Repository logo

Theses and Dissertations

Permanent URI for this collection

https://hdl.handle.net/10217/100385

Browse

Browsing Theses and Dissertations by Author "Alciatore, David, committee member"

Now showing 1 - 1 of 1
  • Thumbnail Image
    ItemOpen Access
    Equine body weight estimation using three-dimensional images
    (Colorado State University. Libraries, 2015) Ku, Kyung-nyer, author; Traub-Dargatz, Josie L., advisor; Salman, Mo, advisor; Alciatore, David, committee member; Hess, Tanja, committee member
    Accurately estimating the body weight (BW) of a horse is important in order to make appropriate management and treatment decisions. Most field equine veterinarians and experienced equine people, however, visually estimate BW because large animal scales are impractical for field use due to the weight (>80 kg), size (length >200 cm), and cost (>$1,000). There are some alternative BW estimation methods such as a weight tape or BW estimation using a combination of heart girth and body length measurements. These methods, however, have 5 - 15% or even higher margin of error. According to human studies, there is a high correlation between BW and body volume (BV). Correlation coefficient (R) between these two variables is 0.996-0.998. Our study was designed to develop methods to estimate the BW of horses by using 3D image based BV measurement. 3D imaging technology allows easy and accurate measurement of diverse indices of an object, including the volume. Recent development of Structure-light 3D scanning technology allows 3D scanning of an object as large as 3 by 3 square meter in a short time. In this study, 3D images of 22 and 11 horses were obtained by using 3D scanning (3DScan) and photogrammetry (2Dto3D), respectively. BV and trunk volume (TV) of the horses were measured from the obtained 3D images. Measurements of BW using five conventional methods (visual estimation, 2 weight tapes (Purina, Shell), estimated BW by using heart girth and body length (Carroll’s formula), and a large animal scale) were also conducted, and the data of body condition score (BCS), sex, coat color, and coat type of the horses were collected. Linear regression models to estimate the BW of the horse based on the volume and other independent variables were developed using regression model stepwise selection procedures (P<0.05). Variables selected in 3DScan method were BV, sex, and coat type, and, in 2Dto3D method, BV (TV) was selected. The coefficient of determination of the developed regression models were 0.95 and 0.78-0.82, respectively, and the average percent errors of the predicted BW compared to the true BW of horses were 2.07 % and 2.67 %, respectively. The accuracy of the 3DScan method was significantly more accurate than WT, Carroll’s formual, and VE (P<0.05). 3D image based BW measurement method had higher accuracy and convenience compared to conventional alternative BW measuring methods. Accurate and easy determination of BW using 3D images will allow for regular BW measurement in the field and allow optimal equine health management by equine stakeholders and practitioners. The 3D images obtained in this study were highly detailed. Further graphical analysis of the obtained 3D images will make it possible to use this technology on automatic evaluation of body condition score, equine conformation evaluation, breed registration, and the study of pharmacokinetics and dynamics of newly developed drugs. This research findings may also have utility for application to wild or zoo animals such as the elephant, rhinoceros, or even the tiger where hands on collection of body weight would be challenging.