Research On Finger Vein Identification Technology Of Coal Mine Personnel

Coal resources account for a considerable proportion in China's energy consumption,which puts forward higher requirements for efficient and orderly mining of coal resources.Face recognition technology is used in the existing identification system of well entry personnel.Due to the bad working environment at the wellhead,the recognition efficiency is low,which will cause the well entry personnel can not quickly arrive at the designated working face to carry out operations,and affect the work efficiency of coal mine.With the continuous development of science and technology,biometric technology is gradually diversified,The finger vein recognition technology stands out in many technologies,because the vein is located in the human body,basically excluding the possibility of being imitated,and the finger vein collection must be in vivo,this is because only the hemoglobin flowing in the human vein can absorb the infrared light of a specific wavelength,so that the finger vein recognition technology has higher security.At the same time,the acquisition of finger vein is different from the image intake of face recognition,which has better affinity.The first step of the identification method based on finger vein recognition technology is to acquire the original finger vein image by CMOS image sensor.The second step is to complete the finger vein image recognition and matching,which is divided into five processes:image quality screening,region of interest extraction(ROI),image enhancement,finger vein feature extraction,finger vein matching and recognition.For the quality screening of finger vein image.In this paper,the quality of finger vein image is screened according to two criteria.Firstly,the gray value of finger vein image region is extracted.Then,the number of vein points is screened by depth threshold,and the quality of finger vein image is screened by normalized quality score.The experimental results show that the EER is 2.01% without image quality screening,and 0.41% after image quality screening.Aiming at the problem of digital vein image enhancement.In this paper,using the characteristics of Gabor filter,the simulation results of 45°,90°,135°,180°.Finally,four images are fused to achieve the purpose of image enhancement.Aiming at the problem of finger vein feature extraction.Because the traditional single feature extraction will lead to low success rate of final matching recognition,this paper uses four kinds of feature extractors to collect finger vein features at the same time,which are maximum curvature,repeated line tracking,principal curvature,Gabor filter feature extraction.Finally,the voting principle is used to determine the extracted features.If more than 50%,it is determined as finger vein line,otherwise it is determined as background.At the same time,this paper compares the success rate of matching recognition of single feature extractor with the success rate of voting for finger vein features with four feature extractors.When voting with four feature extractors,EER decreases to 0.19%.Aiming at the problem of finger vein image matching and recognition.In this paper,MHD algorithm is used to calculate the Hausdorff distance between the fine nodes(endpoint,intersection)to match the collected digital vein image with the data information stored in the database to complete the matching recognition.