Based On Iris Trick Recognition Status Experimental Platform Realization

The iris trick recognition is a technology content high, browses multily-disciplinary, the multi-domain high technology and new technology.Along with the IT development, the superior characteristic and the widespread application prospect which the iris distinguishes, causes it to become the domestic and foreign academic circles and the industrial field research hot spot and the focal point gradually. The iris trick recognition standardized work internationally already obtained the widespread value, but at present domestic works the value to this aspect which also lacked should have, this article in view of the present domestic this domain weak present situation, developed and corresponded the connection aspect to carry on the following exploring research work in the making iris recognition system application:(1) this article first aims at the single frame image pretreatment, the clarity appraisal and the quality appraisal algorithm conducts the research, then synthesizes the single frame image the result and the sequence image general characteristic, with the object-oriented analysis method, the system analysis user demand, the definite iris recognition system function model, the extraction logical organization, has designed the object-oriented lamination iris recognition system logical frame.The logical frame has designed five levels: User level, platform key-course, core algorithm level, characteristic database data level and hardware equipment management level.The level connection design is: System connection, iris processing connection, characteristic data connection and equipment management connection.Has defined various logical level concrete function request and the level the mailing address.(2) pair of logical frame various levels have carried on the function module division, in order to guarantee the system whole the logical uniformity, various functions module in logic also for level structure.This system iris treating processes according to function seal for following function module: Hardware equipment management, picture quality appraisal, iris localization, iris standardization, iris characteristic expression, iris characteristic code, iris characteristic storehouse management, iris characteristic pattern matching.Has in detail defined the function which various functions module must seal and needs the standard the connection, as well as between various modules level information relates alternately.(3) multianalysis system during each kind of level correspondence demand, the basis biology trick recognition technology international standard, this article system interface standard has carried on the standard from two stratification planes: Data stratification plane and connection stratification plane.In view of this system, the data stratification plane carries on the analysis from the characteristic data format standardization and the exchange data format standardization two aspects.The connection stratification plane has defined the software application procedure and the first floor trick recognition technology module/equipment carries on the correspondence an general method, is composed by some transfer function, these transfer function has defined the construction of data and other correlation information.(4) based on this article system logic portal frame construction and the function module division, as well as during each level mailing address, further carries on the description to this article iris recognition system static state construction of data, namely the establishment system object model, the basis object model and the interface specification request, described this system each stratification plane interface standard concrete agreement form through the kind of chart way.Thus has completed the object-oriented lamination iris recognition system construction design.(5) system in order to realize the dynamic load and the convenience carries on the algorithm renewal replace, must follow certain design standard in the algorithm realization.This article uses dynamic link storehouse (dll) the method realization: Each function module algorithm all by dynamic storehouse way realization, moreover each function module only allows to output a syntagma object, in this syntagma object records each function module the basic information, like the module edition number, the name and so on, in which naming rule take as far as possible manifests the module the function as a criterion; Each function module operation connection, namely each function module (kind) defines public function; Each function module construction of data, namely in each function module defines attribute.(6) in this laboratory, to based on this article frame iris recognition system platform each level, has carried on the level function confirmation, the interface specification rational confirmation and the system performance confirmation.In each level selected the specific technical algorithm to carry on the experiment, in order to confirmation system at the same time realized certain degree algorithm research.The system confirmation experiment process, logic linked up linearly has completed the complete function which the iris distinguished.The concrete confirmation work is:The equipment management level confirmation experiment, as a result of the presenttechnical closeness, has only carried on under the platform seal management and theinterface specification condition description after this laboratory patent gatheringinstallment;Core algorithm level confirmation experiment including picture quality appraisal,iris localization, iris standardization, characteristic expression, characteristic code,characteristic data experiments and so on safe code;Picture quality appraisal: Has carried on the confirmation experiment usingsequence picture quality appraisal algorithm seal inserting;Iris localization: The use with asked the fine iris localization method based on thepre-superior voting to carry on the confirmation gradually, the confirmation resulthad simultaneously proven this algorithm had the localization speed quickly, theprecision high characteristic;Iris standardization: Using iris direction energy maximum value method realizationiris texture correction for direction; The iris texture normalization, under in the usepolar coordinate iris coverage the hypothesized circle division and the coordinate transformation form the annulus to radiate the region, then uses the double linear interpolation algorithm to map the reference point under the Cartesian axes to form the image, this article the iris texture normalization for the size rectangular region; The iris texture radiance adjusts, supposes the iris texture along with the light change for the linear variation process, obtains this spot material according to the linear light reflection model the echo strength, again uses the partial histogram equalizing method to strengthen the partial radiance, obtains the equalizing the image brightness, thus reduces the influence which the light change brings; Characteristic expression: Carries on the iris texture using the GHM multi-wavelets the multi-criterion analysis and the expression, seals this algorithm and inserts to this article system platform under, has carried on the module function confirmation and the system performance evaluation;Characteristic code: The choice illumination, the noise insensitive also the recognition separable degree high innertube carries on the iris characteristic to indicate that, establishes the multi-criteria the partial direction characteristic code. This article completes the prime task is:(1) builds a unified experimental platform, uses the existing iris storehouse, withdraws the multi-match algorithm regarding the multi-characteristics to carry on the massive experimental comparison, found the most superior algorithm, from this and carries on the discussion to the iris characteristic, theoretically gives the iris characteristic the model.(2) realized the iris from online gathering, video frequency gathering, regarded the frequency division frame, the image preservation, sequence image step-type screening, the fast quality appraisal, the disturbance withdraws and marks, the iris match recognition, the result analysis and the processing unification tests the platform software; For will be automatic the embedded development which the iris gathered to build the solid foundation and to provide the experimental sample software in the future.This article take the science theoretical analysis as a cornerstone, as stands take the detailed empirical datum according to, designed and confirms the object-oriented lamination iris trick recognition system frame iris to distinguish the platform, the system frame has the good logical rationality, the system platform has certain stability, the reusability, the compatibility and the extendibility.