Hardware Design And Realization Of Embedded Face Attendance System Based On DSP

Digital and concealed is the most character at Information epoch, how tocorrectly identify person and Guarantee the information safety? Biometricrecognition depended on the body characteristic carries on a verification ofbody, which can't be duplicated. The characteristic of the human bodyincludes the fingerprint, voice, face, retina, lines of the palm, framework etc.Obtain the characteristics of the human body, convert them into thenumerical information, storage these information into computer and matchingis the kernel of biometric recognition.The human has the proper unique characteristic that can't be duplicatedor stolen and forgotten. On the other hand, the traditional methods ofidentification such as passwords, certificates, IC cards etc, which often causethe falsification, theft and crack as they can be separated from the owner.They couldn't fully be satisfied with the modern social economic activitiesand social safety. Therefore, the personal identification system based on thehuman body biologic characteristic gradually becomes very well known tothe society.Embedded system is a special computer, which takes application as thecenter and takes the calculator technique as the foundation, software andhardware can be cut. It accord with the consumption, function, credibility,cost, physical volume of applied system.The Digital Signal Processor (DSP) is a kind of specializedmicroprocessor designed for realization of digital signal processing quickly inthe real time, which adopts the Harvard structure. [1] It has some goodcharacteristics such as the convenient interface, the good stability, the highaccuracy, the good reusability and convenient integration etc. Along with afast development of DSP, its function is more and more stronger, the functionprice ratio rises continuously, the method of development has been improvedcontinuously.The traditional face recognition system are achieved depending oncomputers, both drivers of peripherals and image capturing depended on thehardware and the operating system in a computer, which result in somedisadvantages such as the high cost, poor function in real time, and can't befit to common consumer etc. This paper focuses on the design of hardwareframework in a small face recognition attendance system, which based onDSP and image sensor. Applying DSP and the image sensor to the bodyidentification system can achieve such advantages as little volume, quickspeed, low cost, considerable reliability and applied foreground.My main work is that chooses DSP chip, periphery chips and designhardware system that satisfied the face recognition program. Hardwaresystem design include the principle diagram, PCB diagram, the welding anddebug of the system and the driver grogram, lastly give the picture capturedby hardware system.BlackfinADSP-532 is the core of the hardware. On the other hand, theperipheral chips primarily consist of asynchronous FLASH, SDRAM, theimage sensor, serial chip, clock, power, resets, and JTAG.. Firstly the systemis initialized and boots along with electrifying, then it loads programs anddata from FLASH to SDRAM. The data mainly includes the face detectionand location template, mean face matrix, eigenface template and matchtemplate etc. When someone wants to pass the door, he or she gives aninterrupt signal to the system by pressing the interrupt key. The image sensorcaptures the image via PPI interface under DSP controlling. Then thecaptured image is transmitted to the off-chip SDRAM. A feature value ofperson face will be achieved here when programs complete to process theimage information. The following task is matching the extracted feature valuewith data stored in the database of face feature, and calculating the thresholdvalue of similar degree and record the attendance log. The log of recognitionwill be stored in server-named PC through RS232 interface. The JTAG of thesystem is used for emulating, debugging on-line. The user operation interfacecan update the information of person's face feature periodically.Although I realize the hardware system in the lab, now the system isonly a prototype machine. Some places need the further improvement. Suchas, adds the function of displaying the user's face and makes the useroperation interface more humanization...