Font Size: a A A

Research On Bovine Sign Detection Image Based On Sparse Reconstruction Method

Posted on:2020-10-25Degree:MasterType:Thesis
Country:ChinaCandidate:Y Y TianFull Text:PDF
GTID:2393330590981586Subject:Mechanical engineering
Abstract/Summary:PDF Full Text Request
This paper is a non-contact measurement method for bovine sign detection image research based on sparse reconstruction method.This measurement method is non-contact measurement method,and the stress reaction that will not be caused by contact with cow body during measurement may bring injury and danger to the surveyor,which is more effective than traditional measurement tools such as measuring stick,tape measure,semicircle and caliper,thus greatly reducing economic loss,labor force and measurement error.So as to better observe the growth and development of cattle,to promote the development of cattle breeding.The main research contents are as follows:(1)machine vision acquisition device was used to obtain bovine body images: In this paper,the image acquisition can also be obtained by using the drone technology.The UAV acquisition method has also been applied to many fields.This method is more effective,but the drone technology is more suitable for larger occasions,so this article does not consider it first.Use a drone to get an image.In this paper,casio digital camera ex-h15 was used to obtain images with a width of 4320 and a height of3240.(2)preprocessing of the collected images: there will be internal or external interference,such as noise,in the image acquisition stage.Histogram equalization is used to enhance the image quality.(3)extract and match feature points of preprocessed images,and then complete3 d modeling by matching: extract feature points of preprocessed photos with SIFT algorithm and match them.In order to get a better 3d model of cattle,use SFM method to get 3d points of space.(4)finally,the work of calculating the body size of cattle through 3d modeling:the relationship between the pixel and the physical distance is established with the ear tag of cattle as the reference.After establishing the 3d model of cattle,the pixel distance between the height and the straight length of cattle is measured,and then the actual distance between the height and the straight length of cattle is calculated according to the proportion relationship.The measurement of bovine signs using the sparse reconstruction method is a non-contact measurement of the cattle,which has little effect and does not produce a cow's stress response;the measurement accuracy is high,and the image obtained by the image acquisition device is of high quality,and the measurement is high.The precision has little influence;the speed is fast and the efficiency is high.The method does not invest too much labor,and the measurement efficiency is greatly improved;the accuracy is high,and the measurement result measured by the method is objective.The measurement method used in this paper can be said to provide a new measurement method for scientific farming.
Keywords/Search Tags:image processing, non-contact measurement, sparse reconstruction, feature point extraction and matching
PDF Full Text Request
Related items