Vic Goedseels

Development and application of computer vision systems for use in livestock production

Abstract This paper examines the feasibility of applying computer vision systems to improve health, welfare and efficiency in livestock production. Very little directly relevant literature was revealed when reviewing the subject, so it is examined from first principles. After briefly describing the value of computer vision as a sensor with powerful observational and interpretative ability, the different steps in vision system development are identified and explored. Where possible this examination is related to computer vision work on livestock as well as other biological objects which by their typically varied nature offer meaningful paradigms for the livestockrelated work. The analysis suggests that most operations in livestock production tend to be at the complex end of the spectrum of vision-related problems currently being tackled in agriculture. Hence, only applications which have a significant production or welfare effect will be viable. Another vital element necessary for success in this application is a simultaneous understanding by the system designers of a diverse set of mechanisms (the production process, the interaction between process and sensor, vision algorithm building, and software and hardware systems). This calls for a multi-disciplinary, interactive approach to develop optimal solutions.

Image-analysis parameters as inputs for automatic environmental temperature control in piglet houses.

Abstract An interactive software package has been developed in order to integrate parameters of thermoregulatory behaviour of piglets into an image-analysis procedure. This procedure generates input parameters for controlling the ambient temperature in pig houses. The main techniques of image analysis applied are windowing, thresholding and pixel counting as processed by the operator. Within this concept it is also possible to take into account the increasing surface area of the growing piglets. The generated input parameters for temperature control are: (1) the temporal and spatial occupancy of pigs within a window corresponding to the sleeping area, in order to decide the need for a temperature change; and (2) the proportion of pig pixels within a predefined window bordering the sleeping area in order to decide the need for heating or cooling. This image-analysis technique applied to quantify the thermoregulatory behaviour of a pig is complementary to the principle of engineering a gradient of environmental temperature within a pigs pen.

An assessment of optimal air temperatures in pig houses by the quantification of behavioural and health-related problems

Air temperatures were measured every 2 h in 12 growing-finishing pig houses. All houses were operated on the all-in, all-out, principle. Data were collected for two fattening periods in each house. Stocking density, feeding system, pig type, and the farmers skill were standardized. Every 2 weeks, the houses were visited and live weight, mortality rate, the incidence of coughing and tail biting, and the extent of dirty lying areas were recorded. Air temperature limits could be isolated from the complex of factors affecting behavioural and health problems of pigs observed within these experiments. Sensitive periods within the growth period of the pigs seem to exist. At the onset of the fattening period (20 to 30 kg), pigs which have been transported from other farms need special care. During summer, mortality rate was lowered when the periodicity of the temperature cycles was lowered for 40- to 50-kg pigs, whereas for heavier pigs the mean maximal air temperature was important also. With respect to coughing, a statistically significant negative relation with the air temperature in the pig house was found for all weight classes, with interactions from the number of different temperature cycles within a 24-h period. In order to avoid dirty lying areas for 20- to 40-kg animals, air temperatures should be between 20 to 24°C, whereas for diarrhoea, animals of 40 to 50 kg were especially sensitive to the occurrence of low air temperatures. For minimizing tail biting, an optimal air temperature range of 20 to 22°C is suggested.

Influence of floor type and surface temperature on the thermoregulatory behaviour of growing pigs

Despite the application of environmental engineering and control of air temperature within houses for intensively kept pigs, health and behavioural related problems are still observed which reduce the profit of the producer. Since these problems are influenced by the environment within a pen, an examination has been made of the possibilities of controlling floor temperature which must influence the micro-climate within the pig pen. The first objective of these experiments was to influence the behaviour of the pigs so that they did not lie in the dunging area during the growing period. The second was to find a relationship between floor surface temperature and the temperature of the water feeding the floor; the water temperature is easy to measure and control when pigs are present, but floor temperature is not. A series of experiments was set up with growing pigs (10 to 30 kg) within climatic rooms, in which air temperature, air velocity and floor temperature were controlled. The air flow pattern was the same for all rooms and for all experiments. The lying behaviour, feed intake and growth rate were studied in relation to the above environmental parameters. The results show that the preferred floor surface temperature is dependent on pig age and not on the nature of the floor itself. It was found possible to quantify the relationship between water and floor surface temperature, and the time it took for the floor temperature to reach a steady value. Floor surface temperature depended both on the temperature of the water feeding the floor and on the floor material itself.

ENVIRONMENTAL TEMPERATURE CONTROL BY THE PIG'S COMFORT BEHAVIOR THROUGH IMAGE ANALYSIS

ABSTRACT During four experiments with six female piglets, the thermoregulatory behavior of the piglets was quantified by visual observation and by automated image analysis. The collection number of images representing pigs sleeping side by side and touching each other, as well as the occupation ratio of piglets being calculated on pixel values within predefined windows, may serve as set-point values for environmental temperature controllers. But this observational procedure has to be linked to the engineering of a gradient of an effective environmental temperature within a pen.

Development of a heat exchanger system for modern livestock buildings

Abstract Todays requirements in livestock buildings offer a potential market for heat exchanger technology. To realize an economical heat exchanger system, a complex interaction of many factors must be analysed. To solve the different aspects of the problem, a methodical approach is necessary. Firstly a compact heat exchanger for use in modern livestock buildings is selected. Then a calculation method is presented to predict the effectiveness of the selected system. It is shown how this calculation can be used to define an economic criterion by which a well dimensioned and adapted heat exchanger system can be developed.

Measurement of Ear Base Temperature as a Tool for Sow Management

Ear base and rectal temperature of 21 multiparous sows were measured in order to detect estrus for optimal insemination time. A thermistor was implanted in the ear base and wire-connected to a data-acquisition system, allowing time-sampling with a measuring accuracy of 0.1°C. Air temperature in the neighborhood of the sows was measured with the same equipment. Rectal temperature was measured each day in the afternoon with a veterinary thermometer. Following the rectal temperature measurement, a blood sample was taken to determine oestradiol-17 beta content. The combination of the physiological measurement with the observation of the standing reflex of the sows made estrus detection very reliable. A statistically significant rise of ear base (1.1 ± 0.12°C) and rectal (0.65 ± 0.3°C) temperature was observed two days before estrus. Moreover, when relating variation in ear base temperature to environmental temperature, estrus detection and environmental temperature control may be improved. These results offer new possibilities for introduction of injectable electronic identification and monitoring systems for sow management.

Electronic Identification and Physiological Monitoring: Data Collection Efficiency and Powering Requirements

When measuring pigs’ body temperature subcutaneously with a telemetric device during housing and handling, data collection efficiency was 80% and 30%, respectively. With this over-all performance respective 4 and 10 readings/transmissions are necessary to collect at least three physiologically relevant data for calculating a mean value per reading/transmission. For active transponders, the sampling frequency will therefore determine the life time of the available battery.

Process requirements — Technological options

This chapter describes the general context of the application of solar energy in greenhouses. The purpose is to give a theoretical background of solar energy applications in greenhouses and to indicate the possible techniques for incorporating solar energy into the production process. The incorporation of solar energy in greenhouses is bound to a number of plant physiological limits, depending on climatic conditions. Therefore a brief description of the major production parameters is given: photosynthesis, respiration, transpiration and soil nutrition. The reaction of greenhouse crops to a change of the production parameters indicates the limits of the introduction of energy-saving measures in general and of the solar energy applications in particular (8). Special attention is given to solar energy storage systems.

Present research within the EC on the application of solar energy in the drying sector

The overview of research presented in this chapter is based on the materials received from CEC action 2 and 3 participants. This material is not complete: it neither cover all the research within the EC, nor do the given basic descriptions always contain all the elements, essential in the given context.