Study On The Pre-processing Fingerprint Algorithms Based On Embedded System For Criminal Use

While a significant progress has been made in the research and development of automatic fingerprint recognition, the application of the technology in the embedded system does not prevail at present..To improve the performance of automatic fingerprint recognition and its application on criminal use is meaningful both in theory aspect and promoting its application .For this reason, this thesis has discussed the problems of fingerprint pre-processing algorithms.using the knowledge of digital image processing,pattern recognition and computational intelligence,The major contributions of this thesis are listed in the following:Circumgyrating fingerprint images to any arbitrary angles through framing method and mapping method, Put forward two kinds of improved algorithms for rotate bmp images to any arbitary in studying the bmp image rotate algorithms. Revolved a method called "quickly rotate law" to use the add and subtract operations to replace the multiply operation which each pixel's position transformation need to operate. The reflection shot the method exploitation reflection to shoot the rule to resolve lose a true problem, adding the frame rule to limit a reflection to shoot of district, avoid let up to lose effectively operation complications bring to really increase of problem.Segmentalizing the foreground and the background of fingerprint images through variance method, At the foundation that detailed analyzed the existing fingerprint image partitions algorithms, put forward comprehensive partition algorithm. Comprehensive partition algorithm give a law to apply the rule of the dissimilarity partition algorithm respectively according to the different district demarcation, Pass by experiment and contrast analytical certificated that the accuracy of comprehensive partition algorithm is very high, but the time cost is less than other algorithms.the comprehensive partition algorithm is a kind of good method that can apply to each automatic fingerprint recongnizing system.Using the superposition method to construct a directional diagram of fingerprint images which is based on the relevance of their adjacent blocks at the foundation of analyze algorithms of each image filter, put forward a separating filter according to the context of the direction diagram. according to the fingerprint images their special veins partition regulation and the veins, head for different direction of the fingerprint veins applicate different filter. Particularly the quality is very good at the joint of the separating filter, can be better in expiation the contrast degree descend problem which made by Low-Pass filter. the calculation complications compared with the common image smooth filter reduce consumedly, but the result is still ideal. It is a kind of ideal smooth filter under the sistuation that the processing ability of the embeded system is not high.Give binary algorithm, algorithm for thinning the lines of binary images. thus constitute a whole set of fingerprint pre-processing algorithms that suitable for criminal use.This thesis systematically discusses an algorithm of fingerprint image pre-processing used in the embedded system and tries to find an effective way of criminal investigation, which makes good use of fingerprint identification technology in the embedded system.As is shown from many iterative experiments, these algorithms has good effects on fast treatment, operates well in the embedded system and satisfies the requirements of my design. Moreover, the number of distinct fingerprints achieved by using these algorithms is less than that achieved by using other algorithms.