Body Dimension Measurements of Qinchuan Cattle with Transfer Learning from LiDAR Sensing

Abstract

For the time-consuming and stressful body measuring task of Qinchuan cattle and farmers, the demand for the automatic measurement of body dimensions has become more and more urgent. It is necessary to explore automatic measurements with deep learning to improve breeding efficiency and promote the development of industry. In this paper, a novel approach to measuring the body dimensions of live Qinchuan cattle with on transfer learning is proposed. Deep learning of the Kd-network was trained with classical three-dimensional (3D) point cloud datasets (PCD) of the ShapeNet datasets. After a series of processes of PCD sensed by the light detection and ranging (LiDAR) sensor, the cattle silhouettes could be extracted, which after augmentation could be applied as an input layer to the Kd-network. With the output of a convolutional layer of the trained deep model, the output layer of the deep model could be applied to pre-train the full connection network. The TrAdaBoost algorithm was employed to transfer the pre-trained convolutional layer and full connection of the deep model. To classify and recognize the PCD of the cattle silhouette, the average accuracy rate after training with transfer learning could reach up to 93.6%. On the basis of silhouette extraction, the candidate region of the feature surface shape could be extracted with mean curvature and Gaussian curvature. After the computation of the FPFH (fast point feature histogram) of the surface shape, the center of the feature surface could be recognized and the body dimensions of the cattle could finally be calculated. The experimental results showed that the comprehensive error of body dimensions was close to 2%, which could provide a feasible approach to the non-contact observations of the bodies of large physique livestock without any human intervention.

Body measurement of riding horses with a versatile tablet-type 3D scanning device

“…Measuring horses is more difficult than measuring cows because the more an object moves, the more difficult it is to measure. The accuracy of the present study was almost equivalent to that of a study by Cozler et al [4] and slightly lower than those of studies by Huang et al [9,10]. Even a horse that moves frequently can be made to stay still for a short time.…”

“…In a study measuring 30 Holstein dairy cows using a 5-camera system, the correlation coefficients between manual and 3D measurements were 0.89 for ChD, 0.82 for HWi, 0.78 for GiC, 0.76 for backside width, 0.63 for ischial width, and 0.62 for WCr [4]. Huang et al [9,10] used a non-contact body dimension measurement approach for Qinchuan cattle using a LiDAR system and found that the final deviations were close to 2 mm and within approximately 2%. Measuring horses is more difficult than measuring cows because the more an object moves, the more difficult it is to measure.…”

“…In recent years, there has been advancement in the technology for acquiring stereoscopic images with a 3D scanner, even for large animals such as cattle [4,5,9,10] and horses [16]. The key feature of these studies is that the experimenter can take measurements without touching the animal.…”

“…Measurements using completely non-contact conditions are of great significance in carrying out research, particularly that carried out on delicate and expensive racehorses. However, advanced data processing technologies, such as filtering and deep learning, were required to analyze the rear shape of cows [5], the conformation of Qinchuan cattle [9,10] and Andalusian horses [16]. It is currently difficult to say that a horse's conformation can be easily measured in their normal environment.…”

ASAS-NANP SYMPOSIUM: Applications of machine learning for livestock body weight prediction from digital images

“…The volumetric measurements are rough approximations given the limited camera views of either the top or the side of an animal. The shape descriptors reported in the literature include generic ones such as Fourier descriptors ( Fernandes et al, 2019 ) and fast point feature histograms ( Huang et al, 2019 ).…”