China. Zhang Z. A flexible new technique for camera calibration. Technical Report MSRTR-98-71, Microsoft Research, December 1998. Available together with the software at http://research.microsoft.com/zhang/Calib/ Zhang Z., 1999.Flexible Camera Calibration By Viewing a Plane From Unknown Orientations. International Conference on Computer Vision , Corfu, Greece, pp. 666-673.

Biographical notes Suman Karmakar received his bachelors degree in Computer Science and Engineering Department from Bankura Unnayani Institute of Engineering, Bankura under West Bengal University of Technology in 2007. He joined Central Mechanical Engineering Research Institute, Durgapur as a Project Fellow in 2007. He involved in different project like Development of Software for the Optimization of Silicon Rubber
Mould Characteristics, and Development of high speed Braille Printer. His work area includes programming in VC++ and image processing using Matlab and C++. Rudra Prasad Chatterjee (M08) became a Member (M) of IEEE in 2008 of Communication Society. He received his bachelors degree in Electronics and Communication Engineering from Dr.B.C. Roy Engineering College, Durgapur under West Bengal University of Technology in 2005. He joined National Institute of Science and Technology, Berhampura, Orissa as a lecturer in 2005. Later he joined Central Mechanical Engineering Research Institute, Durgapur as a Scientist in 2006. His work area includes wireless control of robotic devices, mobile network etc. He has been involved in SUPRA Institutional Project (SIP-24) Design and Development of Snake Robot for Disaster Management. He has published ten papers including national and international journals. Uma Datta received her bachelors degree in Radio Physics and Electronics from Calcutta University in 1981 and Masters in Electrical from Regional Engineering College, Durgapur in 1994. She joined Central Mechanical Engineering Research Institute, Durgapur in 1984, as a Scientist. She became the Head of the Department of Electronics and Control Lab in the year of 2005. Her work area includes Instrumentation Engineering and Control System. She has supervised the following projects at CMERI: Development of Expander Extruder for processing oil seeds with Automatic Control (sponsored), Design and Development of CTC roller sharpening machine with automatic control (sponsored by: Department of Science and Technology, Govt. of India.), Design and development of Electro-magnetic Stirrer for 1.3 Kg. Molten AL-356 (sponsored by: NMRL, Govt. of India) etc. She has published twenty papers including national and international journals. Received June 2009 Accepted August 2009 Final acceptance in revised form September 2009

Table 9. Technical data of the imaging system (Sony DSC-H50) Max resolution 3456x 2592 Image ratio w:h 4:3, 3:2 Effective pixels 9.1 million Sensor size 1/2.3" 6.16x4.62 mm,0.28 cm Sensor type CCD Normal range 50cm focus Aperture range F2.7 - F4.5
Figure 12. (a) Principal Point of the number for square widths of images of the model plane Check board squares counted from left to right (b) Note that the center of the image is around box no 5.
Figure 13. Braille original gray Image
Figure 14. Braille Undistorted image
Figure 15. Histogram of the gray Braille image
Figure 16. Histogram of the Braille threshold
Figure 17. Shifting the corner point of undistorted model plane
Figure 18. Five images of a model plane, together with the extracted corners (indicated by cross)
Figure 19. Five undistorted images of a model plane