Noninvasive Three-dimensional Localization Of Superficial Veins Based On Near Infrared

Venipuncture is one of the most common medical procedures,which widely used in the medical industry.It includes intravenous injection,intravenous blood transfusion,fluid infusion,venous blood collection.With regard to venipuncture,it is most important to be able to visualize clearer venous vessels or predict their depth in those people with thicker fat.However,the failure rate of venipuncture usually occurs due to the difference of the thickness,shallower,fat thickness and skin color between individuals.At present,venipuncture relies too much on the feeling and experience of medical staff.Especially in the case of the new coronavirus outbreak in 2020,medical staff must wear protective clothing,glasses and two layers of rubber gloves,which will undoubtedly lead to a sharp rise in the failure rate of venipuncture.It may delay rescue and even endanger the lives of patients.Up to now,there is no special medical aid to solve this problem.Therefore,a new medical aid is needed to address these issuesThis paper summarizes the research status of venipuncture auxiliary equipment at home and abroad,analyzes the advantages and disadvantages of the implementation scheme.We explored and designed our own experimental protocols: First,the sensitivity of NIR light was used to obtain NIR light vein images based on hemoglobin in blood vessels is specific for NIR light.However,it cannot be directly applied to assist venipuncture because the original image has problems such as uneven illumination and low contrast.Therefore,the contour of the blood vessel is displayed after the original infrared vein image is enhanced,and the two-dimensional position and the blood vessel diameter of the blood vessel is determined by an appropriate algorithm.A skin model was established combined with the principle of light propagation in biological tissue.Furthermore,a mathematical model is proposed to measure the subcutaneous depth of dorsal hand vein by using the gray value of multi wavelength infrared vein image.As expected,it can provide more detailed auxiliary information for medical staff to carry out venipuncture integrated with the two-dimensional position and depth information of blood vessels,The specific experimental scheme and main results are as follows:(1)Firstly,based on the difference of near-infrared light absorption between vein and its surrounding tissue,the infrared vein image acquisition environment was designed and constructed.It includes installing infrared band-pass filter for the camera,drawing the circuit diagram of multi wavelength infrared light source and PCB board,making the finished product,and controlling the light source cycle lighting with STM32F103 MCU;(2)Secondly,different infrared wavelengths of the dorsal hand vein images were collected for enhancement processing,that is,the image contrast was magnified by CLAHE algorithm.After extracting the blood vessel contour by binarization,the edge of the blood vessel image is smoothed by using the opening and closing operation in morphological processing.The results of the above experiments indicate that the match of CLAHE + binarization + morphological opening and closing operation has some advantages for the processing of two-dimensional infrared vein images,and the detection of blood vessels is in line with the expected results;(3)Finally,the three-dimensional information of blood vessels is recovered from the multi wavelength two-dimensional vein gray-scale images obtained above.The skin model was first established,and the gray scale value at the median of the vessel was second obtained from multi wavelength infrared vein images.The mathematical model was established by the relationship between gray value and vascular depth.Next,the parameter matrix is generated by using the broadening matrix of the vascular depth component of the ultrasound vascular image to obtain the subcutaneous depth value of the dorsal hand vein.Finally,the vessel depth obtained by ultrasound angiography is used as the standard and the selfdesigned vessel depth recovery model,which verifies the calculation results after error correction.There are 91.50%errors less than 0.015 cm and 97.82%errors less than 0.021 cm in the calculation.The results of this experiment part shows that this ingenuity mathematical model can meet the demand in terms of depth recovery and can clearly show the subcutaneous depth of the blood vessel.In this paper,we first locate the two-dimensional position of the vein,and then recover the depth information of the vein from the two-dimensional scene.The three-dimensional position of the blood vessel was further obtained.The results of these experiments provide reference for medical staff in venipuncture.At the same time,the location of blood vessels is helpful to the automation of intravenous injection and the study of the direction of the dorsal vein of the opponent.