Automatic eyes localization in thermal images for temperature measurement in cattle

Infrared thermography technology (IRT) is a non-invasive method that has been used to measure the temperature from Infrared images. The measured temperature can be correlated to evaluate health status. Due to the ability of IRT for measuring temperature remotely, IRT has succeeded widely in detecting a different kind of diseases and inflammations in the animal. Many studies pointed out that eye is the best region for measuring body temperature. Recent work deals with automation of the measurement as a part of the computer-aided diagnosis system. However, such system still based on localization of the eye manually. This paper proposes a novel automated method for eyes localization in cattle. The proposed algorithm is work based on masking the face by modified ellipse detection algorithm. It also extends to remove the consideration for the position and orientation of target animal. Experimental results show that proposed algorithm is working well for localization purpose and temperature measurement as the final result.

I. Introduction

Several studies have been used to test the health status of animals by using a different kind of measurements such as blood test, rectal temperature, respiratory rate and heart rate [1]. These measurement methods are not useful and can be considered as wasting time and effort when having a lot of animals such as the massive amount of cattle in the farm. Therefore, using Infrared thermography technology (IRT) is the best choice for saving time and effort since it does not required from a human being physically close from the animal. Infrared thermography technology (IRT) is a non-invasive method that has been used to calculate and display temperature as an image. IRT can detect the variations in temperature and detect blood flow through determination the changes in body temperature [2], [3]. The concept behind IRT measures heat radiation that emitted from the surface of objects [4]. There are some restrictions associated with this technology that needs from users to care about them while taking an image by using IRT camera. These factors include sunlight, high humidity as well as heat loss as a result of the wind or when the surface of the body is dirty. Also, the emissivity of radiation for objects and reflection (connectivity with another object) of radiation that comes from other object are also considered as other factors that affected the accuracy of captured results by an infrared camera [4]. Moreover, dirt surface of the body will provide various radiation from the clean surface that means the different temperature in both cases[1]. However, the significant factors or parameters that required to provide to the camera are: emissivity of the object, reflected radiation, the distance between the object and camera and humidity [4].