Font Size: a A A

Zebrafish Larva 3D Morphology Modelling Based On Extremum Search And MPC Method

Posted on:2019-08-01Degree:MasterType:Thesis
Country:ChinaCandidate:X X ShangFull Text:PDF
GTID:2370330566498157Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
Zebrafish larva(Danio rario),as one of the three main vertebrate model animals,has the genetic similarity of 70 percent with human being.Due to the attributes that include its short development period(taking about 4 days before the larva could swim by itself),transparent larva body that is very convenient to observe its inner organ development,and its relatively low feeding cost,Zebrafish larva has attracted a lot of biomedical scientists' attention.Zebrafish larva oriented micromanipulation system is designed to assist biomedical scientists to perform fully-automated or selfautomated tasks,which could be divided into two categories: control and observation.Micromanipulation system associated with control objective is capable of injecting foreign material into Zebrafish larva organs(such as its brain and heart),Zebrafish embryos,or just screening Zebrafish embryos according to different requirements;on the contrary,micromanipulation system associated with observation objective aims to acquire and analyze images in specific views and times to reflect Zebrafish development stages.This paper designs a Zebrafish larva imaging system with high performance-cost ratio and high efficiency,which could be used to obtain a three dimensional model of Zebrafish larva,and further assists scientists to study Zebrafish more easily.Considering existing commercial Zebrafish larva imaging systems are expensive,and have complicated structures that greatly limit its useage,this paper provides a novel and easy platform to reconstruct Zebrafish larva.This paper can be mainly organized as follows.In order to observe Zebraifsh larva at any required orientations,we design a 4-DOF(degree-of-freedom)micromanipulator with high precision.The manipulator consists of three translational movement axes and one rotational movement axis.By using this special structure,we further presents a novel method to control the rotation of Zebrafish larva.The rotation method first needs to capture Zebrafish larva's tail to hold its body,then rotates the larva by its texture on its body to control the rotation degree.During this process,we utilize the extremum searching algorithm to fast orient Zebrafish larva to some specific degree(the commonly used orientations).To reduce the perturbation movement of the pipette tip that grasps the Zebrafish larva's tail,we further presents extremum searching algorithm in control area and a series of image pre-processing algorithms in image processing area.Using the present method,the captured Zebrafish larva iamge frame at different orientations could be obtained,and the three dimensional model of larva could be reconstructed.In order to solve the problem of the perturbation movement of the pipette tip,we further presents a reconstruction method based on the model predictive control(MPC)method.In this method,we donot need to use pipette puller to generate specific pipette tip,but only use the original pipette with an appropriate pipette diameter.The pipette is moved to slightly approach Zebraifsh larva tail,and the micropump that connects the pipette with a soft tube sucks he petri dish solution into itself.The negative pressure is capable of inhaling Zebrafish larva into the pipette entirely.Using the image processing algorithm to detect Zebrafish real-time position,the larva could be easily controlled at any desired point in the field of view(FOV).The ‘piston-airwater' model is used to drive the piston to further control the Zebrafish larva position.Experimental results have proved that this method could effectively reduce the perturbation problem generated by the rotation degree of freedom of the manipulator,and could finally give a better three dimensional model of Zebrafish larva.
Keywords/Search Tags:Zebrafish larva, three dimensional reconstruction, model predictive control, extremum searching algorithm
PDF Full Text Request
Related items