Large-scale phenotyping of animal behaviour traits is time consuming and has led to increased demand for technologies that can automate these procedures. Automated tracking of animals has been successful in controlled laboratory settings, but recording from animals in large groups in highly variable farm settings presents challenges. The aim of this review is to provide a systematic overview of the advances that have occurred in automated, high throughput image detection of farm animal behavioural traits with welfare and production implications. Peer-reviewed publications written in English were reviewed systematically following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. After identification, screening, and assessment for eligibility, 108 publications met these specifications and were included for qualitative synthesis. Data collected from the papers included camera specifications, housing conditions, group size, algorithm details, procedures, and results. Most studies utilized standard digital colour video cameras for data collection, with increasing use of 3D cameras in papers published after 2013. Papers including pigs (across production stages) were the most common (n = 63). The most common behaviours recorded included activity level, area occupancy, aggression, gait scores, resource use, and posture. Our review revealed many overlaps in methods applied to analysing behaviour, and most studies started from scratch instead of building upon previous work. Training and validation sample sizes were generally small (mean±s.d. groups = 3.8±5.8) and in data collection and testing took place in relatively controlled environments. To advance our ability to automatically phenotype behaviour, future research should build upon existing knowledge and validate technology under commercial settings and publications should explicitly describe recording conditions in detail to allow studies to be reproduced.

/pdf/recording-behaviour-of-indoor-housed-farm-animals-5aeoe4tyw2.pdf

Recording behaviour of indoor-housed farm animals automatically using machine vision technology: A systematic review

In a surveillance camera environment, the detection of standing-pigs in real-time is an important issue towards the final goal of 24-h tracking of individual pigs. In this study, we focus on depth-based detection of standing-pigs with "moving noises", which appear every night in a commercial pig farm, but have not been reported yet. We first apply a spatiotemporal interpolation technique to remove the moving noises occurring in the depth images. Then, we detect the standing-pigs by utilizing the undefined depth values around them. Our experimental results show that this method is effective for detecting standing-pigs at night, in terms of both cost-effectiveness (using a low-cost Kinect depth sensor) and accuracy (i.e., 94.47%), even with severe moving noises occluding up to half of an input depth image. Furthermore, without any time-consuming technique, the proposed method can be executed in real-time.

https://www.mdpi.com/1424-8220/17/12/2757/pdf

Depth-Based Detection of Standing-Pigs in Moving Noise Environments

Automatic characterization of fracture surfaces of AISI 304LN stainless steel using image texture analysis

The problem of camera calibration is two-fold. On the one hand, the parameters are estimated from known correspondences between the captured image and the real world. On the other, these correspondences themselves—typically in the form of chessboard corners—need to be found. Many distinct approaches for this feature template extraction are available, often of large computational and/or implementational complexity. We exploit the generalized nature of deep learning networks to detect checkerboard corners: our proposed method is a convolutional neural network (CNN) trained on a large set of example chessboard images, which generalizes several existing solutions. The network is trained explicitly against noisy inputs, as well as inputs with large degrees of lens distortion. The trained network that we evaluate is as accurate as existing techniques while offering improved execution time and increased adaptability to specific situations with little effort. The proposed method is not only robust against the types of degradation present in the training set (lens distortions, and large amounts of sensor noise), but also to perspective deformations, e.g., resulting from multi-camera set-ups.

MATE: Machine Learning for Adaptive Calibration Template Detection.

Motion estimation is one of the most demanding and complex tools in block based video encoders. Variable block size motion estimation (ME) and multiple reference frames in H.264/AVC make motion estimation even more complex and time consuming. In HEVC, the complexity is even higher since there are more block sizes. This paper presents an analysis of various tools involved in some fast ME algorithms and proposes some improvements to them in order to achieve a novel fast hybrid algorithm. The proposed algorithm has been tested with HEVC reference software. Simulation results show that the algorithm achieves up to 44.7% decrease in ME complexity when compared to the fast ME algorithm (Test Zone Search or TZSearch) and up to 99% reduction in ME complexity compared to full search algorithm with negligible loss in PSNR and bitrate. Motion estimation is one of the most complex tools in block based video encoders.In HEVC, the motion estimation complexity is even higher since there are more block sizes.TZSearch algorithm is the fast motion estimation algorithm in HEVC reference software.This paper reduces the complexity of TZSearch ME algorithm upto 44.7%.The overall RD performance loss is negligible despite decrease in the complexity.

Complexity reduction methods for fast motion estimation in HEVC

The authors point out the limitations of SUSAN corner detector in detecting chessboard corner, then describe an improved SUSAN(Smallest Univalue Segment Assimilating Nucleus) detector algorithm for detecting chessboard corner on the basis of symmetrical geometry structure of USAN (Univalue Segment Assimilating Nucleus) area. And the algorithm has been applied to the chessboard images on real photos. The improved algorithm can quickly detect corner from real photos shot from every angle. The theory of detecting corner at sub-pixel level is Orthogonal Vector Theory, that is, vector from the corner to its adjacent area pixel point should be vertical to gray grads of the adjacent area pixel point. In order to get the coordinate of corner at sub-pixel level, we establish the neighboring area equation and solve it via iterative method, and propose to check its validity according to cross ratio invariability in perspective projection. Keywords-insert chessboard corner; USAN; sub-pixel; cross ratio invariability

A Fast and Accurate Algorithm for Chessboard Corner Detection

The precipitation behavior of V(C, N) in steels microalloyed with vanadium was researched using a thermal simulator during single-pass deformation at 800–750 °C The V(C, N) precipitates and its nucleation effect on ferrite were investigated by TEM and EDS The experimental results show that there are two remarkable heterogeneous nucleation effects of V(C, N) particles precipitated before γ → α phase change: primary reason is that high coherency between V(C, N) and ferrite promotes V(C, N) to become a nucleating center of intragranular ferrite; secondary reason is that the coarsening of V(C, N) causes locally solute-poor region in austenite, thus expedites the nucleation of intragranular ferrites further Furthermore, the relationship between the size and shape of V(C, N) was studied, and identification method was provided for distinguishing interphase precipitation and general precipitation to avoid erroneous judgment and misguide

Precipitation and hetero-nucleation effect of V ( C, N ) in V -microalloyed steel

Petrographical and geochemical characterization of the Upper Permian Longtan formation and Dalong Formation in the Lower Yangtze region, South China: Implications for provenance, paleoclimate, paleoenvironment and organic matter accumulation mechanisms

In order to explore the relation between steel's dimple size and mechanical properties,it is necessary to measure and classify the dimple size accurately. An improved dimple image quantitative analysis method based on combination of wavelet transform and mathematics morphology was proposed in this paper. It measured and analyzed the dimples by means of edge detection based on combination of B-spline wavelet transform and mathematics morphology, water-growing segmentation and measurement method etc. Application results showed that it had a good processing effect to solve the defects such as dimple edge fuzzy, dimple edge deletion and dimple holes etc, and its quantitative analysis was accurate, effective and convenient.

Research on quantitative analysis method of steel dimples based on wavelet transform

In order to obtain the prospect targets from the top view image of pig barns, the target extraction method of pigs based on adaptive elliptical block and wavelet edge detection is proposed in this paper. Firstly, two-dimensional OTSU global threshold segmentation and mathematic morphology processing were used to get the initial segmentation results. And then, adaptive elliptical block were taken around each pig respectively in the original image according to the initial segmentation results. Finally, the wavelet edge detection for secondary accurate extraction was used in each elliptical block and the complete contours of each pig were obtained. This method is suitable for a complex environment of the light changes in pig barns. This work provides a good foundation for the further research of pig's individual identification and behavior analysis. On the other hand, it also explores a new way to extract target contours of other similar images in complex environments and scenes.

Xincheng Li

Papers

A Fast and Accurate Algorithm for Chessboard Corner Detection

Precipitation and hetero-nucleation effect of V ( C, N ) in V -microalloyed steel

Petrographical and geochemical characterization of the Upper Permian Longtan formation and Dalong Formation in the Lower Yangtze region, South China: Implications for provenance, paleoclimate, paleoenvironment and organic matter accumulation mechanisms

Research on quantitative analysis method of steel dimples based on wavelet transform

Pig target extraction based on adaptive elliptic block and wavelet edge detection