Convolutional neural network (CNN)-based computer vision systems have been increasingly applied in animal farming to improve animal management, but current knowledge, practices, limitations, and solutions of the applications remain to be expanded and explored. The objective of this study is to systematically review applications of CNN-based computer vision systems on animal farming in terms of the five deep learning computer vision tasks: image classification, object detection, semantic/instance segmentation, pose estimation, and tracking. Cattle, sheep/goats, pigs, and poultry were the major farm animal species of concern. In this research, preparations for system development, including camera settings, inclusion of variations for data recordings, choices of graphics processing units, image preprocessing, and data labeling were summarized. CNN architectures were reviewed based on the computer vision tasks in animal farming. Strategies of algorithm development included distribution of development data, data augmentation, hyperparameter tuning, and selection of evaluation metrics. Judgment of model performance and performance based on architectures were discussed. Besides practices in optimizing CNN-based computer vision systems, system applications were also organized based on year, country, animal species, and purposes. Finally, recommendations on future research were provided to develop and improve CNN-based computer vision systems for improved welfare, environment, engineering, genetics, and management of farm animals.

Highlights A broiler mortality removal robot was successfully developed. The broiler shank was the target anatomical part for detection and mortality pickup. Higher light intensities improved the performance of detection and pickup performance. The final success rate for picking up dead birds was 90.0% at the 1000-lux light intensity. Abstract. Manual collection of broiler mortality is time-consuming, unpleasant, and laborious. The objectives of this research were: (1) to design and fabricate a broiler mortality removal robot from commercially available components to automatically collect dead birds; (2) to compare and evaluate deep learning models and image processing algorithms for detecting and locating dead birds; and (3) to examine detection and mortality pickup performance of the robot under different light intensities. The robot consisted of a two-finger gripper, a robot arm, a camera mounted on the robot’s arm, and a computer controller. The robot arm was mounted on a table, and 64 Ross 708 broilers between 7 and 14 days of age were used for the robot development and evaluation. The broiler shank was the target anatomical part for detection and mortality pickup. Deep learning models and image processing algorithms were embedded into the vision system and provided location and orientation of the shank of interest, so that the gripper could approach and position itself for precise pickup. Light intensities of 10, 20, 30, 40, 50, 60, 70, and 1000 lux were evaluated. Results indicated that the deep learning model “You Only Look Once (YOLO)” V4 was able to detect and locate shanks more accurately and efficiently than YOLO V3. Higher light intensities improved the performance of the deep learning model detection, image processing orientation identification, and final pickup performance. The final success rate for picking up dead birds was 90.0% at the 1000-lux light intensity. In conclusion, the developed system is a helpful tool towards automating broiler mortality removal from commercial housing, and contributes to further development of an integrated autonomous set of solutions to improve production and resource use efficiency in commercial broiler production, as well as to improve well-being of workers. Keywords: Automation, Broiler, Deep learning, Image processing, Mortality, Robot arm.

Reducing floor eggs in cage-free (CF) housing systems is among primary concerns for egg producers. The objective of this research was to evaluate the effects of ground robot manipulation on reduction of floor eggs. In addition, the effects of ground robot manipulation on production performance, stress response, bone quality, and behavior were also investigated. Two successive flocks of 180 Hy-Line Brown hens at 34 weeks of this age were used. The treatment structure for each flock consisted of six pens with three treatments (without robot running, with one-week robot running, and with two-weeks robot running), resulting in two replicates per treatment per flock and four replicates per treatment with two flocks. Two phases were involved with each flock. Phase 1 (weeks 35–38) mimicked the normal scenario, and phase 2 (weeks 40–43) mimicked a scenario after inadvertent restriction to nest box access. Results indicate that the floor egg reduction rate in the first two weeks of phase 1 was 11.0% without the robot treatment, 18.9% with the one-week robot treatment, and 34.0% with the two-week robot treatment. The effect of robot operation on floor egg production was not significant when the two phases of data were included in the analysis. Other tested parameters were similar among the treatments, including hen-day egg production, feed intake, feed conversion ratio, live body weight, plasma corticosterone concentration, bone breaking force, ash percentage, and time spent in nest boxes. In conclusion, ground robot operation in CF settings may help to reduce floor egg production to a certain degree for a short period right after being introduced. Additionally, robot operation does not seem to negatively affect hen production performance and well-being.

HIGHLIGHTS  64% of the 96 broiler houses surveyed were at or below the 25 Ω (ohm) ground resistance threshold recommended by the National Fire Protection Association.  58% of traditionally grounded houses with grounding rod and 74% of Ufer grounded houses were at or below the 25 Ω ground resistance threshold, respectively.  Ufer supplemental grounding has become a required or preferred method over traditional ground rods in new broiler house construction to improve earth grounding.  Producers should have their grounding systems inspected annually with an earth ground resistance meter to mitigate both overcurrent and lightning-induced damage. Commercial broiler house construction, maintenance, and insurance costs continue to increase with the adoption of advanced housing control equipment that aid in flock management through automation of environmental control and data collection. Proper earth grounding is essential to protect these systems from lightning strikes or other dangerous electrical surges. However, the quality of grounding systems on broiler farms in Mississippi and Alabama is currently unknown. Therefore, a field survey of electrical grounding systems was performed on 96 commercial broiler houses of varying ages in Mississippi and Alabama. Survey parameters included earth ground resistance (Ω, ohms), house age, and grounding system type (traditional grounding rod or Ufer supplemental ground). 64% of surveyed broiler houses were at or below the National Electrical Code (NFPA 70) recommendation of 25 Ω. 58% of traditionally grounded houses and 74% of Ufer supplemental grounded houses were at or below the 25 Ω ground resistance threshold, respectively. Ufer supplemental grounding has become a required or preferred method over traditional ground rods in new broiler house construction to improve earth grounding. Producers should have their grounding systems inspected annually with an earth ground resistance meter to mitigate both overvoltage and lightning-induced damage.

 Design air temperatures disregard ambient weather effects such as convective heating/cooling and solar radiation.  Design air temperatures may not yield accurate estimates of heat transfer for broiler houses during warm conditions.  The use of sol-air temperature as a design parameter could serve to optimize estimates of heat transfer.  Sol-air temperatures can be predicted with historical meteorological data and used as a design parameter.  This study supports the appropriateness of using sol-air temperature as a design parameter. Thermal stress adversely affects poultry production efficiency, health, and welfare. Poultry house insulation requirements are typically specified based on engineering design air temperatures, which disregards ambient weather effects such as convective heating and cooling, and solar radiation. The objectives of this study were: (1) to monitor external temperatures of a commercial broiler house to verify the suitability of using sol-air temperature as a design parameter for broiler housing design; (2) to use the sol-air temperature to simulate the effects of solar radiation on conductive heat gain during warm weather for a modeled broiler house in varying climatic locations using historical meteorological data. For two 7-day warm season periods (September and May), ambient air, exterior surface temperatures, and solar radiation were recorded for an east facing sidewall of a broiler house in northern Alabama and used to calculate sol-air temperatures. For both periods, maximum daily surface (T surface) and sol-air (T sol-air) temperatures were significantly elevated (P < 0.0001) as compared to maximum ambient air temperatures (T air). Maximum T surface and T sol-air were not significantly different for September (P = 0.2144) and May (P = 0.1544), respectively. Additionally, simulations of conductive heat transfer (gain/loss) using T air and T sol-air were performed for a model structure located in ten different broiler regions in the United States during daytime warm conditions using historical meteorological data. For each simulation, conductive heat transfer calculated using T sol-air was considerably higher when compared to conductive heat transfer calculated using T air. Methods currently used to specify design temperatures for broiler house design and construction can result in inadequate thermal insulation and T sol-air provides improved estimates of conductive heat transfer during daytime conditions.