The ability to detect moisture in broiler feathers for five moisture-sensing devices using varying techniques, an infrared (IR) camera, and one type of moisture-sensitive paper was evaluated in two experiments; using artificial feather beds in laboratory settings and in field conditions using live birds. In the first experiment (lab testing-phase I), seven levels of moisture were applied once per day to the swatches with four different feather densities to account for moisture variation and feather density present in commercial barn conditions. True moisture of the feathers was determined gravimetrically on a daily basis. Using the five devices, 20 readings each were acquired from each of the swatches. The average temperatures of a selected area from images captured using an IR camera along with average device readings were compared with the true moisture content. Moisture-sensitive paper images were analyzed in Photoshop and Matlab prior to statistical analysis. Data from all devices, the IR camera, and moisture-sensitive paper were analyzed using SAS Procedure GLM to define relationships between the true moisture content and the readings. The devices were analyzed based on their accuracy, consistency and sensitivity using adjusted-R2, standard error, and regression slope, respectively. The data from all devices and techniques were significantly correlated with feather swatch moisture content (P<0.05). Results from the first experiment suggested potential to measure feather moisture by several of the tested devices. While feather density presented as a challenge during this experiment, it was not considered as a significant issue when evaluating the devices. The “Hay” and “Construction 1” sensors showed the most promise in detecting feather moisture and
iv
were selected for further testing using live birds. The two devices had relatively higher accuracy, consistency, and sensitivity compared to other devices and techniques.
The second experiment (field testing – phase II) evaluated the two selected devices (Hay and Construction 1 sensors) in various commercial broiler settings. Device readings were acquired from the back, wing, and breast feathers. A sample of back feathers from each bird was collected to determine the true moisture. Statistical analyses of data were the same as in experiment 1. Although the initial study, conducted within a lab setting, denoted a significant relationship between true moisture content and device readings, testing within the field environments showed the devices to perform poorly. Readings from both devices and for all the locations tested demonstrated a lack of sensitivity, accuracy, and consistency for measuring moisture in feathers of live birds.
| Identifer | oai:union.ndltd.org:USASK/oai:ecommons.usask.ca:10388/ETD-2015-03-1986 |
| Date | 2015 March 1900 |
| Contributors | Crowe, Trever G. |
| Source Sets | University of Saskatchewan Library |
| Language | English |
| Detected Language | English |
| Type | text, thesis |
Page generated in 0.0017 seconds