subject: Security Analysis And Image Processing Difficulties Processor Choice - Security - Security Industry [print this page] Security applications, image processing Security applications, image processing
Difficulties and processor selection
- Information provided: Zhanke (ADI Technical Marketing Manager)
Shanghai May 25, 2009 / Xinhua-PRNewswire Asia / - As people required to improve the quality of life and the global trend of anti-terrorism requirements, as well as digital technology advances its own, relying on fingerprint recognition, iris recognition , face recognition and other biometric technology programs and video surveillance programs are becoming the increase of individuals, families, businesses, and an important means of social security. Biometrics program includes four steps: image acquisition, image preprocessing, feature sampling, match analysis; and video surveillance program are mainly including image acquisition, image preprocessing, image processing and transmission, image display and image management. Not difficult to see, whether or biometric video surveillance, image pre-processing is required. In fact, the image pre-processing algorithm flexibility, complexity, image processing chip resources on the occupation degree, and the length of processing time the system will directly affect the operation has a decisive role. Therefore, image pre-processing for the entire security program is a difficult but critical task will determine follow-up image processing and analysis of the accuracy and convenience.
Analysis of image preprocessing
According to different purposes, image pre-processing can be divided into clear images of the collection processing, the image pre-processing before recognition, and image compression before the pretreatment. Among them, a clear image capture processing include image sensor on the CMOS or CCD light-sensitive cells to correct inconsistencies in follow-up on the physical environment and the difference between the image sensor acquisition compensation (such as the backlight), and the collected images to the original noise processing. Although this preprocessing algorithm itself is not difficult, but with the popularity of real-time demand, especially in the larger pixels, this algorithm is on the DSP (http://www.analog.com/embedded-processing- dsp / processors / zh / index.html) presented a very high processing power requirements.
And identify the image pre-processing before the very purpose may need to destroy the original pixels and distribution in order to follow-up feature extraction. This preprocessing algorithm of the difficulty of visual identification of different occasions and different. Recognition algorithm to be integrated later section, select the appropriate DSP. Image preprocessing before compression is mainly refers to the YUV422 into YUV420, YUV to RGB, etc. into. Such treatment is often real-time requirements, if the use of software, processing performance will have higher requirements; if the use of hardware, although the processing performance is guaranteed, but the hardware costs will rise.
The same time, depending on the application of different, image pre-processing can be divided into biometric applications in the image preprocessing and video surveillance applications, the image preprocessing. For biometric applications to fingerprint identification, for example, the pretreatment includes fingerprint image enhancement, binary fingerprint image, fingerprint image thinning, the thinned fingerprint image processing. The video surveillance applications, image preprocessing mainly refers to the image sensor output analysis of continuous images, access to adequate information and through the automatic white balance, gamma (Gamma) correction, auto-focus, auto exposure, backlight compensation to improve the image of the actual results.
