Dear Mr. Sachin Sharma, Naturally you need an embedded hardware. Very conveniently you can use Microcontroller.

There are chips for RADAR applications.

http://www.newscientist.com/article/dn4725-tiny-radar-could-make-driving-safer.html

http://idw-online.de/pages/de/news492745

http://phys.org/news/2012-09-radar-technology-housed-thumbtack-sized-chip.html

http://www.caltech.edu/content/caltech-engineers-design-revolutionary-radar-chip

Regards,

Joachim

Sachin,

I would suggest that you measure the trade off between the following criteria:

1) Whether the sensor will be exposed to weather or not, this will require an enclosure. If you use for example an image processing for detection: will the enclosure limit its visibility?

2) As you get more sophisticated and need more distance coverage, you will need more power to process the signal, but remember that you don't want to drain the power source in 5 minutes.

3) Commercial application or research-This has potential for practical application deployment. If you want to make it commercially viable you will need to keep costs and form factor down.

This will influence which hardware platform you might want to use.

For embedded, real-time image/video processing applications, the best way in my opinion is to adopt a hybrid microcontroller/FPGA-based approach. Field Programmable Gate Arrays (FPGAs) are well-known for their use in computationally heavy tasks involved in handling pixel matrices especially in fast-moving-cars. A best-known FPGA in digital signal processing is the Spartan-6 FPGA family. As you've not mentioned the domain in which you've built the algorithms, I presume OpenCV or Matlab is used. Unfortunately, FPGAs cannot be used directly with code written on PC-based platforms. An alternative is to use Matlab's HDL Code Generation and Verification functionality to program an FPGA.

Alternatively, you can go for a microcontroller-only system where it is generally possible to upload OpenCV-based code. ARM family og microcontrollers with the relevant boards are commonly used in signal/image/video processing applications. The link attached with this post will give you a good idea of various alternatives in this regard.

In the end, I would suggest NOT to go for commonly available PIC, Arduino, Raspberry PI or Netduino families as these hardware boards do not have the computational power and resources required to handle heavy image processing tasks.

http://dsp.stackexchange.com/questions/1437/list-of-all-available-armdsp-processors-for-surveillance

Thanks a lot Syed. Yeah I have built my code in MATLAB so I was thinking whether beagleboard will serve the purpose or not.

Thanks a lot for the appropriate answer Arturo Geigel. I had the same criteria when took this application.

Thanks Joachim Pimiskern. I do have a doubt that RADAR will be an expensive choice though results may get improved by use of it.

Thanks Afaq Ahmad for the information.

Dear Sachin, You are most welcomed.