Applied Mechanics and Materials Online: 2014-05-23 ISSN: 1662-7482, Vols. 556-562, pp 4998-5002 doi:10.4028/www.scientific.net/amm.556-562.4998 2014 Trans Tech Publications, Switzerland Research on the Application of Digital Images Based on the Computer Graphics Jing Li 1, Bin Hu 2 1 Zhejiang Business Technology Institute, Ningbo, China 2 Jiangxi College of Applied Technology, Ganzhou, China lijing424@yeah.net Keywords: ARM; Linux; Digital art images; Frame difference Abstract. Digital art images which is based on the computer graphics is integrated with computer graphics technology and network transmission technology. The design of the system is a new application in the field of digital art. This paper chooses ARM9 (S3C2440) processor as the hardware platform, and computer graphics Linux operating system as the software platform. Then it puts forward an improved algorithm of digital art images based on frame difference, and completes the design of digital art images system that can be applied to the community monitoring or bank safety unattended. The experimental results show that the system can extract the background images from the sequence video, so as to improve the quality of digital art images. Introduction With social progress and economic development, people's sense of security is more and more strong, and the safety of property required is gradually increased [1]. Due to the traditional video surveillance system is very complicated, and has higher power consumption and spending, also needs people on duty, yet it hasn t been unable to meet the modern requirements of video monitoring field. Digital art image system which is on the basis of computer graphics has been widely applied in real life with the development of computer, the mature of computer graphics technology and the innovation of communication technology [2]. The core processor of this system is ARM9, the operating system is computer graphics Linux, and the core algorithm is the improved algorithm of digital art images, which can implement the digital art images for the scene, and achieve the goal of unattended. Software Platform Linux operating system has provided basis for the system software platform, built the cross compile environment in the PC machine, and installed cross compiler arm-linux-gcc, so as to complete cross-compiler [3]. The boot loader is transplanted on the ARM board. Flash storage device has stored almost all of the data in computer graphics system, and obtained the transplanted kernel of the target board, and the kernel is compiled according to the ARM architecture. YAFFS computer graphics file system which is designed for Flash, has a fast storage speed, a less memory consumption, and a better portability, further it can store the program and the module into the root file system, so that the successfully compiled application program in the PC which is ported to the ARM board, has a normal operation, and also can be restored after power-off [4,5]. It can make YAFFS file system image by mkyaffsimage. In the process of system operation, the storage space of Nor Flash can t meet the long time work state with the stored images being more and more, so the detected moving target image should be transmitted to the monitoring terminal through network to achieve storage, which can save resources and improve the utilization rate. The flow chart of system software is shown in Figure 1. All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of Trans Tech Publications, www.ttp.net. (#69793024, Pennsylvania State University, University Park, USA-17/09/16,00:39:58)
Applied Mechanics and Materials Vols. 556-562 4999 Linux PC operating system Network transmission Real time alarm Audio player Image processing Video capture Camera driver Hardware Platform Fig. 1 Flow chart of system software ARM9 processor S3C3440 which is chose as the system processor has the high frequency of 533MHz, the interface of digital camera, the interface of AC97 codec, and independent 16KB instruction cache and 16KB data cache. Processor integrates a complete hardware resources including NOR Flash, NAND Flash, IDE interface, 10M card CS8900, 100M card DM9000, extended serial chip 16C2550, SDRAM, a channel USB equipment, etc. The structure of system hardware is as shown in Figure 2. Fig. 2 The structure of system hardware NAND Flash is used to save the data of system operating in the process of operating system, application program, user data. This system NAND Flash adopts K9F1208U0M that has 64MB capacity and 2.7V-3.6V programming voltage, and can automatically program and erase command, address, and data reuse IO port, protect hardware data, also has a good reliability, and is easily operated [6]. In this system, bootloader guiding operation programs, Linux kernel source code and Yaffs computer graphics file system are stored in the NAND Flash storage device. SDRAM is with a dual storage structure. In the system, after the power is on, the kernel source code is copied from Flash to SDRAM, read, storage and execution of program are in SDRAM, so the speed of reading and writing CPU data can be greatly improved. Algorithm of Digital Art Image Background difference method, optical flow method, frame difference method are commonly used algorithms for digital art images. Generally, optical flow which has a big time-consuming, and
5000 Mechatronics Engineering, Computing and Information Technology complex calculation, can t satisfy the requirements of accuracy and real-time. Background subtraction has a simple principle that can fast and accurately detect moving targets, and can extract a complete moving target, but it is not easy to directly obtain the background images, and it is also influenced by the variation of light, temperature and other external conditions [7]. The algorithm of frame difference is easy to be implemented, and that complexity of programming is low, which can adapt to changes in the external environment condition, and has a good stability. The interval of time between frames decides the extracted moving object region. In the stationary camera, this paper adopts an improved digital art images algorithm which chooses frame difference as the core [8]. The algorithm mainly includes 4 parts: (a) the background image modeling; (b) algorithm detection to obtain the moving target; (c) frame difference between moving target and background image; (d) get the complete moving target image from the initial moving object. Flow chart of the algorithm is shown in Figure 3. Start Background extraction Original image Frame difference Changed area Changed area and background difference Background area Morphological processing Moving target End Background area No To determine whether the foreground area Yes Fig. 3 Flow chart of the algorithm (1) the background image modeling: the RGB color image sequence is treated by graying, so as to obtain the gray-scale image sequence, then it calculates the occurred maximum frequency of each pixel in the gray-scale image sequences, and the maximum frequency is denoted as the gray value, finally it obtains the background image. (2) extraction of moving target: frame difference is carried out between two adjacent frames in the video image sequence to get image of moving target areas, then the image is processed by difference with background image sequence to get moving target image, finally the calculation between moving target image and the selected threshold is carried out to get the binarized image. (a) adjacent two frames of each pixel in Mk(i,j) are carried out the difference, so it gets the frame difference image Nk(i,j): Nk(i,j)= Mk(i,j)- Mk-1(i,j). (1) (b) difference is carried out between frame difference image Mk(i,j) and background image Bk(i,j)to get the moving target Wk(i,j): Wk(i,j)= Mk(i,j)- Nk(i,j). (2) (c) moving target Wk(i,j)is changed into the binarized image S(i,j): S(i,j)= 0 (Wk(i,j) Th). (3) S(i,j)= 1 (Wk(i,j)<Th). (4)
Applied Mechanics and Materials Vols. 556-562 5001 Where S denotes the binarized output image, 0 and 1 respectively denotes the foreground and background, and Th denotes the threshold. (3) to obtain the complete moving target image: the obtained binarized image will have the edge point and the target lost, so it is necessary to get the binarized image to be removed the noise by median filtering, and then using mathematical treatment to eliminate the target hole, finally it can obtain the complete moving target image. This paper uses TCP/IP protocol of the socket to transmit image data, which is a collection of a lot of network protocols. TCP/IP protocol adopts the request / response mode, and the concrete realization process is that firstly, the client calls the socket function to establish a socket, and then it calls bind () function to bind the local IP address and port number, at last, it calls listen () function to monitor the connection request sent by host computer. TCP/IP protocol uses "three handshake" method to complete a connection when it needs to establish the connection. After the image data is transmitted and received through the send () function and recv () function, the pictures stored in the development board are copied to the main contents of client, so the client browser can access the pictures that come from the main contents of server to see the moving target image. Experimental Results The whole system has been made experiments within the scene. After the system power is on, and the system is stable, the image in camera view is changed, which makes the monitoring picture produce contrast, so as to the system would play music alarm [9]. The displayed pictures are shown in Figure 3. Figure4(a) denotes the extracted original background image. Figure4(b) denotes the moving image of the next frame. Figure4(c) denotes the extracted target image after the extraction of target and morphological processing. Figure4(d) denotes the results of detecting Figure(b) algorithm by traditional frame difference method. (a) Background image (b) Moving area image (c) Moving target image (d) Traditional frame difference method Fig. 4 Experimental results
5002 Mechatronics Engineering, Computing and Information Technology The experimental results show the detection of the improved moving target algorithm of random screenshots from video sequences image. The moving target extracted by traditional frame difference method is vague and incomplete [10]. The algorithm proposed in the paper can completely lock the moving targets, and the detection has a good effect, which can realize the intelligent monitoring purposes. In addition, the client can download the pictures in the storage device by HTTP protocol, view the results, and implement remote video monitoring. Summary The digital art image system designed in this paper which chose ARM9 as the core chip, computer graphics Linux as operating system, the traditional frame difference method as the basis, realized the improvement of digital art images algorithm. After the background extraction, target extraction, noise filtering, morphological processing, and repeated experiments, the algorithm could well segment the foreground target of motion, which demonstrated the robustness of the method. This method had a better detection effects and high accuracy, and was easy to implement and achieved unattended goal. The system had some characteristics with low cost, low power consumption, small size, extensible function, which was applicable to the field of intelligent monitoring, video analysis, and had great application prospect. References [1] Dongshan Wei. Computer graphics Linux application development completely manual [M]. Beijing: People's Posts and Telecommunications Press, 2011: 240-350. [2] Jianquan Yang, Hua Liang, Chengyou Wang. Development and current status of video surveillance technology [J]. Modern electronics technique, 2012(1): 84-88. [3] Shichong Dong, Tianzhen Wang, Gang Xu. The motion detection in video images [J]. Journal of Wuhan University of Technology, 2011(3): 65-67. [4] Jie Yu. Development of USB camera driver in Linux system [J]. Application of electronic technique, 2011(11): 42-65. [5] Zhongmei Ma, Guangyun Ma, Yinghui Xu. ARM computer graphics processor architecture and application [M]. Beijing: Beihang University press, 2011:76-134. [6] Gang Li, Shangbin Qiu, Ling Lin, etc. Digital art image method based on background difference method and frame difference method [J]. Chinese Journal of scientific instrument, 2012(1): 96-964. [7] Zhiling Yang, Kai Wang. Digital image acquisition and processing and application [M]. Beijing: People's Posts and Telecommunications Press, 2011: 457-475. [8] Tianmiao Wang. Design and development of computer graphics system [M]. Beijing: Tsinghua University press, 2011:120-158. [9] Hai Han. Computer graphics implementation of digital art images and tracking system [D]. Dalian: Dalian University of Technology, 2012:57-93. [10] Ruijuan Zhang. Research on digital art image and tracking algorithm based on video image [D]. Dalian: Dalian University of Technology, 2011:36-74.