Automatic Lameness Detection in a Milking Robot : Instrumentation, measurement software, algorithms for data analysis and a neural network model

The aim of this thesis is to develop a fully automatic lameness detection system that operates in a milking robot. The instrumentation, measurement software, algorithms for data analysis and a neural network model for lameness detection were developed. Automatic milking has become a common practice in dairy husbandry, and in the year 2006 about 4000 farms worldwide used over 6000 milking robots. There is a worldwide movement with the objective of fully automating every process from feeding to milking. Increase in automation is a consequence of increasing farm sizes, the demand for more efficient production and the growth of labour costs. As the level of automation increases, the time that the cattle keeper uses for monitoring animals often decreases. This has created a need for systems for automatically monitoring the health of farm animals. The popularity of milking robots also offers a new and unique possibility to monitor animals in a single confined space up to four times daily. Lameness is a crucial welfare issue in the modern dairy industry. Limb disorders cause serious welfare, health and economic problems especially in loose housing of cattle. Lameness causes losses in milk production and leads to early culling of animals. These costs could be reduced with early identification and treatment. At present, only a few methods for automatically detecting lameness have been developed, and the most common methods used for lameness detection and assessment are various visual locomotion scoring systems. The problem with locomotion scoring is that it needs experience to be conducted properly, it is labour intensive as an on-farm method and the results are subjective. A four balance system for measuring the leg load distribution of dairy cows during milking in order to detect lameness was developed and set up in the University of Helsinki Research farm Suitia. The leg weights of 73 cows were successfully recorded during almost 10,000 robotic milkings over a period of 5 months. The cows were locomotion scored weekly, and the lame cows were inspected clinically for hoof lesions. Unsuccessful measurements, caused by cows standing outside the balances, were removed from the data with a special algorithm, and the mean leg loads and the number of kicks during milking was calculated. In order to develop an expert system to automatically detect lameness cases, a model was needed. A probabilistic neural network (PNN) classifier model was chosen for the task. The data was divided in two parts and 5,074 measurements from 37 cows were used to train the model. The operation of the model was evaluated for its ability to detect lameness in the validating dataset, which had 4,868 measurements from 36 cows. The model was able to classify 96% of the measurements correctly as sound or lame cows, and 100% of the lameness cases in the validation data were identified. The number of measurements causing false alarms was 1.1%. The developed model has the potential to be used for on-farm decision support and can be used in a real-time lameness monitoring system.

The importance of modelling heterogeneity in complex disease: application to NIMH Schizophrenia Genetics Initiative data

As for other complex diseases, linkage analyses of schizophrenia (SZ) have produced evidence for numerous chromosomal regions, with inconsistent results reported across studies. The presence of locus heterogeneity appears likely and may reduce the power of linkage analyses if homogeneity is assumed. In addition, when multiple heterogeneous datasets are pooled, inter-sample variation in the proportion of linked families (alpha) may diminish the power of the pooled sample to detect susceptibility loci, in spite of the larger sample size obtained. We compare the significance of linkage findings obtained using allele-sharing LOD scores (LOD(exp))-which assume homogeneity-and heterogeneity LOD scores (HLOD) in European American and African American NIMH SZ families. We also pool these two samples and evaluate the relative power of the LOD(exp) and two different heterogeneity statistics. One of these (HLOD-P) estimates the heterogeneity parameter alpha only in aggregate data, while the second (HLOD-S) determines alpha separately for each sample. In separate and combined data, we show consistently improved performance of HLOD scores over LOD(exp). Notably, genome-wide significant evidence for linkage is obtained at chromosome 10p in the European American sample using a recessive HLOD score. When the two samples are combined, linkage at the 10p locus also achieves genome-wide significance under HLOD-S, but not HLOD-P. Using HLOD-S, improved evidence for linkage was also obtained for a previously reported region on chromosome 15q. In linkage analyses of complex disease, power may be maximised by routinely modelling locus heterogeneity within individual datasets, even when multiple datasets are combined to form larger samples.

Zygosity diagnosis in the absence of genotypic data: An approach using latent class analysis

For zygosity diagnosis in the absence of genotypic data, or in the recruitment phase of a twin study where only single twins from same-sex pairs are being screened, or to provide a test for sample duplication leading to the false identification of a dizygotic pair as monozygotic, the appropriate analysis of respondents' answers to questions about zygosity is critical. Using data from a young adult Australian twin cohort (N = 2094 complete pairs and 519 singleton twins from same-sex pairs with complete responses to all zygosity items), we show that application of latent class analysis (LCA), fitting a 2-class model, yields results that show good concordance with traditional methods of zygosity diagnosis, but with certain important advantages. These include the ability, in many cases, to assign zygosity with specified probability on the basis of responses of a single informant (advantageous when one zygosity type is being oversampled); and the ability to quantify the probability of misassignment of zygosity, allowing prioritization of cases for genotyping as well as identification of cases of probable laboratory error. Out of 242 twins (from 121 like-sex pairs) where genotypic data were available for zygosity confirmation, only a single case was identified of incorrect zygosity assignment by the latent class algorithm. Zygosity assignment for that single case was identified by the LCA as uncertain (probability of being a monozygotic twin only 76%), and the co-twin's responses clearly identified the pair as dizygotic (probability of being dizygotic 100%). In the absence of genotypic data, or as a safeguard against sample duplication, application of LCA for zygosity assignment or confirmation is strongly recommended.

A need for re-examination of the fibre diffraction data of A-DNA

A-DNA pattern, obtained using a flat plat camera, was indexed by Fuller Image on the basis of a c-face centred monoclinic cell with A = 22.24 Å, B = 40.62 Å, C = 28.15 Å and β = 97.0°. A precession photograph of A-DNA which gives an undistorted picture of the lattice, showed that the unit cell parameters as given by Fuller Image were not quite correct. The precession photograph showed a strong meridional reflection (R = 0.00 Å−1) on the 11th layer line. But the occurrence of the meridional reflection on the 11th layer line could not be explained on the basis of the cell parameters given by Fuller Image ; using those cell parameters the reflection which comes closest to the meridian on 11th layer line is at R = 0.025 Å−1. However, a simple interchange of a and b values accounted for the meridional reflection on 11th layer line. The corrected cell parameter refined against 28 strong spots are A = 40.75 Å, B = 22.07 Å, C = 28.16 Å and β = 97.5°. In the new unit cell of A-DNA, the packing arrangement of the two molecules is different from that in the old one. Nonetheless, our earlier contention is again reaffirmed that both right and left-handed A-DNA are stereochemically allowed and consistent with the observed fibre pattern.

Correlation between NIRS generated and chemically measured feed quality data in barley (Hordeum vulgare), and potential use in QTL analysis identification

A study was performed to investigate the value of near infrared reflectance spectroscopy (NIRS) as an alternate method to analytical techniques for identifying QTL associated with feed quality traits. Milled samples from an F6-derived recombinant inbred Tallon/Scarlett population were incubated in the rumen of fistulated cattle, recovered, washed and dried to determine the in-situ dry matter digestibility (DMD). Both pre- and post-digestion samples were analysed using NIRS to quantify key quality components relating to acid detergent fibre, starch and protein. This phenotypic data was used to identify trait associated QTL and compare them to previously identified QTL. Though a number of genetic correlations were identified between the phenotypic data sets, the only correlation of most interest was between DMD and starch digested (r = -0.382). The significance of this genetic correlation was that the NIRS data set identified a putative QTL on chromosomes 7H (LOD = 3.3) associated with starch digested. A QTL for DMD occurred in the same region of chromosome 7H, with flanking markers fAG/CAT63 and bPb-0758. The significant correlation and identification of this putative QTL, highlights the potential of technologies like NIRS in QTL analysis.

Choosing a fishery's governance structure using data poor methods

Multi-species fisheries are complex to manage and the ability to develop an appropriate governance structure is often seriously impeded because trading between sustainability objectives at the species level, economic objectives at the fleet level, and social objectives at the community scale, is complex. Many of these fisheries also tend to have a mix of information, with stock assessments available for some species and almost no information on other species. The fleets themselves comprise fishers from small family enterprises to large vertically integrated businesses. The Queensland trawl fishery in Australia is used as a case study for this kind of fishery. It has the added complexity that a large part of the fishery is within a World Heritage Area, the Great Barrier Reef Marine Park, which is managed by an agency of the Australian Commonwealth Government whereas the fishery itself is managed by the Queensland State Government. A stakeholder elicitation process was used to develop social, governance, economic and ecological objectives, and then weight the relative importance of these. An expert group was used to develop different governance strawmen (or management strategies) and these were assessed by a group of industry stakeholders and experts using multi-criteria decision analysis techniques against the different objectives. One strawman clearly provided the best overall set of outcomes given the multiple objectives, but was not optimal in terms of every objective, demonstrating that even the "best" strawman may be less than perfect. © 2012.

Learning from insurance data: Injuries to other road users in motorcyclist at-fault crashes

In multi-vehicle motorcycle crashes, the motorcycle rider is less likely to be at-fault but more commonly severely injured than the other road user. Therefore, not surprisingly, crashes in which motorcycle riders are at-fault and particularly the injuries to the other road users in these crashes have received little research attention. This paper aims to address this gap in the literature by investigating the factors influencing the severity of injury to other road users in motorcyclist-at-fault crashes. Five years of data from Queensland, Australia, were obtained from a database of claims against the compulsory third party (CTP) injury insurance of the at-fault motorcyclists. Analysis of the data using an ordered probit model shows higher injury severity for crashes involving young (under 25) and older (60+) at-fault motorcyclists. Among the not at-fault road users, the young, old, and males were found to be more severely injured than others. Injuries to vehicle occupants were less severe than those to pillions. Crashes that occurred between vehicles traveling in opposite directions resulted in more severe injuries than those involving vehicles traveling in the same direction. While most existing studies have analyzed police reported crash data, this study used CTP insurance data. Comparison of results indicates the potential of using CTP insurance data as an alternative to police reported crash data for gaining a better understanding of risk factors for motorcycle crashes and injury severity.

Dingoes at the doorstep: Preliminary data on the ecology of dingoes in urban areas

Wild carnivores are becoming increasing common in urban areas. In Australia, dingoes exist, in most large cities and towns within their extended range. However, little empirical data is available to inform dingo management or address potential dingo–human conflicts during urban planning. From GPS tracking data, the nine dingoes, predominately juvenile and female, we tracked lived within 700 m of residential homes at all times and frequently crossed roads, visited backyards and traversed built-up areas. Home range sizes ranged between 0.37 km2 and 100.32 km2. Dingoes were mostly nocturnal, averaging 591 m/h between dusk and dawn. Juvenile and adult dingoes spent up to 19% and 72% of their time in urban habitats. Fresh scats from most areas surveyed tested positive to a variety of common zoonoses. These data suggest dingoes are capable of exploiting peri-urban areas and might contribute to human health and safety risks, the significance of which remains unknown.

NEESTIMATOR v2: re-implementation of software for the estimation of contemporary effective population size (N-e) from genetic data

NeEstimator v2 is a completely revised and updated implementation of software that produces estimates of contemporary effective population size, using several different methods and a single input file. NeEstimator v2 includes three single-sample estimators (updated versions of the linkage disequilibrium and heterozygote-excess methods, and a new method based on molecular coancestry), as well as the two-sample (moment-based temporal) method. New features include the following: (i) an improved method for accounting for missing data; (ii) options for screening out rare alleles; (iii) confidence intervals for all methods; (iv) the ability to analyse data sets with large numbers of genetic markers (10000 or more); (v) options for batch processing large numbers of different data sets, which will facilitate cross-method comparisons using simulated data; and (vi) correction for temporal estimates when individuals sampled are not removed from the population (Plan I sampling). The user is given considerable control over input data and composition, and format of output files. The freely available software has a new JAVA interface and runs under MacOS, Linux and Windows.

Improving the prediction of odour impacts from meat chicken farms with detailed ventilation and odour data

Odour from meat chicken (broiler) farms is an environmental issue affecting the sustainable development of the chicken meat industry but is a normal part of broiler production. Odour plumes exhausted from broiler sheds interact with the environment, where dispersion and dilution of the odours varies constantly, especially diurnally. The potential for odour impacts is greatest when odour emission rates are high and/or when atmospheric dispersion and dilution of odour plumes is limited (i.e. during stable conditions). We continuously monitored ventilation rate, on-site weather conditions, atmospheric stability, and estimated odour concentration with an artificial olfaction system. Detailed inspection of odour emission rates at critical times, i.e. dawn, dusk and night time, revealed that maximum daily and batch odour emission rates are not necessarily the cause of odour impacts. Periods of lower odour emission rates on each day are more likely to correspond with odour impacts. Odour emission rates need to be measured at the times when odour impacts are most likely to occur, which is likely to be at night. Additionally, high resolution ventilation rate data should be sought after to improve odour emission models, especially at critical times of the day. Consultants, regulators and researchers need to give more thought to odour emission rates from meat chicken farms to improved prediction and management of odour impacts.

Abundance indices for long-lived tropical snappers: estimating standardized catch rates from spatially and temporally coarse logbook data

The Northern Demersal Scalefish Fishery has historically comprised a small fleet (≤10 vessels year−1) operating over a relatively large area off the northwest coast of Australia. This multispecies fishery primarily harvests two species of snapper: goldband snapper, Pristipomoides multidens and red emperor, Lutjanus sebae. A key input to age-structured assessments of these stocks has been the annual time-series of the catch rate. We used an approach that combined Generalized Linear Models, spatio-temporal imputation, and computer-intensive methods to standardize the fishery catch rates and report uncertainty in the indices. These analyses, which represent one of the first attempts to standardize fish trap catch rates, were also augmented to gain additional insights into the effects of targeting, historical effort creep, and spatio-temporal resolution of catch and effort data on trap fishery dynamics. Results from monthly reported catches (i.e. 1993 on) were compared with those reported daily from more recently (i.e. 2008 on) enhanced catch and effort logbooks. Model effects of catches of one species on the catch rates of another became more conspicuous when the daily data were analysed and produced estimates with greater precision. The rate of putative effort creep estimated for standardized catch rates was much lower than estimated for nominal catch rates. These results therefore demonstrate how important additional insights into fishery and fish population dynamics can be elucidated from such “pre-assessment” analyses.

What kind of data would ascertain real periodicity in the terrestrial impact crater record

Maan törmäyskraaterien ikäjakauman mahdollinen ajallinen jaksollisuus on herättänyt laajaa keskustelua sen jälkeen, kun ilmiö ensimmäistä kertaa raportoitiin joukossa arvostettuja tieteellisiä artikkeleita vuonna 1984. Vaikka nykytiedon valossa on kyseenalaista perustuuko havaittu jaksollisuus todelliseen fysikaaliseen ilmiöön, on kuitenkin mahdollista, että jaksollisuus on todella olemassa ja se voitaisiin havaita laajemmalla ja tarkemmalla törmäyskraateriaineistolla. Tutkimuksessa luotiin simuloidut kraaterien ajalliset tiheys- ja kertymäfunktiot tapauksille, jossa kraaterit syntyvät joko täysin jaksollisella tai satunnaisella prosessilla. Näiden kahden ääritapauksen lisäksi luotiin jakaumat myös kahdelle niiden yhdistelmälle. Nämä mallit mahdollistavat myös erilaisten kraaterien iänmäärityksen epätarkkuuksien huomioonottamisen. Näistä jakaumista luotiin eri pituisia simuloituja kraaterien ikien aikasarjoja. Lopulta simuloiduista aikasarjoista pyrittiin Rayleigh'n menetelmän avulla etsimään jakaumassa ollutta jaksollisuutta. Tutkimuksemme perusteella ajallisen jaksollisuuden havaitseminen kraateriaikasarjoista on lähes mahdotonta mikäli vain yksi kolmasosa kraatereista on jaksollisen ilmiön aiheuttamia, vaikka nykyistä kraateriaineistoa laajempi ja tarkempi aineisto olisi tulevaisuudessa saatavilla. Mikäli kaksi kolmasosaa meteoriittitörmäyksistä on jaksollisia, sen havaitseminen on mahdollista, mutta vaatii huomattavasti tämän hetkistä kattavamman kraateriaineiston. Tutkimuksen perusteella on syytä epäillä, että havaittu kraaterien ajallinen jaksollisuus ei ole todellinen ilmiö.

Crinkles in connectivity: combining genetics and other types of biological data to estimate movement and interbreeding between populations

Marine species generally have large population sizes, continuous distributions and high dispersal capacity. Despite this, they are often subdivided into separate populations, which are the basic units of fisheries management. For example, populations of some fisheries species across the deep water of the Timor Trench are genetically different, inferring minimal movement and interbreeding. When connectivity is higher than the Timor Trench example, but not so high that the populations become one, connectivity between populations is crinkled. Crinkled connectivity occurs when migration is above the threshold required to link populations genetically, but below the threshold for demographic links. In future, genetic estimates of connectivity over crinkled links could be uniquely combined with other data, such as estimates of population size and tagging and tracking data, to quantify demographic connectedness between these types of populations. Elasmobranch species may be ideal targets for this research because connectivity between populations is more likely to be crinkled than for finfish species. Fisheries stock-assessment models could be strengthened with estimates of connectivity to improve the strategic and sustainable harvesting of biological resources.

Improved estimation of size-transition matrices using tag-recapture data

We derive a new method for determining size-transition matrices (STMs) that eliminates probabilities of negative growth and accounts for individual variability. STMs are an important part of size-structured models, which are used in the stock assessment of aquatic species. The elements of STMs represent the probability of growth from one size class to another, given a time step. The growth increment over this time step can be modelled with a variety of methods, but when a population construct is assumed for the underlying growth model, the resulting STM may contain entries that predict negative growth. To solve this problem, we use a maximum likelihood method that incorporates individual variability in the asymptotic length, relative age at tagging, and measurement error to obtain von Bertalanffy growth model parameter estimates. The statistical moments for the future length given an individual’s previous length measurement and time at liberty are then derived. We moment match the true conditional distributions with skewed-normal distributions and use these to accurately estimate the elements of the STMs. The method is investigated with simulated tag–recapture data and tag–recapture data gathered from the Australian eastern king prawn (Melicertus plebejus).

Survival analysis of time-to-event data in respiratory health research studies

This article provides a review of techniques for the analysis of survival data arising from respiratory health studies. Popular techniques such as the Kaplan–Meier survival plot and the Cox proportional hazards model are presented and illustrated using data from a lung cancer study. Advanced issues are also discussed, including parametric proportional hazards models, accelerated failure time models, time-varying explanatory variables, simultaneous analysis of multiple types of outcome events and the restricted mean survival time, a novel measure of the effect of treatment.