Research On Image Reconstruction Algorithms Of Electrical Resistance Tomography System For Tuber Monitoring

Electrical resistance tomography(ERT)technology is an important branch of Electrical Tomography.It is being widely applied in many different fields including petroleum,chemical industry and geological exploration,due to its outstanding advantages like fast,non-disturbance,non-radiation and high performance-price ratio.And it has become a research focus in recent years.In this paper,the application of ERT technology has been expanded to the field of agriculture,and the concept of TERT(Tuber Electrical Resistance Tomography)is proposed.TERT system can provide much information about the growth state of plant tuber through the electronic imaging technology,which can provide great help to agricultural irrigation,fertilization and tuber planting and has wide application prospect and practical value.Firstly,through the deep research of the two-dimensional image reconstruction algorithm,the simulation platform based on Matlab software is designed for theoretical study;then,the TERT PC software is developed to make experimentally study with the sensitive field boundary voltage signal,the three-dimensional image is reconstructed and the entire TERT system is finally completed.The main work includes:1.Research of basic imaging algorithm.The comparative research of basic algorithms,especially the iteration and direct method,aging and accuracy,was made to get the appropriate basic algorithm for TERT system.2.The study on combination algorithm.The selected basic algorithms are analyzed and combined to get a kind of new hybrid algorithm,which can make the tuber section image more real and effective.3.Border processing and noisy pixel detection.The imaging precision has been improved through the hybrid algorithm;however,the border is still relatively vague.In order to solve this problem,the Otsu threshold algorithm is used to extract the gray transition region around the image,and then the genetic algorithm and noisy pixel method are introduced into TERT system to binary process border elements and eliminate image noisy pixel,to get the tuber section picture with higher accuracy.4.The contour curve fitting and 3D model reconstruction based on the comparison of physical tubers.The image obtained by the hybrid algorithm is compared with the plant tuber and the tuber contour is found to be kind of hard which does not match the actual situation.So the cubic B spline curve and double cubic B spline surtace technique are used to fit tuber outline and build 3D contour surface,so as to obtain the algorithm which makes 3D model reconstruction to be more similar to the real plant tuber.5.The software design and experimental verification of tuber reconstructed algorithm.The experimental researches are carried out on the real sensitive field boundary voltage signal collected by TERT-I data acquisition system,to reconstruct the 3D model.According to different situations,the software platform and algorithm of the host computer are compared and optimized,so that it can be more approximately with the actual situation.