Abozar Nasirahmadi

Using machine vision for investigation of changes in pig group lying patterns

Group pig lying change detection using Delaunay triangulation is presented.Ellipse fitting algorithms were used to localize each pig body in the image.Delaunay triangulations were changed as the room temperature increased.Pig lying location is determined through ellipse centroids in the pen. Pig lying patterns can provide information on environmental factors affecting production efficiency, health and welfare. The aim of this study was to investigate the feasibility of using image processing and the Delaunay triangulation method to detect change in group lying behaviour of pigs under commercial farm conditions and relate this to changing environmental temperature. Two pens of 22 growing pigs were monitored during 15days using top view CCD cameras. Animals were extracted from their background using image processing algorithms, and the x-y coordinates of each binary image were used for ellipse fitting algorithms to localize each pig. By means of the region properties and perimeter of each Delaunay Triangulation, it was possible with high accuracy to automatically find the changes in lying posture and location within the pen of grouped pigs caused by temperature changes.

Automatic detection of mounting behaviours among pigs using image analysis

Automatic mounting events detection among pigs is presented.Ellipse fitting algorithms were used to localize each pig in the image.The Euclidean distances between head, tail and sides of pigs were obtained.Major and minor axis lengths were altered during mounting events. Excessive mounting behaviours amongst pigs cause a high risk of poor welfare, arising from skin lesions, lameness and stress, and economic losses from reduced performance. The aim of this study was to develop a method for automatic detection of mounting events amongst pigs under commercial farm conditions by means of image processing. Two pens were selected for the study and were monitored for 20days by means of top view cameras. The recorded video was then visually analysed for selecting mounting behaviours, and extracted images from the video files were subsequently used for image processing. An ellipse fitting technique was applied to localize pigs in the image. The intersection points between the major and minor axis of each fitted ellipse and the ellipse shape were used for defining the head, tail and sides of each pig. The Euclidean distances between head and tail, head and sides, the major and minor axis length of the fitted ellipse during the mounting were utilized for development of an algorithm to automatically identify a mounting event. The proposed method could detect mounting events with high level of sensitivity, specificity and accuracy, 94.5%, 88.6% and 92.7%, respectively. The results show that it is possible to use machine vision techniques in order to automatically detect mounting behaviours among pigs under commercial farm conditions.

A new approach for categorizing pig lying behaviour based on a Delaunay triangulation method

Machine vision-based monitoring of pig lying behaviour is a fast and non-intrusive approach that could be used to improve animal health and welfare. Four pens with 22 pigs in each were selected at a commercial pig farm and monitored for 15 days using top view cameras. Three thermal categories were selected relative to room setpoint temperature. An image processing technique based on Delaunay triangulation (DT) was utilized. Different lying patterns (close, normal and far) were defined regarding the perimeter of each DT triangle and the percentages of each lying pattern were obtained in each thermal category. A method using a multilayer perceptron (MLP) neural network, to automatically classify group lying behaviour of pigs into three thermal categories, was developed and tested for its feasibility. The DT features (mean value of perimeters, maximum and minimum length of sides of triangles) were calculated as inputs for the MLP classifier. The network was trained, validated and tested and the results revealed that MLP could classify lying features into the three thermal categories with high overall accuracy (95.6%). The technique indicates that a combination of image processing, MLP classification and mathematical modelling can be used as a precise method for quantifying pig lying behaviour in welfare investigations.

Influence of Moisture Content, Variety and Parboiling on Milling Quality of Rice Grains

Abstract The effects of moisture content (8%, 10% and 12%), variety (Tarom and Fajr) and parboiling on milling quality of rice as a function of milling recovery (MR), head rice yield (HRY), degree of milling (DOM) and whiteness were investigated. The parboiled grains was prepared with three soaking temperatures of 25 °C, 50 °C and 75 °C and three steaming times of 10, 15 and 20 min. As a result of parboiling, the increasing rates of MR and HRY values were 7.8% and 14.3% for Tarom and 9.8% and 10.0% for Fajr, respectively, and the decreasing rates for DOM and whiteness were 6.6% and 10.8% for Tarom and 6.8% and 10.5% for Fajr, respectively. Moreover, decreasing moisture content to 8% maximized MR (75.8% for Tarom and 74.3% for Fajr) and HRY (65.8% for Tarom and 57.0% for Fajr) while increasing that to 12% revealed maximum values of DOM (6.1% for Tarom and 6.2% for Fajr) and whiteness (24.8% for Tarom and 28.2% for Fajr).

Modelling and analysis of compressive strength properties of parboiled paddy and milled rice

Abstract The present investigation deals with analyzing the compressive strength properties of two varieties (Tarom and Fajr) of parboiled paddy and milled rice including: ultimate stress, modulus of elasticity, rupture force and rupture energy. Combined artificial neural network and genetic algorithm were also applied to model these properties. The parboiled samples were prepared with three soaking temperatures (25, 50 and 75°C) and three steaming times (10, 15 and 20 min). The samples were then dried to final moisture contents of 8, 10 and 12% (w.b.). In general, Tarom variety had higher compressive strength properties for paddy and milled rice than Fajr variety. With increase in steaming time from 10 to 20 min, all mentioned properties increased significantly, whereas these properties were decreased with increasing moisture content from 8 to 12% (w.b.). Coupled artificial neural network and genetic algorithm model with one hidden layer, three inputs (soaking temperature, steaming time and moisture content), was developed to predict the compressive strength properties as model outputs. Results indicated that this model could predict these properties with high correlation and low mean squared error.

Docking piglet tails: How much does it hurt and for how long?

Tail docking in pigs has the potential for evoking short- as well as long-term physiological and behavioural changes indicative of pain. Nonetheless, the existing scientific literature has thus far provided somewhat inconsistent data on the intensity and the duration of pain based on varying assessment methodologies and different post-procedural observation times. In this report we describe three response stages (immediate, short- and long-term) through the application of vocalisation, behavioural and nociceptive assessments in order to identify changes indicative of potential pain experienced by the piglets. Furthermore, we evaluated the following procedural differences: (1) cautery vs. non-cautery docking; (2) length of tail removal. Sound parameters showed a significantly greater call energy and intensity exhibited by docked vs. sham-docked piglets (P<0.05). Observations of general activity of the animals in a test situation failed to detect a difference among treatments (P>0.05) up to 48h post-tail docking. Similarly, no difference in mechanical nociceptive thresholds indicative of long term pain was observed at 17weeks following neonatal tail docking (P>0.05). The present results highlight the potential for the use of measures of vocalisation to detect peri-procedural changes possibly associated with evoked pain. Nonetheless, activity and nociceptive measures failed to identify post-docking anomalies, suggesting that alternative methodologies need to be implemented to clarify whether tail docking is associated with short- and long-term changes attributable to pain experienced by the piglets.

A neural network based model to analyze rice parboiling process with small dataset

In this study, milling recovery, head rice yield, degree of milling and whiteness were utilized to characterize the milling quality of Tarom parboiled rice variety. The parboiled rice was prepared with three soaking temperatures and steaming times. Then the samples were dried to three levels of final moisture contents [8, 10 and 12% (w.b)]. Modeling of process and validating of the results with small dataset are always challenging. So, the aim of this study was to develop models based on the milling quality data in parboiling process by means of multivariate regression and artificial neural network. In order to validate the neural network model with a little dataset, K-fold cross validation method was applied. The ANN structure with one hidden layer and Tansig transfer function by 18 neurons in the hidden layer was selected as the best model in this study. The results indicated that the neural network could model the parboiling process with higher degree of accuracy. This method was a promising procedure to create accuracy and can be used as a reliable model to select the best parameters for the parboiling process with little experiment dataset.