Abstract ‘ This paper focuses on denoising the Gaussian noise in natural image using raspberry pi.Gaussian noise is removed using spatial filter. The spatial filter employed in this paper is bilateral filter. The Bilateral Filter is a nonlinear filter that does spatial averaging without smoothing edges; it has shown to be an effective image denoising technique. The proposed system aims to substitute PC with Raspberry Pi which will not only drastically reduce the cost involved, but also will help achieving quality of service as the system will consume a smaller amount of power, yet will provide the same functionality as any other similar system does. The performance is evaluated in terms of Peak Signal to Noise Ratio. The experimental results are obtained and then compared with the results of Gaussian.
Keywords: Image denoising ‘Bilateral filter ‘Raspberry pi
Digital image processing focuses on two major tasks: Improvement of pictorial information for human interpretation and Processing of image data for storage, transmission and representation for autonomous machine perception. One of the most common uses of digital image processing techniques: improve quality; remove noise etc .
The aim of image denoising algorithm is to reduce the noise with the preservation of image features as much as possible. The images are corrupted by different types of noises. For example, dark current noise is due to the thermally generated electrons at sensor sites. Shot noise is due to the quantum uncertainty in photoelectron generation; and it is characterized by a Poisson distribution. Amplifier noise and quantization noise occur during the conversion of the number of electrons generated to pixel intensities.
The overall noise characteristics in an image depends on many factors, including sensor type, pixel dimensions, temperature, exposure time, and ISO speed. Images are often corrupted by additive noise, which is mostly modelled as Gaussian, during acquisition and transmission. Several methods have been proposed to remove noise and try to recover the ‘true’ image. Bilateral Filter is a nonlinear, edge-preserving, smoothing filter .The intensity value at each pixel in an image is replaced by a weighted average of intensity values from nearby pixels. This weight is based on a Gaussian distribution. In  blend of Gaussian/ bilateral filter and its method noise thresholding using wavelets were proposed.
In the multiresolution bilateral image denoising scheme, Bilateral Filter is applied on the approximation band of wavelet coefficients and wavelet thresholding is applied on the detail subbands . In a new hybrid image denoising scheme, Bilateral Filter is employed as pre-filter and post-filter for wavelet thresholding . In , two new image enhancement filters have been developed; to remove and enhance the appearance of an image according to the distance measure between adjacent pixels. These filters are Far Distance Filter (FDF) and Near Distance Filter (NDF). Compared to the well-known mean filter, the proposed filters can achieve better results in visual and quantitative measures.
In , an improved block matching and 3-D filtering (BM3D) image denoising algorithm used. Algorithm parameter values according to noise level, removes pre-filtering, which is used in high noise level are changed to modify the algorithm. In , a hybrid denoising algorithm which combines spatial domain bilateral filter and hybrid thresholding function in the wavelet domain was proposed. In a multiresolution multilateral filtering, Gaussian noise is reduced based on the idea of regional similarity .In  Wiener filtering on the wavelet coefficients was proposed to denoise an image degraded by an Additive White Gaussian Noise (AWGN).In , Multi wavelets can be considered as an extension of scalar wavelets. The main aim is to modify the wavelet coefficients in the new basis, the noise can be removed from the data.
Image processing in embedded system is critical, as this will decide the accuracy, speed & performance of entire system. This embedded system should have easy interface, scalability & support for end libraries for image processing. For this paper, Raspberry Pi is used as basic embedded system. Configuration and usage of Raspberry Pi is mainly highlighted in this paper.
The rest of the paper is structured as follows: Section II describes theoretical background, Section III describes raspberry pi hardware, Section IV describes the proposed method, Section V illustrates the experimental results and Section VI concludes the paper
A .GAUSSIAN FILTER
Gaussian filter is a linear, smoothing filter whose impulse response is a Gaussian function and it is not edge-preserving.
‘GF[I]_p=G’_?? (‘p-q’ ) I(q) (1)
where G_?? (x)=1/(2′?^2 ) e^(-x^2/(2??^2 )) (2)
Gaussian filtering is a weighted average of the intensity of the adjacent positions with a weight decreasing with the spatial distance to the centre position p. This distance is defined by G_?? (‘p-q’ ) where ?? is a parameter defining the extension of the neighbourhood. As a result, image edges are blurred.
B. BILATERAL FILTER
Bilateral filter overcomes gaussian filter problem by filtering the image in both range and domain (space). Bilateral filter is a local, nonlinear and non-iterative technique which considers both gray level (color) similarities and geometric closeness of the neighboring pixels. Mathematically, the bilateral filter output at a pixel location p is calculated as follows
BF[I]_p=1/W ‘_(q’S)”G_(??_s ) (‘p-q’ ) G_(??_r ) (|I(p)-I(q) | ) ‘ I(q) (3)
Where G_(??_s ) (‘p-q’ ) is a geometric closeness function and it is given by
G_(??_s ) (‘p-q’ )=e^(-(p-q^2)/(2??_s^2 )) (4)
G_(??_r ) (|I(p)-I(q) | ) is a gray level similarity function and it is defined as,
G_(??_r ) (|I(p)-I(q) | )=e^(-|I(p)-I(q) |^2/(2??_r^2 )) (5)
Wis a normalization constant and it is defined as
W=’_(q’S)”G_(??_s ) (‘p-q’ ) G_(??_r ) (|I(p)-I(q) | ) ‘ (6)
‘p-q’ is the Euclidean distance between p and q, and S is a spatial neighborhood of p.
The two parameters ??s and ??r control the behaviour of the bilateral filter. As range parameter ??r increases, the bilateral filter gradually approaches Gaussian convolution more closely because the range Gaussian widens and flattens, which means that it becomes nearly constant over the intensity interval of the image. On increasing the spatial parameter ??d, the larger features get smoothened.
The Raspberry Pi is a low cost, credit-card sized computer that plugs into a computer monitor or TV, and uses a standard keyboard and mouse.
The original Raspberry Pi is based on the Broadcom BCM2835 system on a chip (SoC), which includes an ARM1176JZF-700 MHz processor, VideoCore IV GPU, and was originally shipped with 256 megabytes of RAM, later upgraded (models B and B+) to 512 MB. The system has Secure Digital (SD) (models A and B) or MicroSD (models A+ and B+) sockets for boot media and persistent storage.
Fig 1: raspberry pi hardware
The SoC used in the first generation Raspberry Pi is somewhat equivalent to the chip used in older smartphones. The Raspberry Pi is based on the Broadcom BCM2835 system on a chip (SoC) which includes an 700 MHz ARM1176JZF-S processor, VideoCore IV GPU,and RAM. It has a Level 1 cache of 16 KB and a Level 2 cache of 128 KB. The Level 2 cache is used primarily by the GPU. The SoC is stacked underneath the RAM chip, so only its edge is visible.
There are two main connection options for the RPi display, HDMI (high definition)and Composite (low definition).
HD TVs and most LCD Monitors can be connected using a full-size ‘male’ HDMI cable, and with an inexpensive adaptor if DVI is used. HDMI versions 1.3 and 1.4 are supported, and a version 1.4 cable isrecommended. The RPi outputs audio and video via HMDI, but does not support HDMI input.
Older TVs can be connected using Composite (a yellow-to-yellow cable) or via SCART (using a Composite to SCART adaptor). PAL and NTSC TVs are supported.When using composite video, audio is available from a 3.5mm (1/8 inch) socket, and can be sent to your TV, to headphones, or to an amplifier. To send audio your TV, you will need a cable which adapts from 3.5mm to double (red and white) RCA connectors.
Though the model A and A+ do not have an 8P8C (“RJ45″) Ethernet port, they can be connected to a network using an external user-supplied USB Ethernet or Wi-Fi adapter. On the model B and B+ the Ethernet port is provided by a built-in USB Ethernet adapter.
D. POWER SUPPLY
The unit uses a Micro USB connection to power itself (only the power pins are connected ‘ so it will not transfer data over this connection). A standard modern phone charger with a micro-USB connector will do, but needs to produce at least 700mA at 5 volts.
You will probably need a number of cables in order to connect your RPi up.
1. Micro-B USB Power Cable
2. HDMI-A or Composite cable, plus DVI adaptor or SCART adaptor if required, to
connect your RPi to the Display/Monitor/TV of your choice.
3. Audio cable, this is not needed if you use a HDMI TV/monitor.
4. Ethernet/LAN Cable
Table 1 Raspberry Pi specification
Chip Broadcom BCM2835 SoC
Core architecture ARM11
CPU 700 MHz Low Power ARM1176JZFS Applications Processor
Memory 512MB SDRAM
Operating System Boots from SD card, running a version of the Linux operating system
Dimensions 85.6 x 53.98 x 17mm
Power Micro USB socket 5V, 1.2A (l)
F. KEYBOARD & MOUSE
Most standard USB keyboards and mice will work with the RPi. Wireless keyboard/mice should also function, and only require a single USB port for an RF dongle. In order to use a Bluetooth keyboard or mouse you would need to use a Bluetooth dongle, which again uses a single port.
A. DENOISING ALGORITHM
1. Let the image x is degraded by independent and identically
distributed (iid) zeromean white Gaussian noise with
standard deviation ??n .
2. The image is recovered from noisy observation by applying
3. The parameter is estimated in terms of PSNR
4. Experimental results are then compared with state-of-the-art
of Guassian filter
The major problem with Bilateral Filter isthe selection of the parameters photometric spread ??r and geometric spread ??d . The problem is analyzed in . if ??r is smaller than ??_n, noisy data could remain isolated and untouched in bilateral filter . When ??r is sufficiently large, ??d becomes important apparently, increasing the value of ??d too much results in over-smoothing and decrease of MSE. The work done by Yu et al.  and  concluded that the value of ??_r=2*??_n and ??d = 0.7 in the design of Bilateral Filter to get better performance in terms of PSNR. The performance of this image denoising scheme is evaluated using PSNR.
B.WORK FLOW FOR IMPLEMENTING
BILATERAL FILTER IN RASPBERRY PI
1. Download the rasbian os image from the website
2. Copy the Rasbian os image to SD card using win32 disk
3. Insert SD card into raspberry Pi module.
4. Raspberry pi is connected with mouse, keyboard and
5. Plug the power source into the main socket.Fig 2 shows the
6. In order to develop the software for image processing, it is
required to install Octave in the raspberry Pi. For
installing this, following this procedure.
i) Go menu->accessory->LXTerminal, it will open terminal in Raspberry Pi graphical interface .
ii) Type the following
sudo apt-get update
sudo apt-get upgrade
sudo apt-get install octave
sudo apt-get install octave-control
sudo apt-get install liboctave-dev
7. For python ,install the following
sudo apt-get install python
sudo apt-get install python-dev
sudo apt-get install libjpeg-dev
sudo apt-get install libfreetype6-dev
sudo apt-get install python-setuptools
sudo apt-get install python-pip
sudo easy_install -U distribute
sudo pip install RPi.GPIO
sudo pip install pySerial
sudo pip install nose
sudo pip install cmd2
sudo apt-get install python-matplotlib
sudo apt-get install python-mpltoolkits.basemap
sudo apt-get install python-numpy
sudo apt-get install python-scipy
sudo apt-get install python-pandas
8. Open menu ‘ programming– GNU octave
9. LX terminal will open, programming can done and execute.
Fig 2: Raspbian GUI
Experiments are carried out on various standard gray scale images of size 512 ?? 512 which are corrupted by a simulated Gaussian white noise with zero mean and five different standard deviations ??n ‘ [10, 20, 30, 40, 50].Denoising process have been performed on these images using bilateral filter and the PSNR results are calculated
Peak signal-to-noise ratio, often abbreviated PSNR This ratio is often used as a quality measurement between the original and a compressed image. The higher the PSNR, better the quality of the compressed or reconstructed image. The Mean Square Error (MSE) and the Peak Signal to Noise Ratio (PSNR) are the two error metrics used to compare image compression quality.
The MSE represents the cumulative squared error between the compressed and the original image, whereas PSNR represents a measure of the peak error. The lower the value of MSE, the lower the error.
The PSNR (dB) is defined as
PSNR(dB)=10log_10 ‘255’^2/MSE (11)
where Mean Squared Error (MSE) is defined as
MSE= 1/(N??N) ‘_(i=0)^(N-1)”_(j=0)^(N-1)’|x_(i,j)-x ??_(i,j) |^2
TABLE II PSNR Comparison of various denoising methods versus proposed method for different ??_n
??_n Gaussian bilateral
10 34.2567 77.9554
20 34.1459 73.1971
30 33.9876 70.3012
40 33.8987 68.5769
60 27.5679 66.7271
10 35.7689 72.2506
20 33.9876 71.9109
30 29.7654 68.7791
40 27.8923 67.4657
60 25.7893 65.9916
10 30.2340 71.0727
20 28.5438 69.2894
30 26.5281 67.5553
40 24.8901 66.4124
60 22.6781 65.0670
Laptops or personal computer consume a lot of power, as well as are very costly. Thus, this paper proposes replacement of PC with a low-cost as well as low-power consuming storage device, which can be designed using Raspberry Pi. This device can be used for presentations used in seminars, or for educational purposes in schools or colleges, or for entertainment purposes. Fig 4 shows the matlab output of Bilateral Filter. Fig 5 shows the output after programming done raspberry pi.
Fig3: Octave GUI
Fig 4: matlab BF output for lena image
Fig 5: BF output of lena image
after implemented in Rpi
CONCLUSION AND FUTURE WORK
Raspberry Pi OS was installed successfully on Raspberry Pi Kit. Octave program developed for denoising the Gaussian noise image. This method exploits the edge-reserving property of BF. The proposed method gives outperformed results than Gaussian in terms of PSNR .Personal computer consumes a lot of power, as well as are very costly. Thus, this paper proposes replacement of PC with a low-cost as well as low-power consuming storage device, which can be designed using Raspberry Pi.The proposed framework will inspire further research towards better understanding and working of raspberry pi.
 Gonzalez, R.C.,Woods,R.E.: Digital Image Process. Pearson Education (Singapore) Pte. Ltd., Delhi, India (2004).
 Tomasi, C., Manduchi, R.: Bilateral filtering for gray and color images. In: Proceedings of International Conference on Computer Vision, pp. 839’846 (1998)
 B.K.Shreya Kumar.:Image Denoising based on Gaussian/bilateral filter and its method noise thresholding. ?? Springer-Verlag London Limited 2012.
. Zhang, M., Gunturk, B.K.K.: Multiresolution bilateral filtering for image denoising. IEEE Trans. Image Process. 17(12), 2324’2333 (2008)
 Roy, S., Sinha, N., Sen, A.K.: A new hybrid image denoising method. Int. J. Inf. Technol. Knowl. Manag. 2(2), 491’497 (2010).
 Raghad Jawad AHMED. : Image Enhancement and Noise Removal by using New Spatial Filters. U.P.B. Sci. Bull., Series C, Vol. 73, ISSN 1454-234x ,Iss. 1, 2011.
 G. Kanchana, K. Suresh.:BM3D-Based Denoising of CFA Images for Single-Sensor Digital, IJERA, Vol. 2, Issue 6, December 2012, pp.1055-1059  Sabahaldin A. Hussain, Sami M. Gorashi.:Image Denoising based on Spatial/Wavelet Filter using Hybrid Thresholding Function, International Journal of Computer Applications (0975 ‘ 8887) Volume 42’ No.13, March 2012
 Rajpoot,N., Butt, I.: Multiresolution framework for local similarity based image denoising. PatternRecognit. 45(8), 2938’295 1(2012).
 Nevine Jacob and Aline Martin.: Image Denoising In The Wavelet Domain Using Wiener Filtering, December 2004.
Sachin D Ruikar, Dharmpal D Doye : Wavelet Based Image Denoising Technique. (IJACSA) International Journal of Advanced Computer Science and Applications, Vol. 2, No.3, March 2011
 Quick Start Guide, the Raspberry Pi ‘ Single Board Computer
...(download the rest of the essay above)