Design And Implementation Of Surgical Navigation And Positioning System Based On Machine Vision

At present,intracranial surgery mainly relies on the experience of doctors,the operation is complicated,and the trauma after surgery is large,which greatly prolongs the recovery period of patients.With the rapid development of artificial intelligence technology,surgical navigation and positioning system has become a current research hotspot.Brain surgery positioning requires the navigation system to have high precision and high real-time performance.At present,the navigation and positioning technology of foreign surgery is relatively high,but the instruments are expensive,and the positioning accuracy of domestic surgical navigation systems is poor.The surgical navigation system has become an important problem that needs to be solved urgently in China.In order to improve the accuracy and real-time performance of surgical navigation and positioning,a surgical navigation and positioning system based on machine vision is studied in this paper.The specific research contents are as follows:Firstly,build an optical surgical navigation system based on machine vision.In order to solve the problems of low puncture accuracy and poor postoperative effect in craniotomy,an optical surgical navigation system for craniotomy was constructed in this paper,and auxiliary aids such as binocular camera,red marker ball,self-made surgical instruments,and patient head mold were designed.Secondly,the three-dimensional reconstruction method of brain images is studied.By comparing the advantages and disadvantages of the four methods of surface rendering and volume rendering,the ray casting method is selected for 3D reconstruction.The ray-casting method used in this paper can produce high-quality images and can well reflect the changes at different tissue boundaries.By carrying out volume rendering experiments,the experimental results show that the ray casting method in this paper has a better effect on 3D reconstruction,which meets the accuracy requirements of medical image processing.Thirdly,patient registration methods for surgical navigation are investigated.By using the red marker sphere as the marker point,the marker sphere can be displayed in the patient space and the image space,and the patient space registration method is realized based on the marker sphere.The patient registration method is performed in two steps: The first step is to realize the rough registration of the patient space and the image space.The coarse registration method of the multi-feature point method is proposed to replace the traditional three-point method.This method can generate better initial transformation values in the initial registration stage;The second step is to achieve precise registration between patient space and image space,and a point-to-surface ICP algorithm is proposed to optimize the traditional ICP algorithm.This method improves the traditional point-to-point distance construction error function to point-to-plane.The distance constructs the error function,which not only further improves the registration accuracy,but also reduces the possibility of the solution falling into the local optimal solution.In this paper,a series of experiments are carried out using the patient head model.Through the actual experimental verification,the results show that the registration error of the method in this paper is less than 2.5mm,which can meet the accuracy requirements of general surgical navigation and positioning systems.