Study On Extraction Of Grain Feature And Parameter Measurement Of Hand Image

Image-based measurement is widely used in medical,industrial,agricultural,military and national defense fields due to its high detection accuracy,high speed and low cost.Therefore,it has become one of the hottest research topics in the field of computer vision.Elastic gloves are a common means of rehabilitation in hospitals for hand burns.The traditional method is to make elastic gloves by taking measurements of the patient's hands in the ward.In the manual measurement method,the error of measured data is different according to different standards.And the manpower training time is long,the cost is high.Therefore,the use of image-based measurement technology to take photos of patient's hands and measure the parameters of their hands has important practical value in the preparation of elastic gloves.In order to solve the problems of image-based feature detection and geometric parameter measurement,this paper focuses on feature extraction and parameter measurement based on the patient's hand image.The main work of this paper is as follows:1?This paper presents a method to extract palmprint features by combining Block Weighted Locally Binary Pattern(BW-LBP)with MFRAT(modified linear filter of finite Radon transform).This method firstly by using high frequency emphasis filtering image is combined with limited contrast adaptive histogram equalization method of image enhancement,and then and then use the original local Binary characteristics(LBP,Locally Binary Pattern)by a single pixel point and neighborhood pixels in contrast to implement for multiple blocks between the average grey value of comparison,block pixel area average processing can reduce the disturbance of noise,according to the neighborhood within a block point and the size of the center pixel distance weighted arithmetic for LBP values,get a BW-LBP palm texture images after processing;The main line feature image of the current linear template is obtained by convolving each pixel with the multi-directional linear template of MFRAT.Finally,the feature images of all directions are summed to obtain the palm lines and edge feature information.Finally,the feature images of each direction are summed to obtain the palm lines and edge feature information.Experimental results verify the feasibility and effectiveness of the method.2?Aiming at the problem of large error difference of measured data due to different standards of manual measurement,a set of geometric parameter measurement system based on image was designed and implemented.The system first calibrates camera parameters.Then,the acquired patient hand image is filtered to improve the image quality,and the edge feature information of the hand is acquired in real time.Finally,the parameter measurement in the image is realized by taking a point with the mouse.The results of laboratory test and field hospital test show that the system has higher measurement accuracy and efficiency than manual measurement.