Computers and Electronics in Agriculture
Volume 210,
July 2023
, 107908
Author links open overlay panel, , , , , , , , ,
Abstract
Blood vitamin A levels during fattening have a crucial influence on beef marbling. As an alternative of conventional blood assays, eye imaging method has been raised as a potential non-invasive way to monitor vitamin A levels in cattle. However, current image capture systems merely acquire cattle eye surface images, which cannot adequately reflect the chromatic information of eye retina that indicates vitamin A level changes. Based on previous research, we propose a novel, automatic double imaging system that combined two illuminating system as well as the installation of polarising filters to eliminate eye surface reflection and consequently, obtain clear fundus images. This greatly facilitated subsequent image processing and feature extraction for vitamin A prediction. The proposed imaging system demonstrated a stronger correlation between vitamin A levels and chromatic features in fundus images than previous systems. An improved classification (accuracy 82%) of vitamin A deficiency level than previous study (56.6%) strongly supports this system’s advantage. Being non-invasive cost-effective and rapid, this double imaging system is expected to play an important role in precision livestock farming as well as clinical diagnosis in veterinarian.
Introduction
Marbling of meat is the appearance of white flecks or streaks of fatty tissue that spread within the muscle, located between individual muscle fibres (Harper and Pethick, 2001). Positively correlated with juiciness, tenderness and palatability (Killinger, 2004, Wheeler et al., 1994), it is also one of the most important features determining beef quality, also referred to as the beef marbling standard. It is subject to not only genetic factors (Peng et al., 2021, Yamada et al., 2020), but also nutritional factors (Motoyama et al., 2016, Pethick et al., 2004). With the discovery of significantly increased marbling in non-vitamin A supplemented group of young wagyu cattle (Oka et al., 1998), dietary restriction of vitamin A has become common practice for farmers to manipulate beef marbling in Japan and abroad. (Bryant et al., 2010, Gorocica-Buenfil et al., 2007, Peng et al., 2019, Siebert et al., 2006). An example of the desired vitamin A levels in cattle for good marbling is shown in Fig. 1. Currently, cattle producers can only monitor and control vitamin A levels through regular blood assays. Although effective, the process of collecting blood from cattle is laborious and potentially stressful. Moreover, blood assays are expensive, time consuming, and therefore limit the usage of these assays by cattle farmers.
Vitamin A and its derivatives are known to play key roles in eye vision (Morshedian et al., 2019). It cannot be synthesised by the animals, rather it is obtained from food source carotenoids, such as β-carotene (Wyss and Lintig, 2008). A continued supplementation of vitamin A is essential in the retinoid cycle of eye vision (Kaylor et al., 2017). From a physiological perspective, vitamin A level affects pupillary light reflex features (Purves et al., 2008), as well as the amount of chromophore retinal that turns to photopigment (Palczewski and Kiser, 2020), which macroscopically is reflected in the colour of retina (Aranda and Schmidt, 2021). For this reason, dietary regulation of vitamin A during fattening period impacts the chromatic features of cattle eyes. Below a critical level of vitamin A concentration, early clinical symptoms will be observed (O'Donoghue, 1955). Based on the above considerations, researchers have turned to ophthalmological and photographic methods as a mean to monitor vitamin A levels.
Matsuda et al. (1999) investigated pupil shrinking speed and its relation with blood vitamin A based on camera imaging. They found a delay in constriction time for cattle with lower vitamin A levels. Our group furthered this research by using computer vision for vitamin A prediction. Takahashi et al. (2011) developed an ultraviolet camera to investigate eye tapetum. They found a negative correlation between serum vitamin A level and tapetum reflection observed in ultraviolet eye images. This initial research was conducted on dissected eyes after slaughter, rather than monitoring live cattle. To explore the possibility of real time monitoring, Takao et al., 2011, Han et al., 2013 developed an imaging system composed of 2 CCD multi-spectral cameras to capture colour and near-infrared images of cattle eyes in situ. Although the near infrared images were not able to provide much information, they still found a negative correlation between vitamin A concentration and red component ratio (in RGB components) of cattle eyes in colour images. In addition, Han et al. (2014) investigated how blood vitamin A level influenced pupillary light reflex features. However, in their images, there was reflection from white LED ring (Fig. 2) on the eye surface and, despite masking or use of other image processing techniques, this still caused intensity gradient problems, as well as unclear boundaries with the region of interest (ROI). Furthermore, their imaging process required manual operation and it would be laborious for farmers and stressful to cattle when repeated many times. Then Zhou et al. (2017) developed an automatic system that acquired colour photographs of cattle eyes. This has greatly reduced the manual workload, but it still captured eye surface images that met with the same problems from the LED light source, as well as the resultant issues with image processing.
Although previous research has set the foundation and enabled a possible way to predict vitamin A level by eye photography, there are still several issues that need to be addressed. Firstly, changes in blood vitamin A level are reflected in chromatic variances in the retina. But these eye surface images (Fig. 2) observed indirectly the retina colour from “outside of the eye” and therefore involved systematic noise and errors, including reflected light from light source. What we pursue is a direct observation of the retina from an “inside the eye” perspective, or a fundus image, to be exact. Secondly, in all eye surface images, the LED light results in inevitable surface reflection, which interfered with the chromatic information extracted for vitamin A prediction. Furthermore, the dark bluish pupil area sometimes also hurdled the ROI extraction during image processing. Therefore, a new image acquisition setup was needed to eliminate interference from intense LED reflection of the eye surface and to capture the fundus of the eye.
In this work, we propose a novel automatic double imaging system and mainly focused on the construction and evaluation of this camera system. The usage of a combined illuminating system of ring LED and centre LED as well as the installation of perpendicularly oriented PL filters is the key to suppressing eye surface reflection while enabling the fundus reflection to be observed. In the subsequent investigation, accurate prediction model of vitamin A is to be established through combining multiple factors such as pupil shrinking speed, constriction amplitude, latency time as well as various colour features from fundus and eye surface images captured by this system and the result will be reported in the next paper.
Section snippets
Experiment and measurement
Forty clinically healthy Japanese Black Cattle (Tajima strain) were fed grain-based diet low in β-carotene (under 0.04mg/kg) in Tajima Agricultural High School, Hyogo Prefecture, Japan. Four steers of them, with similar 10.2±0.88month-age, were chosen for eye photography experiment that was conducted over 14months during which, blood vitamin A level changes were tracked. Although only four cattle were used for the experiment, their eye images were captured continuously throughout the whole
A comparison of eye surface and fundus images
With new double imaging system, we captured cattle eye surface and fundus images, as shown in Fig. 5. Since the scattering effect of other ocular tissues, such as cornea and lens, is negligible in comparison with that of the retina (Sardar et al., 2009), and these effects are present in both eye surface and fundus images, optical properties of ocular tissues, as well as their influence on distribution and propagation of light were not taken into account. What is crucial is the specular
Conclusion
In this work, we proposed novel automatic eye surface-fundus double imaging system that combined two illuminating system as well as the installation of PL filters. Clear distinct fundus images facilitate the image processing and feature extraction for vitamin A assessment. PCA analysis and classification of VAD outperformed previous study in the overall classification accuracy (from 56.6% to 82%) and demonstrated this system’s advantage.
Funding
This work was supported by JSPS KAKENHI [Grant Number 17H01500, 20K15628, 20H00439, 23H00350] from Japan Society for Promotion of Science.
CRediT authorship contribution statement
Nanding Li: Conceptualization, Methodology, Software, Formal analysis, Writing – original draft, Writing – review & editing. Otieno Samuel Ouma: Methodology, Formal analysis, Data curation, Software. Dimas Firmanda Al Riza: Methodology, Formal analysis. Mizuki Shibasaki: Methodology, Visualization. Wulandari: Methodology, Validation. Moriyuki Fukushima: Methodology. Tateshi Fujiura: Validation, Writing – review & editing. Yuichi Ogawa: Validation. Naoshi Kondo: Conceptualization, Resources,
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
Authors want to thank Takehiko Ohmae and Norio Nishiki for assistance in cattle management. We are grateful to Professor Garry John Piller (Graduate School of Agriculture, Kyoto University, Japan) for English proof reading.
References (31)
- S. Han et al.
Estimation of serum vitamin a level by color change of pupil in japanese black cattle
Eng. Agric. Environ. Food
(2013)
- S. Han et al.
Feasibility of pupillary light reflex analysis to identify vitamin A deficiency in Japanese black cattle
Comput. Electron. Agric.
(2014)
- S. Han et al.
Classification of vitamin A deficiency levels by ocular changes in Japanese black cattle
Biosyst. Eng.
(2018)
- Z.A. Kruk et al.
Effects of vitamin A on growth performance and carcass quality in steers
Livest. Sci.
(2008)
- T.D. Lamb et al.
Dark adaptation and the retinoid cycle of vision
Prog. Retin. Eye Res.
(2004)
- A. Morshedian et al.
Light-Driven Regeneration of Cone Visual Pigments through a Mechanism Involving RGR Opsin in Müller Glial Cells
Neuron
(2019)
- M. Motoyama et al.
Wagyu and the factors contributing to its beef quality: A Japanese industry overview
Meat Sci.
(2016)
- A. Oka et al.
Influence of vitamin A on the quality of beef from the Tajima strain of Japanese black cattle
Meat Sci.
(1998)
- Y. Qiao et al.
Intelligent perception for cattle monitoring: A review for cattle identification, body condition score evaluation, and weight estimation
Comput. Electron. Agric.
(2021)
- N. Takahashi et al.
Effects of serum vitamin a level on tapetum in Japanese black cattle eye
Eng. Agric. Environ. Food
(2011)
Automated pupil image acquisition to estimate serum Vitamin A levels in cattle using pupil color analysis
1st Asian-Australasian Conf Precis. Pastures Livest. Farming
(2017)
Diversity of intrinsically photosensitive retinal ganglion cells: circuits and functions
Cell. Mol. Life Sci.
(2021)
Effect of dietary supplemental vitamin A concentration on performance, carcass merit, serum metabolites, and lipogenic enzyme activity in yearling beef steers
J. Anim. Sci.
(2010)
Effect of low vitamin A diets with high-moisture or dry corn on marbling and adipose tissue fatty acid composition of beef steers
J. Anim. Sci.
(2007)
Cited by (0)
Recommended articles (6)
Research article
Development of improved flow analysis prototype method for measuring and understanding agricultural non-point source (NPS) nitrogen (N) and phosphorus (P) pollution
Computers and Electronics in Agriculture, Volume 210, 2023, Article 107894
The loss of nitrogen (N) and phosphorous (P) through continuous surface runoff by rainfall after fertilizing and leaching into groundwater has become a critical issue that causes eutrophication of rivers, lakes, and even drinking water, resulting in agricultural non-point source (NPS) pollution of up to 50% of pollution load. Traditionally, N and P were measured using the laboratory's flow injection analysis (FIA) method before evaluating the pollution load, which could meet urgent requirements in the fast response, prevention, and control of agricultural NPS pollution. Herein, we propose an improved flow analysis (perturbation dispersion) method by integrating chemical reaction chambers to measure nitrogen and phosphorus in water automatically and continuously. The as-fabricated system was validated with real samples from Qianjiang, Hubei Province. The experimental results drew a correlation coefficient of 0.91449 for ammonium nitrogen between the existing lab-based flow analysis method and the innovative (improved) flow analysis method. The correlation coefficients of 0.99828, 0.85129, and 0.99206 for nitrate nitrogen, ammonium nitrogen, and phosphorus were calculated from the measurement results. The detection limits of 0.55, 0.20, and 0.025mg/L with relative standard deviations (RSD) of 0.8%, 5.7%, and 0.9% were observed for NH4+–N, NO3––N, and PO43− measurements, respectively. This work facilitates the online and continuous control of the NPS pollution load and timely expands the flow injection analysis (FIA) techniques for understanding the pollutants in water and farmland environments.
Research article
Analysis of jujube movement characteristics under positive and negative pressure airflow based on CFD-DEM
Computers and Electronics in Agriculture, Volume 210, 2023, Article 107902
Current study was conducted to explore the jujube kinematic characteristics in the pneumatic pickup device's operation under positive and negative pressure airflow to optimize its working parameters. This study designed a CFD-DEM coupling simulation system of the device in operation. Here, we investigated the kinematics of internal airflow and the jujube equivalent under its action. Following that, the pickup rate and loss rate of the equipment were used as response indicators, and the weight analysis method was used to calculate the forward speed and negative and positive pressure airflow velocity. The optimal parameter combination were 0.26m/s, 35m/s, and 20m/s. At the same time, the pickup rate and loss rate of the model predicted were 99.20% and 0.75%, respectively. Results from the field experiment showed that the pickup rate and loss rate were 98.43% and 0.80%, which error with the predicted values were 0.78% and 6.67%. The research is not only of great significance for research on pneumatic landing jujube pickers but has also great reference value to the construction and methods of the CFD-DEM coupling simulation systems.
Research article
Combined fluorescence-transmittance imaging system for geographical authentication of patchouli oil
Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, Volume 218, 2019, pp. 155-160
Recently, demand for authentication technology is growing rapidly in an attempt to overcome counterfeiting of high-value agricultural products, such as patchouli oil. Fingerprinting methods based on spectroscopy are one such technology being used for authentication. However, the spectral datasets obtained are multivariate in nature; containing thousands of data points for a single sample, making data acquisition and processing time-consuming. Therefore, reduction and simplification in the number of variables used required is needed to provide a more rapid and applicable method.
Color cameras, which can capture image in the visible region light, could be such an alternative spectral data acquisition approach. In this research, a simplified spectroscopy method was developed for origin authentication of patchouli oil. The system consists of front ultraviolet light induced (365 nm) fluorescence and a white LED-based backlighting imaging system that consecutively captures the fluorescence and transmittance characteristics of the oil in the visible region. From the captured images, features were extracted and analyzed using Principle Component Analysis (PCA) to identify important image features for discrimination of origin. From the samples measured, the samples clustered around three islands of origin in the PCA space. A classification model based on fluorescence and transmittance image features (color values) could discriminate origin classes with a total accuracy of 88.46%. A lower accuracy was found for the Java class due to low sample numbers. This result demonstrates that the proposed system has the potential to be a rapid authentication tool for determining the geographical origin of patchouli oils.
Research article
Where am I heading? A robust approach for orientation estimation of autonomous agricultural robots
Computers and Electronics in Agriculture, Volume 210, 2023, Article 107888
Reliable knowledge of the vehicle heading plays a significant role in the autonomous navigation of agricultural Unmanned Ground Vehicles (UGVs), especially in the context of unstructured outdoor environments such as rural and forestry scenarios. However, achieving this information with an acceptable degree of confidence is a non-trivial task and still an open field of research. Expensive solutions are available on the market, but they often discourage most farmers due to the large investments needed for the startup. This paper introduces a novel algorithmic solution for reliable evaluation of the absolute vehicle heading, grounded on adaptive Kalman filtering with input evaluation via linear regression analysis. The proposed approach provides a functional and affordable solution to the heading estimation problem that can be used in real-world applications. The system is validated through an extensive experimental campaign using an all-terrain tracked rover operating in agricultural settings, showing good accuracy compared to other approaches, such as a dual GPS method found in the literature.
Research article
Classification of raw Ethiopian honeys using front face fluorescence spectra with multivariate analysis
Food Control, Volume 84, 2018, pp. 83-88
Front face fluorescence measurements were carried out to classify raw honeys as such based on their floral origins. The excitation-emission matrix patterns of mixed flower, pseudoacacia, arabica, fials-indica, and amygdalina raw honeys along with fake honey sample from the market were examined by recording emission wavelength from 250 to 600nm with excitation wavelength in the range of 200–550nm. The spectra of fake honey samples demonstrated low intensity and did not fit within any one of the classified raw honeys. The multivariate analyses of the spectra were performed using principal component analysis and soft independent modeling of class analogy (SIMCA). The SIMCA model showed that the adulterate honey samples were detected with 100% sensitivity and specificity.
Research article
Prediction of seed distribution in rectangular vibrating tray using grey model and artificial neuralnetwork
Biosystems Engineering, Volume 175, 2018, pp. 194-205
To maintain good continuous working performance in a vacuum plate seeder, it is important to monitor the distribution of seeds in real time and automatically adjust vibration parameters accordingly. Seed motion in a rectangular vibrating tray with vibration varying with time and interference by direction angle was simulated using the discrete element method (DEM). A plane model P was used to describe the variation of seed layer thickness. Four square areas on the bottom of the tray were divided symmetrically near the four corners to measure seed layer thickness, and a monitoring plane model Pm was established. DEM simulation results showed that the models Pm and P had the similar change rules, although there were some differences in fitting parameters. There was obvious time delay in the change of Pm compared with P. Therefore, a grey system model (GM) was adopted to predict the change of Pm, and two back-propagation (BP) neural networks which take GM prediction results as input parameters were developed respectively. Then, according to the BP neural networks outputs, a prediction plane model Pp was proposed to predict the seed distribution. Experiments were carried out on a test-rig to validate these predictions. The seed distribution plane P was measured manually, the monitoring plane Pm was established using seed layer thickness measurement results and the prediction plane Pp was established using the GM and BP neural networks. The results indicated that the proposed method had good precision and stability and provides the basis for the design of an automatic control system for the vibrating tray to promote a uniform seed distribution.
© 2023 Elsevier B.V. All rights reserved.