Research On Deformation Detection Of Electro Active Polymer Based On Image Analysis

Electro Active Polymer(EAP)has a great research and application value.It can transform mechanical energy and electric energy into each other.The manufacture of flexible machinery through EAP is a newly emerging field in industry.The EAP can be applied to many projects,including flexible robots,lithium-ion battery,ele-activate diffusion barrier,portable generator,drive material and other fields.However,the deformation of EAP is a small deformation under high voltage.How to measure this tiny deformation without touching is the first difficult problem to solve.In order to solve the problem above,this paper proposes a method based on image analysis to detect the deformation of EAP.The digital images of the deformation under different experimental conditions of EAP can be taken by digital camera.Through the improved digital image enhancement algorithm,the deformed region and the detail information of the scale as reference standard can be highlighted.Through the improved image segmentation algorithm,the deformed region can be segmented from the background.Through the image analysis algorithm,we can extract the scale region from the enhanced experimental image,and calibrate the actual physical size of a single pixel through the calibration information of scale.Finally,the accurate electro-deformation data of EAP can be obtained.The works are completed as follows:(1)During the collection process of experimental images,too dark image,uneven illumination and unclear details are all unavoidable,which will seriously affect the accuracy of subsequent operations.In order to solve those problems,this paper performs image enhancement on experimental images.This paper has improved the traditional multi scale Retinex algorithm and proposed a multi scale Retinex image enhancement algorithm based on the guidance filter and Gamma correction.Experiments show that the algorithm in this paper can not only suppress uneven illumination,but also enhance the details and edges of images.(2)In order to avoid the over segmentation of the traditional watershed segmentation algorithm,we combine preprocessing before segmentation and over segmentation regions merging after segmentation.In this paper,we select the suitable structural elements to do morphological gradient reconstruction and segment the reconstructed image by image watershed segmentation.Finally,the segmented images can be clustered by SOM-K algorithm and the similar regions are merged to reduce the effect of over segmentation.The experiment shows that the segmentation method not only solves the problem of over segmentation of the traditional watershed segmentation algorithm,but also has a clear boundary of deformed region and scale region.(3)Through the segmentation algorithm in this paper,the deformed region of EAP has been separated from the background.But the scale region and the noise are still affecting the result of the segmentation.In order to not affect the follow-up measurement,this paper extracts the deformed region of EAP through the connected region extraction algorithm based on morphology.Through the binary morphology operation,we can remove smaller interference areas,smoothen the edges of the images and disconnect the deformed region from the narrow regions of interference.Finally,the deformed region is extracted by the maximum connected region extraction algorithm,which is convenient for the subsequent operation and measurement.(4)The actual size of the pixel can be calibrated by the scale in the experimental image.The region of the scale is extracted by the feature of scale line density combining with the edge detection,interfering line filtering and so on.The scale line can not be guaranteed to be absolutely vertical in experiments.In order to ensure the accuracy of experiments,in this paper,we combine edge detection and morphological thinning algorithm with Radon transformation and propose an improved algorithm to calibrate the scale.Then,the region of corrected scale is scanned and the actual size of a single pixel is determined by the trip points 's information.Finally,the area of the EAP's deformed region can be calculated by the formula.