মঙ্গলবার, ১৮ ডিসেম্বর, ২০১২

ARM Based Video Surveillance System & Pattern recognition


                                 Introduction
            The  video  surveillance  system  based  on  computer technology, embedded  technology, and network   technology can  carry  on  all-day  automatic  real-time  monitoring  to  the important department or place of the various professions, and it  has  already  become  the  important  component  of  peaceful guard  domain.  In  recent  years,  the  research  on  video surveillance  system  has  caused  widespread  attention  of  the academic  circle  and  industrial  field  in  home  and  abroad, which  also  becomes  the  front  topic  of  domestic  and  foreign research. However,  the  alarm  system  applied  to  family safety protection  is  static and based on  the  signal at present, so the video information is unable to be viewed. The merit often embedded system makes it suits the family monitoring. Therefore,  in  the  study,  the  embedded  video  surveillance system  and  the  alarm  system  was  integrated,  and  the embedded  remote  video  surveillance  system  based  on  the ARM  was  studied  and  designed.  The system of this study realized real-time monitoring to the family security.

Due to the advantages of small size, stable performance, convenient communication, embedded devices are frequently used in the video surveillance system. In this system, ARM Cortes M-3 (mbed LPC1768) microprocessor is selected as the hardware platform. Under the Windows operating system, a video surveillance systems based on ARM is realized. The system is able to provide a flexible network and a feature-rich portable terminal. The video surveillance system can be loaded to living organisms, through the controlling of the living organisms to accomplish search and rescue purposes.

In machine learning, pattern recognition is the assignment of a label to a given input value. An example of pattern recognition is classification, which attempts to assign each input value to one of a given set of classes (for example, determine whether a given email is "spam" or "non-spam").

Overall Scheme Design
The overall block diagram of the system is shown in Figure 1. The system is divided into two parts: video acquisition terminal and handheld terminal. In the video acquisition terminal system, the video capture module collects video information and the Wi-Fi module receives the network information. Then the collected data is sent to the LPC1768 microprocessor. After compression, the information is sent to the handheld terminal by the Wi-Fi wireless communication module. In the handheld terminal system, the video and information is displayed on the TFT LCD display module after extracted. The sending terminal is controlled via the key control module. Combined with the closed-loop control scheme between the sending terminal and handheld terminal, wireless video transmission and Wi-Fi function has been achieved.
                                 Overall block diagram of the system.

Hardware Composition of the System
                                       
                                       mbed LPC1768 ARM (microprocessor/µP)

LPC1768 processor has been adopted as the main control unit in the hardware platform. LPC1768 is based on the construction of ARM7 Cortex M-3 core. It is a low-power, high-performance processor. It has a powerful hardware encoding and decoding unit, perfect peripheral and rich interface resources. The system consists of two parts: the video acquisition terminal and the handheld terminal. The sending end terminal is designed to collect required data, compress and send the video via Wi-Fi. The handheld terminal is designed to receive, decompress and display the information sent from the sending terminal. Wi-Fi wireless communication module is used to realize the transmission of the remote data transmission between the sending terminal and the handheld terminal. These two parts are introduced in detail below.

Hardware Structure Design of the Sending Terminal
The hardware block diagram of the sending terminal is shown in Figure 2. It is composed of power supply unit, LPC1768 microprocessor with on chip flash and SRAM. Flash memory can store the user application program. As a system memory, SDRAM can store the system information. LPC1768 microprocessor receives the video data from the video capture module and the Wi-Fi network information from the Wi-Fi module. After processed by the microprocessor, the information is sent to the handheld terminal via the Wi-Fi wireless transmission module. The serial part is mainly used for human-computer interaction and low speed data transceiver. The conversion between the TTL/CMOS level and the ordinary serial port level is achieved by using the level conversion chip (MAX3232). The JTAG port section is mainly used for the Boot loader download.

Hardware Structure Design of the Handheld Terminal

Hardware block diagram of the handheld terminal is shown in Figure. Compared with the sending end terminal, the handheld terminal removes the video capture module. At the same time, it increases the key control module and TFT LCD display module. The LCD display module is designed to display the data sent from the sending terminal. The key control module is used to control the sending terminal via the Wi-Fi wireless communication module. The photograph of the hardware platform is shown in Figure.

2.5 Software Design
Windows operating system is adopted as the software platform of the system. For firmware development of mbed LPC1768 online compiler of mbed is used according to the algorithm. While the firmware development is completed then it is loaded into mbed LPC1768 ARM. Once the firmware is loaded into the mbed LPC1768 then it starts to work as desired according the algorithm.

                      Algorithms and Flow Chart






                Pattern Recognition
In machine learning, pattern recognition is the assignment of a label to a given input value. An example of pattern recognition is classification, which attempts to assign each input value to one of a given set of classes (for example, determine whether a given email is "spam" or "non-spam").
                Pattern recognition algorithms generally aim to provide a reasonable answer for all possible inputs and to do "fuzzy" matching of inputs. This is opposed to pattern matching algorithms, which look for exact matches in the input with pre-existing patterns. A common example of a pattern-matching algorithm is regular expression matching, which looks for patterns of a given sort in textual data and is included in the search capabilities of many text editors and word processors. In contrast to pattern recognition, pattern matching is generally not considered a type of machine learning, although pattern-matching algorithms (especially with fairly general, carefully tailored patterns) can sometimes succeed in providing similar-quality output to the sort provided by pattern-recognition algorithms.

                There are many pattern matching algorithms like………..
1.                   K-Nearest Neighbor Algorithm.
2.                   Support vector machine Algorithm.
3.                   Fuzzy Neural Algorithm.
4.                   Regression Analysis etc.







                










       

কোন মন্তব্য নেই:

একটি মন্তব্য পোস্ট করুন