The Research And Implementation Of 3D Surface Measurement Technology Based On Kinect

In China,the number of people who seek medical treatment for burns every year is very large,second only to traffic accidents.In the treatment of burns,the estimation of burn area has important implications for shock recovery,on-site first aid,nutritional support and choice of surgical methods.It is crucial to accurately identify and estimate areas of burn pathology.At present,the area of burns is mainly estimated by manual,the accuracy is poor,and the subjectivity is strong.This topic comes from the National Natural Science Foundation "Rapid Assessment of ComputerAssisted Burn Area in Damage Control Resuscitation" to study three-dimensional surface measurement techniques applied to the assessment of pathological regions on the human body,and to improve the accuracy of the measurement of burned areas in the clinic by medical personnel.Sex,ease of use and convenience provide an objective and reliable basis for the treatment of burn patients,facilitating the evaluation of medical staff.The existing statistical methods for burn area in China and abroad have some problems such as large error,complicated operation and limited use.The three-dimensional body surface measurement technology based on Kinect is proposed.This technology uses Kinect to obtain the original depth and color data as input,generates a real 3D human model using 3D color point cloud registration technology,and then uses a 3D color point cloud segmentation algorithm to obtain the burned area,thereby assessing burn area.The technology is portable,accurate,inexpensive,non-invasive,and simple to operate.The medical staff can directly use this system to perform burn area statistics without special training.In view of the condition of a burn patient,the patient is not required to pose a particular posture when acquiring the human body model,and the patient can lie flat on the bed.In addition,the system can also be used for trauma,a range of skin lesions(scars,skin diseases,etc.)to measure the range of lesions and obesity diagnosis and treatment.Based on the current popular registration algorithm,this paper improves the traditional ICP(Iterative Closest Point)algorithm,mainly including down sampling,bounding box delineation of search space,elimination of false matching points,and matching point weighting.And Kd-tree-based search acceleration.The paper aims at the problem of salt noise being generated in the boundary of the model after the hand-held Kinect is used to scan the patient in the current experiment.This paper adopts an adaptive threshold method to eliminate the background information and adopts an improved edge detection method to target the white edge of the target object.Excluding,the cumulative error of the model is optimized.Aiming at the problem that the current three-dimensional color point cloud segmentation algorithm is not suitable for medical clinical application scenes,a three-dimensional color point cloud segmentation algorithm is proposed.The algorithm can accurately segment the region based on the color difference.Finally,the surface area is calculated by human body mapping and compared with the traditional measurement methods to illustrate the practicability and accuracy of the system.