Jay D. Harmon

Comparison of image feature extraction for classification of swine thermal comfort behavior

Abstract Interactive, behavior-based environmental control for swine production has inherent advantages over the conventional, temperature-based control methods. Correctly distinguishing the thermal comfort behavior of swine is critical for the success of interactive environmental control and appropriate feature selections are the basis of the correct classification. This paper compares four feature selection methods using postural images of young pigs subjected to a range of environmental conditions. Programmable cameras were used to capture the behavioral pictures which were then processed into binary images. Fourier coefficients (8×8); moments (first eight); perimeter and area; and combination of perimeter, area and moments of the binary behavioral images were used as the input patterns to a neural network. With the respective feature selection, the neural net correctly classified 96, 92, 96 and 99% training images and 78, 73, 86 and 90% testing images. Thus, the combination of perimeter, area and moments of the binary images as neural network features gave the best performance in the behavioral classification.

RESPONSES OF GROUP-HOUSED NEONATAL CHICKS TO POSTHATCH HOLDING ENVIRONMENT

The posthatch energetics of neonatal chicks exposed to a variety of environmental conditions was investigated. Specifically, moisture production (MP), sensible heat production (SHP), total heat production (THP), respiratory quotient (RQ), and body mass loss (BML) of breeder chicks during a 50 h posthatch holding period in shipping containers at the ambient temperatures of 20, 25, 30, and 35°C with the concomitant relative humidity of 40%, 30%, 22%, and 17%, respectively, were determined. The average responses of the chicks under these environmental conditions were respectively, 3.8, 4.0, 4.1, and 4.6 g·(kg·h)–1 MP; 7.8, 6.4, 5.7, and 5.2 W·kg–1 SHP; and 10.3, 9.1, 8.5, and 8.4 W·kg–1 THP. RQ was independent of the temperature (P = 0.59) and averaged 0.75. BML increased with temperature (P < 0.01) and averaged 5.3, 5.7, 5.9, and 6.3 g·chick–1, or 13.5%, 14.4%, 14.9%, and 15.9% of the initial body mass, respectively. Chick mortalities at 20°C and 25°C (1.09% and 0.71%) were higher than those at 30°C and 35°C (0.28% and 0.33%) (P < 0.01). Based on the criteria of least thermoregulatory efforts and mortality, the thermal neutrality for the unfed, group-housed neonatal chicks was between 30°C and 32°C. Moreover, the metabolic rate of the fasting chicks at thermal neutrality, 77 kcal·day–1/kg0.75 of this study agreed well with the literature value of 70 kcal·day–1/kg0.75 for fasting homeotherms. This study provides unique data for the design and operation of ventilation and environmental control systems during transportation of baby chicks.

Partial Biofiltration of Exhaust Air from a Hybrid Ventilated Deep-Pit Swine Finisher Barn

A strategy for providing partial biofiltration of a critical minimum amount of ventilation air (CMVR) from a hybrid ventilated swine finishing facility was developed and tested. The CMVR, defined as the minimum treated exhaust air that suppressed nighttime curtain opening movement, was set at 81 m3 h-1 pig-1 with the intention of providing enough fan ventilation to suppress inlet curtain movement during stable atmospheres, providing biofiltering for a high percentage of exhaust air. Two side-by-side 300-head hybrid ventilated deep-pit swine finishing rooms were used for this research, one room as the control (CTL) with the other treatment (TRT). The TRT room was fitted with a wood-chip based biofilter for scrubbing the CMVR. In terms of total room emissions, the TRT room had an average odor emission 37% less than the CTL room. Ammonia emission was 58% lower for the TRT room as compared to the CTL room. The results presented indicate that a strategy of partial biofiltration can result in significant reductions in odor and ammonia emissions when applied to hybrid ventilated swine finishing barns.

Environmental Impacts & Bio-security of Composting for Emergency Disposal of Livestock Mortalities

Carcass degradation rate, environmental impacts, and bio-security of windrow-type composting test units were monitored in replicated seasonal trials to assess the feasibility of using composting for emergency disposal of cattle and other large livestock carcasses. Internal temperatures were highest in test units constructed with corn silage. Test units constructed with ground cornstalks or straw and manure were generally 10-20 °C cooler. O2 concentrations in the core of ground cornstalk test units typically exceeded 15%, while those in corn silage and straw/manure test units were in the 5-10% range during the initial weeks of the trials. Despite differences in core temperature and O2 concentration, soft tissue degradation rates were the same in all test units, taking 4-6 months in units constructed during warm weather, and 8-10 months during cold-weather. It is believed that the less favorable (lower) temperatures in the cornstalks may have been offset by significantly higher O2 concentrations which favor rapid aerobic decomposition. Thirty to 45 cm of cover material proved effective in absorbing and retaining odorous gases and leachate. Odors samples collected from the surface of the mortality composting piles typically had low threshold values (< 1500) that differed little from odors emitted by stockpiles of the cover material alone. Leachate volumes were <2% of the precipitation falling on the test units, and preliminary analyses of 1.2 m soil cores show only slight increases in total C and N concentrations in the top 45 cm. Biosecurity tests indicated that pathogens were effectively retained and inactivated: vaccine strains of two avian viruses were inactivated in <21 days; and <2% of sentinel poultry located near the test units exhibited an immune system response to these viruses.

Energetics of Segregated Early Weaned Pigs

The energetics of early weaned (13 to 16 days old) pigs were examined under four temperature regimes. The pigs, in groups of 10 (0.28 m2/pig), were housed in the indirect calorimeter chambers at initial air temperatures of 31.1, 28.9, 26.7, or 24.4°C which then were decreased by 1.1°C per week for the three weeks of the post-weaning period. The feed efficiency and average daily gain during the three-week trial for the four temperature regimes were 1.30, 360 g/day; 1.33, 380 g/day; 1.38, 370 g/day; and 1.27, 370 g/day, respectively. Overall feed efficiency was significantly lower (P < 0.05) for the coolest and warmest treatments. Average daily gain was not significantly different for any of the treatments (P > 0.05). The average moisture production for the four treatments was: 4.85, 4.26, 4.13, and 3.85 g H2O/kg-h, respectively. The total heat production and sensible heat loss were: 5.53, 2.21; 5.66, 2.71; 5.83, 3.01; and 5.87, 3.22 W/kg, respectively. Respiratory quotient averaged 0.95 to 0.96 for all regimes. All average energetic responses, except respiratory quotient, had significant differences among treatments (P < 0.05). Energetic responses of this study were substantially higher than data currently used by the ASAE Standard for swine nursery design.

Neural Network Analysis of Postural Behavior of Young Swine to Determine the IR Thermal Comfort State

A novel method that may be used to interactively control the micro-environment for young swine was investigated to classify the thermal comfort state of animals. Early weaned pigs at 13 to 16 days of age were housed in groups of 10 pigs in four environmentally controlled chambers (1.52 m × 1.83 m floor space per chamber). Air temperatures inside the four chambers were set at 24.4°C, 26.7°C, 28.9°C, and 31.1°C, respectively, for the first week , and were reduced by 1.1°C each of the following two weeks. Postural behaviors of the pigs (huddling or spreading) were captured every 40 min with programmable cameras installed above the transparent false ceilings of the chambers. The raw behavioral images were processed by thresholding, edge detection, and morphological filtering techniques to separate the pigs (objects) from their background. The processed images were further subjected to Fourier transformation. The Fourier coefficients of the processed images (8×8 features) were then used as the inputs to a neural network, which classified the environment into cold, comfortable, or too warm category for the pigs. The neural network analysis worked quite well, with 131 out of 136 training images (96%) and 51 out of 65 testing images (78%) properly classified. This study demonstrates that an innovative environmental controller which uses the animal behavior, instead of the conventionally used air temperature, as the input variable, is possible for swine production. It is anticipated that the behavior-based automatic controller would lead to improved animal well-being and production efficiency. Future research needs include development of algorithms for automatic image segmentation of the pigs, exploration of alternative feature extraction methods to improve classification accuracy of the neural network, and development and evaluation of the behavior-based controller prototype.

Ammonia, Methane, and Carbon Dioxide Concentrations and Emissions of a Hoop Grower-Finisher Swine Barn

Hoop structures have been used quite widely for animal production in the U.S. due to their lower capital costs and multi-purpose versatility. Hoop barns for grower-finisher (G-F) swine production have attracted attention in China as an alternative, environmentally friendly, and water-saving production system. This study was conducted to assess concentrations and emissions of ammonia (NH3) and greenhouse gases (GHGs) for a hoop G-F pig barn at a commercial pig operation in suburban Beijing, China. The NH3 and GHG concentrations and emissions of the facility were measured for three consecutive days during spring and summer seasons. The results revealed the following hourly gaseous concentrations (mean ±SD, mg m-3): 5.9 ±2.7 NH3, 2,183 ±1,376 CO2, and 4.0 ±2.5 CH4 in spring, and 6.8 ±3.4 NH3, 1,530 ±364 CO2, and 5.0 ±2.3 CH4 in summer. The estimated gaseous emissions averaged, in g pig-1 d-1, 22.7 NH3, 2,003 CO2, and 6.7 CH4, or in g AU-1 d-1 (AU = 500 kg), 124 NH3, 11,264 CO2, and 36.2 CH4. The emission values from this study, while being generally comparable with those reported in the literature, add new information concerning emissions from alternative swine housing system.

Ammonia and PM Emissions from a Tom Turkey Barn in Iowa

Considerable progress has been made toward collection of baseline data on air emissions from U.S. animal feeding operations. However, limited data exist in the literature regarding turkey air emissions. The project described in this paper continuously monitors ammonia (NH3) and particulate matter (PM) emissions from turkey production houses in Iowa (IA) and Minnesota (MN) for one year, with IA monitoring Hybrid tom turkeys and MN monitoring Hybrid hens. Mobile Air Emission Monitoring Units are used in the continuous monitoring. Data collection and analysis has been ongoing since May 2, 2007 for the IA site and October 9, 2007 for the MN site. Based on the one-year measurement at the IA site involving three flocks, daily NH3 emissions (g/d-bird) from the IA turkey house varied from 0.04 to 6.4 (mean of 1.9) for flock 1 (May-Aug), 0.2 to 3.4 (mean of 1.3) for flock 2 (Aug-Dec), and 0.16 to 3.8 (mean 1.4) for flock 3 (Dec-Apr). The PM10 emissions (g/d-bird) were 0.04 to 1.6 (mean of 0.58), 0.04 to 0.39 (mean of 0.2), and 0.04 to 0.82 (mean of 0.37) for flocks 1, 2, and 3, respectively; and the concomitant PM2.5 emissions (g/d-bird) were 0 to 0.11 (mean of 0.048), 0 to 0.05 (mean of 0.021), and 0 to 0.14 (mean of 0.053) for flocks 1, 2, and 3, respectively. Annual mean emissions from the tom turkeys (including downtime emission), expressed as grams of constituent per bird marketed, were 169 g NH3, 40 g PM10, and 4.3 g PM2.5 per bird marketed. Data collection and analysis at the MN site are ongoing.

Air Emissions from Tom and Hen Turkey Houses in the U.S. Midwest

Limited data exist in the literature regarding air emissions from U.S. turkey feeding operations. The project described in this article continuously monitored ammonia (NH3) and particulate matter (PM) emissions from turkey production houses in Iowa (IA) and Minnesota (MN) for 10 to 16 months, with IA monitoring Hybrid tom turkeys (35 to 143 d of age, average market body weight of 17.9 kg) for 16 months and MN monitoring Hybrid hens (35 to 84 d of age, average market body weight of 6.7 kg) for 10 months. Mobile air emission monitoring units (MAEMUs) were used in the continuous monitoring. Based on the approximately one-year measurement, each involving three flocks of birds, daily NH3, PM10, and PM2.5 concentrations (mean ±SD) in the tom turkey barn were 8.6 ±10.0 ppm, 1104 ±719 µg m-3, and 143 (±124) µg m-3, respectively. Daily NH3 and PM10 concentrations (mean ±SD) in the hen turkey barn were 7.3 ±7.9 ppm and 301 ±160 µg m-3, respectively. Daily NH3 concentrations during downtime (mean ±SD) were 38.4 ±20.5 and 20.0 ±16.3 ppm in the tom and hen barns, respectively. The cumulative NH3 emissions (mean ±SE) were 141 ±13.1 and 1.8 ±0.9 g bird-1 for the tom turkeys during 108 d growout and 13 d downtime, respectively, and 52 ±2.1 and 28.2 ±2.5 g bird-1 for the hen turkeys during 49 d growout and 32 d downtime, respectively (the extended downtime for the hen house was to ensure monitoring of one flock per season). The cumulative PM10 emission (mean ±SE) was 28.2 ±3.3 g bird-1 for the tom turkeys during 108 d growout and 4.6 ±2.2 and 0.3 ±0.06 g bird-1 for the hen turkeys during 49 d growout and 32 d downtime, respectively. Downtime in the hen house was of greater duration than would be typically observed (32 d vs. 7 d to 14 d typical). The cumulative PM2.5 emission (mean ±SE) was 3.6 ±0.7 g bird-1 for the tom turkeys during 108 d growout (not monitored for the hen turkeys). Because farm operations will vary in flock number, growout days, and downtime; annual emissions can be calculated from the cumulative emissions and downtime emissions per bird from the data provided. Air emissions data from this study, presented in both daily emission and cumulative per-bird-marketed emission, contribute to the improved U.S. national air emissions inventory for animal feeding operations.

Nonsolar energy use and one-hundred-year global warming potential of Iowa swine feedstuffs and feeding strategies.

Demand for nonsolar energy and concern about the implications of fossil fuel combustion have encouraged examination of energy use associated with agriculture. The United States is a global leader in pig production, and the United States swine industry is centered in Iowa. Feed is the largest individual input in pig production, but the energy consumption of the Iowa swine feed production chain has yet to be critically examined. This analysis examines nonsolar energy use and resulting 100-yr global warming potential (GWP) associated with the swine feed production chain, beginning with cultivation of crops and concluding with diet formulation. The nonsolar energy use and accompanying 100-yr GWP associated with production of 13 common swine feed ingredients are estimated. Two diet formulation strategies are considered for 4 crop sequence x ingredient choice combinations to generate 8 crop sequence x diet formulation scenarios. The first formulation strategy (simple) does not include synthetic AA or phytase. The second strategy (complex) reduces CP content of the diet by using L-lysine to meet standardized ileal digestibility lysine requirements of pigs and includes the exogenous enzyme phytase. Regardless of crop sequence x diet formulation scenario, including the enzyme phytase is energetically favorable and reduces the potential excretion of P by reducing or removing inorganic P from the complete diet. Including L-lysine reduces the CP content of the diet and requires less nonsolar energy to deliver adequate standardized ileal digestible lysine than simply feeding soybean meal. Replacing soybean meal with full-fat soybeans is not energetically beneficial under Iowa conditions. Swine diets including dried distillers grains with solubles and crude glycerol require approximately 50% more nonsolar energy inputs than corn-soybean meal diets or corn-soybean meal diets including oats. This study provides essential information on cultivation, processing, and manufacture of swine feed ingredients in Iowa that can be coupled with other models to estimate the nonsolar energy use and 100-yr GWP of pig production.