Determination of live weight of dairy buffalo (Bubalus bubalis) using visual image analysis

Date

3-2011

Degree

Bachelor of Science in Agricultural Engineering

Major Course

Major in Agrometeorology & Farm Structures Engineering

College

College of Engineering and Agro-Industrial Technology (CEAT)

Adviser/Committee Chair

Ronaldo B. Saludez

Committee Member

Vicente G. Ballaran, Jr., Moises A. Dorado

Request Access

To request access of this material, please email the administrator at uscs-mainlib.uplb@up.edu.ph

Abstract

Two mathematical models were developed to rapidly estimate the live weight of dairy buffalo using visual image analysis (VIA). A total of 23 dairy buffaloes from the Philippine Carabao Center (PCC) were used in the study. Live weight and digital images of each buffalo were taken every two weeks for model development and validation. The first model (Body Dimension Model) established a relationship between the live weight and selected body parameters of dairy buffalo. Measurements of the body length (BL), wither height (WH), fore girth depth (FD), hip height (HH), rear depth (RD), and diagonal length (DL) were performed using the digital images of dairy buffalo. Stepwise regression analysis was carried to determine the significant body parameters affecting the live weight. On the other hand, the second model (Body Pixel Model) related the dairy buffalo live weight to its lateral body surface area. The digital images were filtered and body pixel count was determined to represent the body surface area. Simple linear regression was carried out to establish the relationship between the live weight and body surface area. Results showed that body length (BL), wither height (WH), and foregirth depth (FD) significantly affected changes in live weight. The Body Dimension Model has an R2 of 0.934. Validation of the Body Dimension Model showed an average percent error of 17.69 %. On the other hand, the Body Pixel Model has an R2 and % error of 0.878 and 13.79%, respectively. T-test showed significant relationship between live weight and body pixel count at 10% level of significance. However, studying the effect of different camera angles on the body linear measurements and pixel counting is recommended.

Language

English

LC Subject

Dairy cattle

Location

UPLB College of Engineering and Agro-Industrial Technology (CEAT)

Call Number

LG 993.5 2011 A2 N53

Document Type

Thesis

This document is currently not available here.

Share

COinS