Quantification of tick infestation in cattle using Infrared thermal image processing

Introduction

Awareness of the economic losses associated with tick infestation remains limited; however, ticks and tick-borne diseases such as Tick pyemia, Toxicosis, Paralysis, Babesiosis, Theileriosis and Anaplasmosis are globally recognized as significant constraints in livestock production (Ragulraj, S et al., 2023) [17] . Economic losses due to ticks primarily arise from reduced productivity, deterioration of hide quality and disease transmission (Gracia, 2003) [9] . Each tick causes an average live weight loss of 1.37 ± 0.25 grams in cattle (Jonsson, 2006; Holroyd and Dunster, 1978) [12,10] , and infestations of approximately 40 ticks per day can result in annual body weight losses of up to 20 kg per animal (Frisch et al., 2000) [8] . The risk for younger animals is higher if their growth gets compromised at an early age due to stress; their growth and production over their lifetime are affected. The collective economic loss from these effects on the global livestock industry is in the billions of dollars (Rodriguez-Vivas et al., 2014) [19] . Therefore, it is crucial to quantify the tick infestation level in dairy animals to make effective management decisions. Ticks are the second most common vectors of infectious diseases worldwide, after mosquitoes, in both livestock and humans. Globally, these ectoparasites affect the livestock industry as approximately one billion bovines are at risk of infestation (Rodriguez et al., 2011) [18] . Mortality incidence is less associated with tick infestation, but growth-and production-related losses are much more evident (Marcelino et al., 2012) [15] . Generally, tick infestation assessment is carried out manually, which is labor-intensive and time-consuming. With technological advancements, researchers can measure the variation in surface temperature of the tick and the host animal using IRT. For the first time, Cortivo et al. (2016) [5] reported that the surface temperature of the host animal body is>34 °C, and the temperature of the tick body is 33° C to 34° C, and the difference between tick and skin surface temperature would be around 1 °C. Therefore, there is a need to explore the potential Cortivo et al. (2016) [5] also compared visual count methods with an image analysis program for automated counting of ectoparasites (ticks and horn flies) captured via infrared thermography.

The visual count and thermography, coupled with an autocounting program, were highly correlated for horn flies (0.82) and distal-side ticks (0.84), but a correlation of 0.32 was observed for lateral-side ticks. The variation may be due to technical hitches in counting small ticks (diameter < 4 mm), hair length, skin coat colour, and environmental temperature. The surface temperature difference between the tick and the animal's body is used as a principle for detecting ticks. Environmental temperature can also influence tick body temperature. So, it may affect the accuracy of the tick count. Long, excessive hair makes it difficult to distinguish between ticks. A lack of temperature contrast reduces the actual tick count in the animal body (Cortivo et al., 2016) [5] . Sometimes the software may consider the cluster of ticks as a single tick. Capturing the angle of the IRT image can help improve the contrast between the cattle and the tick. Hair length interference can be mitigated by increasing the camera's sensitivity and resolution (Barbedo et al., 2017) [3] . Barbedo et al. (2017) [3] also proposed a new algorithm based on colour differences using mathematical morphology operations. They reported that IRT images can provide a rough estimate of tick infestation level, which can be better quantified by manual or proposed algorithms. Still, accurate quantification remains difficult due to various extrinsic and intrinsic factors. It is worth exploring the optimal body location for accurate tick quantification using the IRT image, coupled with automated analysis using a programme under tropical climatic conditions. Further, the concept of detecting and quantifying tick infestation using Infrared Thermography and auto-counting ticks in ImageJ software in crossbred animals under tropical climatic conditions, accounting for differences in climate, was validated.

Materials and Methods

The study was conducted at the Livestock Research Centre (LRC) of the ICAR-National Dairy Research Institute (ICAR-NDRI), Karnal, Haryana, at an elevation (attitude) of 250 meters above mean sea level, latitude 29°43" North and longitude 77°20" East. This is a tropical climate region, and the annual minimum and maximum temperatures range from 4 to 45 °C. During the hot-dry season (April to June), the relative humidity varies from 41 to 85% and the vapour pressure from 7.0 to 25 mmHg. Ethical approval (number 45-IAEC-19-3) was sought from the Institutional Animal Ethics Committee (IAEC) at ICAR-National Dairy Research Institute for the protocols followed and the animals' management in the experiment. A total of 112 crossbred animals aged 12-18 months, infested with ticks, were randomly selected. Then, ticks were counted manually, and values from visual and thermal images, coupled with automated analysis in ImageJ, were evaluated to determine the best region for capturing Infrared thermography images to detect ticks with higher accuracy. Surface temperature data from the animal and the tick were used to determine where the most significant temperature difference occurred between them.

Statistical analysis

Statistical analysis was performed using SPSS 20.0 (IBM Corporation, Armonk, New York, USA). Data on tick and skin surface temperature were analysed by using an independent t-test. Then, the temperature difference between the tick and skin surface was analysed using one-way analysis of variance (ANOVA), followed by the post hoc Tukey test to compare the temperature differences between regions. The data on manual tick counts and ImageJ analyses of tick counts across various regions were analysed using an independent t-test.

Results and Discussion Predilection site of ticks in crossbred animals

The maximum population of ticks in crossbred animals was recorded in the lateral side, followed by the fore limb, inner thigh, brisket, perineum and abdomen region (Table 1). The variable results were documented regarding the predilection site of the ticks (Katuri et [13,6,2] as the predilection site of tick infestation and site of tick attachment to the animal body varies with tick species (Abiso et al., 2019; Huruma et al., 2015; Fantahun and Abdul, 2012) [1,11,7] , crowdedness of the host species, season, time and accessibility of grooming kit (Solomon and Kassa, 1998) [20] . Katuri et al. (2013) [13] reported maximum tick infestation under the scrotum, followed by the ano-vulva, perineum, dewlap and brisket region. Eyo et al. (2014) [6] reported that the inner thigh was most infested, followed by the armpit, udder, penis, scrotum and vulva in indigenous cattle. Akthar et al. (2019) [2] reported the maximum number of tick infestations in the inguinal region (310), followed by abdomen (300), head (200), tail (100) and thorax (40) in cattle. Apart from the lateral side of the body, the forelimb and inner thigh regions have a greater number of ticks because, in general, ticks prefer sites with thin skin, less hair, and a rich blood supply for easy penetration. Ticks prefer moist, warm sites as well as sites protected from environmental temperatures and from birds (Muchenje et al., 2008) [16] . The odour of various predilection sites, especially the armpit and inner thigh, may also be an additional factor for higher tick infestation in the region (Wanzala et al., 2004) [21] .

The tick and skin surface temperature (°C) in various regions of the animal body

Based on the findings involving 112 animals, significantly (p < 0.05) higher skin surface temperatures (°C) in the dewlap (36.77 ± 0.16 vs. 36.06 ± 0.15), perineum (36.06 ± 0.13 vs. 34.95 ± 0.11), and inner thigh (35.75 ± 0.12 vs. 34.27 ± 0.13) regions were observed compared to the tick surface. In contrast, skin surface temperature was higher on the lateral side (35.93 ± 0.12 vs. 35.62 ± 0.12), but this difference was not significant (Figure 1). Cortivo et al. (2016) [5] also reported that tick surface temperature was lower compared to skin surface temperature. The scientific reason behind it has already been discussed in the present paper.

Difference between the tick and skin surface temperature (°C) in various regions of the animal body:

Based on the findings involving 112 animals, the maximum temperature (°C) difference recorded between the tick and skin surface temperature (1.48 ± 0.04) was in the inner thigh region, followed by perineum (1.11 ± 0.06), dewlap (0.71 ± 0.03) and lateral (0.30 ± 0.02) aspect of body (p < 0.01) (Figure 2).In a similar line, Cortivo et al. (2016) [5] reported that the temperature difference between the tick and the skin surface was approximately 1 °C. The scientific reason behind it has already been discussed in the present paper.

Tick count by Image J and manual of various regions of the animal body

No significant difference was recorded in tick count between Manual count and Image J analysis in the inner thigh region (42.89 ± 1.59 vs. 43.83 ± 1.61). Still, it was found to be significantly (p < 0.01) different in dewlap (21.01 ± 1.59 vs. 15.04 ± 1.25), perineum (19.9 ± 1.32 vs. 15.58 ± 1.34) and lateral side (7.15 ± 0.94 vs. 4.47 ± 0.81) of the body (Figures 3 &4). The correlation between visual count and IRT image, coupled with ImageJ analysis of the inner thigh region (0.85), dewlap (0.68), perineum (0.75), and lateral side (0.59), was estimated. In a similar line, Cortivo et al. (2016) [5] reported the highest correlation (0.82) between Image J and manual tick count in the distal part of the animal body. This might be due to lower hair density and a greater surface-temperature difference between the tick and the animal's body in the inner thigh region. A significant difference (p < 0.01) of Image J count and manual tick count was found in dewlap, perineal region and lateral side of the body, further supported by the findings of Cortivo et al. (2016) [5] who reported lowest correlation (0.32) between Image J counting and Manual tick count in the lateral side of the animal body and Barbedo et al. (2017) [3] reported poor correlation (0.285) between visual count and developed algorithm in neck region. This might be due to the presence of a lengthy and dense hair coat, as well as direct environmental influences, causing a minor temperature difference between the tick and the skin coat; therefore, the lack of temperature contrast reduced the actual tick count (Cortivo et al., 2016; Barbedo et al., 2017) [5,3] . The surface temperature difference between the tick and the animal's body was used as a principle for detecting ticks. Tick temperature may be influenced by environmental temperature, potentially affecting the accuracy of the tick count. A wide variation in tick surface temperature was also evident in the present study, but it was always lower than the animal's skin surface temperature. It had also been reported that an error in counting of ticks might be associated with the size of the ticks (less than 4 mm in diameter), skin and hair colour, as well as length of the hair (Cortivo et al., 2016) [5] . Sometimes the software may consider a cluster of ticks as a single tick. Capturing the angle of the IRT image can help improve the contrast between the cattle and the tick. Hair length interference can be mitigated by increasing the camera's sensitivity and resolution (Barbedo et al., 2017) [3] . The quality of the image in capturing the thermal signature of ticks may also influence tick quantification, and image quality can sometimes be compromised by the inevitable movement of animals even after proper restraint (Barbedo et al., 2017) [3] . Barbedo et al. (2017) [3] found a low correlation between visual counting of thermographic images and manual tick counts, corroborating our study. Besides that, no part of the body can be captured by IR images. The present study demonstrated that tick count using IRT image capture, coupled with automated analysis in ImageJ, especially in the inner thigh region, holds greater promise for identifying and quantifying ticks. Another benefit of assessing ticks using an IRT camera is that ticks may fall due to licking or scratching by the animal, as well as at different stages of the tick's life cycle; they may also be assessed, as demonstrated by Cortivo et al. (2016) [5] . Despite its shortcomings, IRT can serve as a valuable tool for detecting ticks across various body regions. The noninvasive nature, lower labour intensity, and easier counting are the key points of interest. The visual quantification of ticks has its own limitations; ticks smaller than 4 mm in diameter, skin colour of the animal, length and density of hair, and the skill of the observer may influence the actual tick infestation rate. Similar to our observation, IRT image capture and programme-based auto-analysis quantify tick and horn flies (Cortivo et al., 2016) [5] , and the filming method for horn fly quantification has been reported to be superior to visual count (Lima et al., 2002) [14] . Image-based quantification of ticks has several advantages, as described by Cortivo et al. (2016) [5] . Besides, ticks can exhibit contrast due to temperature variation. IRT can also capture inflammation caused by the tick bite, as the tick may be detached from the animal's body before image capture due to licking, scratching, the tick's lifecycle stage, or the use of grooming brushes on commercial farms. The study's limitation is that the threshold for quantifying tick infestation in each image is independent, and tick species identification is another challenge.

Conclusion

The present study has identified that thermal imaging, coupled with automated analysis using ImageJ, is an effective method for accurate quantification of ticks in the inner thigh region of crossbred animals.

Acknowledgement

AcknowledgementsICAR-National Dairy Research Institute provided research funding under the Institute Research Project (B-54).

References

1. Abiso M, Yohannes B, Alemu B. Ixodidae ticks of bovine: prevalence and major species identification in Soddo Zuria Districts of Wolaita Zone, Ethiopia. Approaches Poult Dairy Vet Sci. 2019;6: APDV000632.
2. Akthar Z, Anwar Z, Adnan M, Imtiaz N, Rashid HU, Gohar F. Collection and prevalence of ticks in cattle and buffaloes from free-range management systems of Islamabad. J Basic Appl Zool. 2019;80:12.
3. Barbedo JGA, Gomes CCG, Cardoso FF, Domingues R, Ramos JV, McManus CM. The use of infrared images to detect ticks in cattle and proposal of an algorithm for quantifying the infestation. Vet Parasitol. 2017;235:106-112.
4. Castro E, Gil A, Solari MA, Farias NA. Validation of a subjective counting method for a horn fly (Haematobia irritans irritans) population in a cattle herd. Vet Parasitol. 2005;133:363-367.
5. Cortivo PD, Dias E, Barcellos JOJ, Peripolli V, Costa JBG Jr, Dallago BSL, et al. Use of thermographic images to detect external parasite load in cattle. Comput Electron Agric. 2016;127:413-417.
6. Eyo JE, Ekeh FN, Ivoke N, Atama CI, Onah IE, Ezenwaji NE, et al. Survey of tick infestation of cattle at four selected grazing sites in the tropics. Glob Vet. 2014;12:479-486.
7. Fantahun B, Abdu M. Survey on the distribution of tick species in and around Assosa Town, Ethiopia. Vet Sci Res J. 2012;5:32-41.
8. Frisch JE, O'Neill CJ, Kelly MJ. Using genetics to control cattle parasites-the Rockhampton experience. Int J Parasitol. 2000;30:253-264.
9. Gracia Z. Integrated control of Boophilus microplus in cattle. In: Proceedings of the 11 th International Congress of the International Society for Animal Hygiene; 2003; Mexico City, Mexico.
10. Holroyd RG, Dunster PJ. The effect of the cattle tick on growth rates and reproductive rates of Bos indicus cross heifers in North Queensland. Proc Aust Soc Anim Prod. 1978;12:277.
11. Huruma G, Abdurhaman M, Gebre S, Deresa B. Identification of bovine tick species and their prevalence in and around Sebeta Town, Ethiopia. J Parasitol Vector Biol. 2015;7:1-8.
12. Jonsson NN. The productivity effects of cattle tick (Boophilus microplus) infestation on cattle, with particular reference to Bos indicus cattle and their crosses. Vet Parasitol. 2006;137:1-10.
13. Katuri RN, Das G, Chalhotra SK, Singh AK. Analysis of site of predilection for the ticks of family Ixodidae. Int J Food Agric Environ. 2013;3:42-46.
14. Lima LGF, Prado AP, Perri SHV. Comparison of two methods (visual estimates and filming) for counts of horn flies (Haematobia irritans irritans). Vet Parasitol. 2002;103(3):227-235.
15. Marcelino I, De Almeida AM, Ventosa M, Pruneau L, Meyer DF, Martinez D, et al. Tick-borne diseases in cattle: applications of proteomics to develop new generation vaccines. J Proteomics. 2012;75(14):42324250.
16. Muchenje V, Dzama K, Chimonyo M, Raats JG, Strydom PE. Tick susceptibility and its effects on growth performance and carcass characteristics of Nguni, Bonsmara and Angus steers raised on natural pasture. Animal. 2008;2:298-304.
17. Ragulraj S, Bhakat M, Fernandes A, Nandhini PB, Sahu C. Haematological investigation and identification of tick infestation in crossbred cattle. Pharm Innov J. 2023;12(8):811-813.
18. Rodríguez R, Ojeda M, Pérez L, Rosado J. Epidemiology and control of Rhipicephalus (Boophilus) microplus in Mexico. In: Quiroz H, Figueroa J, Ibarra F, López M, editors. Epidemiología de enfermedades parasitarias en animales domésticos. Mexico: UNAM; 2011. p. 477-504.
19. Rodríguez-Vivas R, Rosado-Aguilar J, Ojeda-Chi M, Pérez-Cogollo L, Trinidad-Martínez I, Bolio-González M. Integrated control of ticks in bovine production. Ecosist Recur Agropecu. 2014;1(3):295-308.
20. Solomon G, Kaaya GP. Development, reproductive capacity and survival of Amblyomma variegatum and Boophilus decoloratus in relation to host resistance and climatic factors under field conditions. Vet Parasitol. 1998;75:241-253.
21. Wanzala W, Sika NFK, Gule S, Hassanali A. Attractive and repellent host odours guide ticks to their respective feeding sites. Chemoecology. 2004;14:229-232.