Use of thermal drone in detection and assessment of larval mass temperature in decomposed rabbit carcasses

Manual ground searches and cadaver dogs are traditional methods for locating remains, but they can be time‐ and resource‐intensive, resulting in the decomposition of bodies and delay in victim identification. Therefore, thermal imaging has been proposed as a potentially useful tool for detecting remains based on their temperature. This study investigated the potential of a novel search technique of thermal drones to detect surface remains through the detection of maggot mass temperatures. Two trials were carried out at Selangor, Malaysia, each utilizing 12 healthy male Oryctolagus cuniculus European white rabbits and DJI Matrice 300 RTK drone China, equipped with a thermal camera; Zenmuse H20T to record the thermal imaging footage of the carcasses at various heights (15, 30, 60–100 m) for 14 days for each trial. Our results demonstrated that the larval masses and corresponding heat emissions were at their largest during the active decay stage; therefore, all the carcasses were observable in thermal images on day 5 and remained until day 7. Statistical analyses showed that (1) no statistically significant differences in thermal images between clothed and unclothed subjects (p > 0.05); (2) 15 m above ground level was proven to be the optimal height, as it showed the greatest contrast between the carcass heat signature and the background (p < 0.005). Our data suggested the potential window of detection of thermal signatures was detectable up to 7 days post‐deposition. This could be an important guideline for the search and recovery teams for operational implementation in this tropical region.


| INTRODUC TI ON
In forensic entomology, analysis of insects can provide evidence to assist legal investigations by establishing some circumstances of death [1].The insect evidence can be used to determine the cause, place, and time of death through careful examination of the larval gut content, species, and developmental stage, respectively [2].The presence of insects as first colonizers within decomposed corpses has been used widely to estimate the minimum PMI (mPMI) by analyzing insect developmental stages or by estimating insect succession patterns on the corpse [3,4].
In tropical environments, shortly after death occurs, the dead body will entice various species of insects and arthropods as food sources, and thousands of fly larvae will soon colonize the ephemeral resource.During the active decay stage, large aggregations of fly larvae or maggot masses will be formed as they feed on decomposing carcasses [5,6].Hence, a significant amount of heat will be produced by maggot masses, and this phenomenon is also known as larval mass effect [7].The heat produced is commonly higher than ambient temperature by more than 3°C as a result of microbial metabolism or from the continuous frenetic movement and high metabolism rate of larval mass due to larval digestion processes [8,9].
Recently, this phenomenon has piqued the interest of forensic entomologists to utilize the heat generated from larval mass to locate decomposing remains.The earliest findings reported the detection of larval heat mass in pig carcasses using a handheld thermal imaging camera over a short distance (35 m) [10].Thermal imaging is a process where a thermal camera captures and creates an image of an object by using infrared radiation emitted from the object based on its temperature [11].Thus, thermal imaging has been used to detect and differentiate the corpse's temperature from the surrounding area [11].Furthermore, it has been reported that thermal imaging was used in the detection of clandestine graves in various environments, for example, woodland [12] and arid [13].
In addition, the time of death could be determined using the corpse's temperature, especially in the fire investigation [14].
According to [14], thermography can also be used to locate injured survivors or victim remains.However, only limited research has been reported on thermal imaging in searching for decomposed bodies associated with the larval mass effect.One of the researchers demonstrated the successful detection of heat signatures from two decomposing pig carcasses using an aircraft-mounted thermal imaging (AMTI) [15].Although both animal carcasses were successfully detected by thermal imaging for 21 days at the height of 457 m, the authors recommended further studies using larval masses thermal imaging in various environmental conditions such as geographic regions, habitats, and climatic conditions to investigate the efficiency of thermal imaging in locating decomposing remains.
The heat signatures from carcasses originating from insect larval masses depend on various factors in order to be detectable on thermal imaging.Thus, one experiment was conducted to locate both clothed and unclothed pig carcasses in two different seasons, autumn and winter [9], using AMTI.It was found that despite the climatic variation, the temperature difference (8°C) was sufficient to detect larval masses across seasons, thus supporting the findings of the earlier study [10].
Furthermore, AMTI was also able to detect the heat generated by cadavers, either clothed or unclothed at a great distance even further than 457 m and at a longer decomposition stage (i.e., advanced decay stage).However, the use of a helicopter is costly and limited depending on the availability of a pilot, weather, and geographic conditions [16].
As an alternative, a drone (Unmanned Aerial Systems, UASs; often called Unmanned Aerial Vehicles, UAVs) equipped with thermal imaging cameras have demonstrated successful detection of heat signatures from decomposing bodies associated with larval aggregations [8,16].A drone is defined as an aircraft that operates without human intervention either piloted remotely or via a computer [17,18].Instead of using visible light to generate a picture, a thermal imaging camera, on the other hand, employs radiated heat (infrared radiation) to create one [19].With low cost yet easy to operate, drones with a thermal imaging camera provide a cost-effective solution to previous limitations.A drone equipped with a thermal camera has the potential of a more effective search technique than the use of a helicopter or conventional search.Drones could be useful tools for crime scene investigators because of their great accessibility, lower cost, smaller size, and capability to fly into difficult yet risky areas without putting investigators at risk.Moreover, a drone equipped with a thermal camera to detect decomposing remains via the heat signature from larval aggregations could be a new area to explore in forensic entomology.
The decomposition rate and insect succession to carcasses are affected by a number of intrinsic and extrinsic factors such as carcass body mass, age, nature of death, climatic conditions such as ambient temperature and relative humidity, and insect activity [20,21].In addition, the clothing of carcasses also plays an important role in insect fauna activities.Earlier studies have demonstrated that clothed carcasses have a faster decomposition rate due to greater insect diversity and feeding activity than unclothed carcasses [22][23][24].However, some studies have reported contrasting results that clothing affects insect activities negatively and hence delays the decomposition process [22,25].The inconsistency and paucity of data detailing insect activities associated with clothed and unclothed carcasses in tropical climates remain to be investigated.
Each geographical region is unique in terms of temperature, moisture, habitat, type of vegetation and soil, and ambient environmental conditions [20,26].To date, no specific literature has been conducted on the utility of aircraft-mounted thermal cameras, whether on UAVs or helicopters in tropical climate regions.Hence, there is a need to investigate the use of drone-mounted thermal cameras in detecting decomposing remains with maggot masses under this climate.Recently, Butters et al. [16] demonstrated that the dronemounted thermal camera could successfully detect heat signatures from body fragments of pig cadavers.They suggested that future research should focus on the minimum size or mass of cadavers that allow detection by the drone-mounted thermal camera.
Therefore, the present study aims to: (i) determine whether a thermal drone-assisted procedure can be utilized to detect any heat signatures by fly larval aggregations from small animal cadavers (both clothed and unclothed) in a tropical region, (ii) detect thermal signatures produced from decomposing carcasses and live subjects placed on the ground surface, and (iii) determine the minimum altitude of a thermal drone in the detection of heat generated from both maggotcolonizing and live subjects.The outcome of this study is to provide the optimum operating guidelines for thermal drone-assisted procedures for better detection of human remains in tropical regions.

| Study site
Experiments were conducted in a soccer field; the study site was flat, treeless, and covered with carpet grass (Axonopus compressus) within the campus of Universiti Teknologi MARA, Sungai Buloh Campus, Selangor, Malaysia with GPS coordinates; 3.224630, 101.593335, 68 m from sea level and 25.6 km northwest from the capital city of Kuala Lumpur, Malaysia.Permission for the land use had been granted from the university and no public access was allowed during the experiment.The soil at the study site belonged to the red, yellow podzolic type which is also known as forest soil with carpet grass.This region experiences a tropical climate which recorded a monthly rainfall of between 50 mm and 400 mm (Malaysian Meteorological Department, www.met.gov.my) and minimum and maximum daily temperatures between 21.4°C and 46.6°C.

| Animal carcasses
Twenty-seven healthy male European white rabbits (Oryctolagus cuniculus) with an approximate weight of 220-270 g purchased locally were used as proxies for human cadavers.Due to dietary restrictions and the difficulty in obtaining pigs, rabbits were chosen as proxies for human cadavers.Additionally, rabbits, being larger than rats, proved to be more suitable.A total of 24 rabbits were placed on the soil surface within 2 hours of death after being euthanized by 5 mL of 200 mg/mL phenobarbital injection intracardiacally and the remaining three rabbits were kept alive as control to serve as a baseline for normal body temperature and to differentiate thermal signatures between natural variations and carcasses.Twelve rabbit carcasses were utilized in each experimental trial; six of the carcasses were wrapped in 100% cotton fabric and six were left unclothed.The size of the cotton fabric was 0.08 × 1 m and of white color with RGB code (255,255,255).Cotton was chosen as the fabric to wrap the rabbit carcasses as it is the most common fabric material used by Malaysians due to its affordable price, lighter materials, cooler, and more durable quality.Animal ethics approval was granted by the Universiti Teknologi MARA Animal Research and Ethics Committee (UiTM CARE).
Each carcass was spaced 20 m apart from each other as recommended by Lee et al. [9] and was well visible from above using drones (Figure 1).The carcasses were placed alternately between clothed and unclothed.A PVC-coated wire mesh measuring 0.381 × 0.381 m was placed over each carcass to prevent any scavenging activities by larger vertebrate scavengers but allowed insect access (Figure 2).
The cages also were fixed to the ground with eight pieces of metal pegs to prevent movement by scavenging animals.The cages were retained during data collection since the PVC coating on the cage had no impact on the thermal signature of the carcasses, due to its stability over a wide range of temperature.Carcasses were left in the field for 14 days continuously during each trial.Trial 2 commenced a week after the completion of Trial 1 with the complete removal and disposal of the skeletal remains and was located 20 m away from the Trial 1 location to ensure independency of the generated outcomes.
The live rabbits also were placed within individual PVC-coated wire mesh enclosures, without being secured to the ground, allowing for their return to the animal house for resting and feeding.

| Sampling procedures
Two types of temperature recordings were obtained throughout the experimental period.A data logger, RC-4 (Elitech Temperature Data Logger, UK), was used independently to capture the hourly ambient temperature at the trial site for 24 days continuously while the larval mass temperatures on all rabbit carcasses were obtained by placing a mercury thermometer directly into the largest maggot mass observed at an approximate depth of 1 cm into the body cavity for 1 minute and repeated for three times.The temperature was recorded daily until there were no more maggot masses seen.
A DJI Matrice 300 RTK drone, equipped with a thermal camera: Zenmuse H20T (DJI, Shenzen, China), was utilized to record the thermal imaging footage of the carcasses and their larval masses.
The Matrice 300 RTK can capture and store data from its RGB zoom and wide-angle cameras, and thermal images were taken by the Zenmuse H20T camera on the Matrice 300 RTK all at once with just one click, featuring a remarkable zoom ratio of up to 200 times.

F I G U R E 1
The aerial layout of the study site comprises unclothed, clothed rabbit carcasses (numbers 1-12), and live rabbits (numbers 13-15).Twelve carcasses were spaced 20 m apart from each other.Carcasses in trial 2 were placed 20 m north of the previous carcass site.
In addition, the images captured can be analyzed using DJI analysis software which can provide temperature information such as the lowest, highest, and average temperatures.The thermal data colors can also be adjusted based on the needs of specific scenarios.All thermal images presented in the paper were set to the "iron bow" palette as it was able to quickly identify objects that emit heat in lighter and warmer colors while colder objects were in darker and cooler colors.
Images were captured directly at 90 degrees above the rabbit carcasses using the positioning mode drone program with zoom 20× and thermal zoom 4× at various altitudes (i.e., 15, 30, 60, 70, 80, 90 and 100 m above ground level) without removing the cages.
Multiple flight altitudes were conducted to determine the optimal height for efficient searching with high accuracy.A minimum altitude of 15 m was chosen, despite the H20T camera's 5-meter focal range due to the presence of nearby trees.The thermal cameras measured heat on an object's surface without any disruption from the drone.Hence, minimal or no thermal interference occurs as the prop downwash lacks the force to significantly affect the carcass temperatures [27].The image recordings were made on days 0, 1, 2, 5, 7, 9, and 14 post-mortem for each trial, from 9 a.m. to 11 a.m.As recommended by Lee et al. [9], the most favorable time for detection is when the temperature differences between carcasses and the ambient temperature are the greatest: early morning or late evening.
All drone pilots were licensed with flying permits obtained from the Civil Aviation Authority of Malaysia (CAAM).Throughout the experiment, each carcass was recorded photographically, and the stage of decomposition was documented each time the site was visited.
There were five stages of carcass decomposition observed in this experiment, namely, fresh, bloated, active decay, advanced decay, and the remains stage [8,28].

| Data analysis
Thermal images obtained from the thermal drone were analyzed using DJI Thermal Analysis Tool 3 software (DJI, Shenzen, China) for temperature measurement: minimum, maximum, and average temperature using a single image.In order to determine the optimum height for thermal detection, the contrast between the carcasses and the surrounding environment was determined using ImageJ Version Next, a repeated-measures ANOVA (RM-ANOVA) was used to determine the optimum height of the thermal drone over time.A pvalue of 5% or lower is considered statistically significant.

| Carcass decomposition
Observations on clothed carcasses were carried out after unwrapping the clothed carcasses to have a more detailed and accurate examination of the bodies.No differences were observed between clothed and unclothed carcasses in the duration of each stage in trials 1 and 2 (Figures 3 and 4).Both types of carcasses progressed through all previously mentioned stages of decomposition: fresh, bloated, active, advanced decay, and remains stage.Clothed carcasses were unwrapped before observations were made.In trial 1, the fresh stage was observed on day 0 (Figure 3A) and then progressed into a bloated stage on the next day, day 1 (Figure 3B), for both clothed and unclothed carcasses.Large egg masses belonging to blow flies (Diptera: Calliphoridae) with significant larval aggregations observed in the chest, genital, and stomach areas on day 2 (Figure 3C) for both clothed and unclothed carcasses indicating they have reached the active decay stage.The active decay stage lasted only 3-4 days before the carcasses started to dry and their skin had darkened marking the start of the advanced decay stage on day 7 (Figure 3E).The carcasses were continuously observed until day 14 (Figure 3F) where the rabbit carcasses were at the dry and remains stage.However, larval masses were still present in clothed carcasses on day 14 even at the remains stage.Differences were noted between trials in the total durations of each carcass decomposition stage.In trial 2, the carcasses progressed from the bloated stage to the active decay stage on day 2 which was similar to trial 1 (Figure 4A-F).However, the decay rate was higher after reaching the active decay stage with both clothed and unclothed carcasses entering the dry and remains stage after 5 days (Figure 4D).The dry and remains stage ended when the remaining dipteran larvae migrated away from the body.Newly emerged blow flies were observed resting on both clothed and unclothed carcasses on day 9 (Figure 4F) and none was observed on day 14.At the end of trial 2, all carcasses were at the remains stage (Figure 4G).

| Climatic conditions
Mean ambient temperatures during trial 1 ranged between 27.8 and 30.8°C (Figure 5) while they were slightly lower during trial 2 (ranging between 26.2 and 28.8°C) (Figure 6).The maximum daytime temperature was recorded on day 2 of trial 1 (43.2°C) and the minimum daytime temperature was recorded on day 5 of trial 2 (23.5°C).
Night-time temperatures were much cooler during trial 2, ranging between 22.5°C and 25.6°C.Precipitation was recorded during trial 2 on day 5 and day 8, with day 5 being recorded as the lowest temperature throughout the study period (i.e., 28 days) (Figure 6).

| Thermal images
In both trials, we observed that 15 m above ground level showed the greatest contrast between the carcasses' heat signature and the background heat, although 30 m and 60 m may also be adequately discernible (Figure 7A-G).From Figure 3, images taken at 15 m altitude with a ground sample distance (GSD) of approximately 1.333 cm/pixel were able to provide the highest contrast that made the carcass clearly observable, with a 158 intensity for 60 m (Table S1).Above 70 m onward (up to 100 m), lower thermal contrasts were observed and showed an overlap of contrasts between the carcasses and the background, thus making it difficult to detect the decomposing carcasses.There were no significant differences between 70 m (GSD 6.22 cm/pixel) to 80 m (7.11 cm/pixel) (p > 0.05) and 80 m to 100 m (GSD 8.89 cm/pixel) (p > 0.05).Based on these findings, 15 m altitude was proven to be the optimal altitude, and hence, the imaging results reported throughout this paper were only focused on 15 m altitude.Detectability of the thermal signature from carcasses varied depending on the decomposition stage, the associated level of insect activity present, and the clothing of the carcasses.While live rabbits show consistent thermal signatures in both trials 1 and 2 ranging between 27 and 35°C (Figures 5 and 6).During the fresh stage (day F I G U R E 4 Representative photographic images (rabbit #7, unclothed) for rabbit carcass decomposition (left) and corresponding thermal images (right) over days post-mortem during Trial 2. The dry and remains stage was achieved on day 5 (earlier than trial 1), and the temperature recorded was the highest (31.5°C).(F) showed the carcass temperature increase on day 9 but no heat signature was detected on day 14.0), all carcasses were clearly observable on thermal images for both trials 1 and 2 (Figure 8).Although the temperatures of the carcasses were recorded as slightly higher (±0.5°C) than the ambient temperature as shown in Figures 5 and 6, all the carcasses were still clearly detectable in contrast to the ambient ground surface.A two-way ANOVA (Table S2) revealed that there was a significant difference in carcass temperature between trials 1 and 2 on day 0 (F(1) = 6.59, p < 0.05), although the interaction between these trials (F(2) = 1.457, p > 0.05) and between types of carcasses was not significant.
Trial 2 recorded slightly higher carcass' temperature than trial 1: 3.1°C for unclothed and 0.7°C for clothed carcasses.No heat signatures were produced in all carcasses for both trials 1 and 2 during the bloated stage on day 1 (Figure 8).The mean temperature of carcasses for both clothed and unclothed recorded in trial 1 ranged between 23.2 and 23.9°C, lower than the mean ambient temperature, 30.8°C.This is consistent with trial 2 although the mean temperature of carcasses recorded was significantly different than trial 1 and ranged between 24.1 and 25.1°C (F(1) = 10.391,p < 0.05).We also found a significant difference in mean carcass temperatures between types of carcasses (F(2) = 33.15,p < 0.05), and further analysis using a Tukey's HSD Test revealed that the mean temperature of carcasses was significantly different between "live and unclothed" (p = 0.000, 95% CI = [3.764,8.069]) and "live and clothed" (p = 0.000, 95% CI = [4.606,8.911]) on day 1.
Although large egg masses with significant larval aggregations were observed on day 2 of both clothed and unclothed carcasses in trials 1 and 2, all carcasses still failed to produce a thermal signature (Figure 8).Prior to the formation of larval aggregations, carcasses were not readily detectable by the thermal images due to the lower temperature of the larval masses and average carcass temperature.
Nevertheless, the two-way ANOVA revealed that there was a statistically significant interaction between the carcass temperature of clothed and unclothed carcasses in both trials (F(2) = 9.99, p < 0.05).
Interestingly, this was the only time point that recorded a significant interaction between these trials.Thermal signatures were clearly observed for both clothed and unclothed rabbit carcasses during the peak period of insect activity, which was on day 5 for both trials 1 and 2. When fly larval aggregations increased, they generated the highest temperature observed either in or around the carcasses.On day 5 of trial 1, the maximum carcass temperature was 35.8°C for unclothed and 37.4°C for clothed carcasses (Figure 3d).The values were almost 7.2-8.8°Cabove mean ambient temperatures (±26.2°C).This observation was positively correlated with the larval mass temperatures on day 5 where the temperatures were the highest at 36.8°C for unclothed and 38.2°C for clothed carcasses; therefore, all rabbit carcasses were clearly observable on thermal images on day 5 until day 7 (Figure 8).In trial 2, fly larvae were present on the surface of carcasses until day 5 although the carcasses had progressed into the dry and remains stage, the thermal signature of the decomposing remains was still clearly detectable.
Figure 6 shows the mean larval mass temperatures of clothed and unclothed carcasses at 35.2°C and 37.7°C, respectively, during the dry and remains stages in trial 2. The two-way ANOVA showed that there was no significant interaction between clothed and unclothed carcasses and trials, indicating a similar mean carcass temperature on day 5 of trial 2 (F(2) = 0.504, p > 0.05).
We also found similar observations for day 7 where an insignificant interaction was obtained between clothed and unclothed carcasses and trials (F(2) = 1.806, p > 0.05).In trial 1, the carcass temperature was like live rabbits indicating higher temperature of larval mass (29 ± 1.0°C for both clothed and unclothed carcasses) was still recorded on day 7; therefore, no significant difference was observed between live rabbits, clothed, and unclothed rabbit carcasses (p > 0.05).There were also insignificant differences between trials 1 and 2 despite the lower carcasses' temperature on day 7 of trial 2 which was 24.3 ± 4.7°C for unclothed and 23.75.3°C for clothed carcasses (p > 0.05).
After the larval aggregations were no longer evident on the carcasses on day 14, the mean carcass temperature decreased to ambient temperatures or even lower.Therefore, imaging was difficult due to temperature similarity between ambient and carcass surface.This can be seen from the unclothed carcasses in trial 1 and both clothed and unclothed carcasses in trial 2 (Figure 8).The temperature recorded by the thermal drone showed the mean carcass temperatures were lower than the ambient temperatures on day 14 for both trials (Figures 5 and 6).Unlike clothed carcasses of trial 1, larval masses were observed and there were sufficient heat signatures produced from the carcasses.Even on the dry and remains stage on day 14, the carcasses could be visualized by the thermal images with the presence of larval masses (Figures 3 and 8).There was a significant difference between live rabbits and carcasses both clothed and unclothed despite no significant interaction between these factors were detected (F(2) = 1.036, p > 0.05).A Tukey post-hoc test revealed significant pairwise differences between live and unclothed rabbit carcasses (p = 0.000, 95% CI = [3.444, 10.189]), between live rabbits and clothed rabbit carcasses (p = 0.001, 95% CI = [2.603,9.347]), and between trial 1 and 2 (F(1) = 28.505,p > 0.05).
An interesting observation was made on day 9 of trial 2 where the temperature of the carcasses increased unexpectedly higher F I G U R E 7 (A-G) Thermal images of rabbit #2 from Trial 1 on day 5 taken from different flight altitudes started from 15 m until 100 m (GSD range from 1.333 cm/pixel to 8.89 cm/pixel).Fifteen meters above ground level showed the greatest contrast between the carcass heat signature and the background heat.The contrast between the carcasses and the surrounding environment was determined using ImageJ software by measuring the intensity of the heat signature produced by the carcasses.(H) The greatest contrast was observed at 100 m on day 7 where low ambient temperature was recorded, resulting in greater resolution of a thermal image.
(2.2°C) than day 7.The thermal signature of the decomposing remains was clearly detectable on day 9 compared to day 7 (Figure 4).This could be due to precipitation that occurred a day before which softened the dried skin and facilitated larval aggregations.No thermal images were recorded on day 9 for trial 1 due to data corruption during image capturing.A one-way ANOVA was conducted on day 9 of trial 2 to compare the temperature differences of clothed, unclothed, and live rabbits on thermal images.There was a significant difference in mean temperatures between at least two groups

| DISCUSS ION
The current study was conducted to extend previous research on forensically important fly larval masses associated with decomposing remains using a thermal drone.Analysis of the thermal images demonstrated that rabbit carcasses (Oryctolagus cuniculus) were successfully detected with a thermal drone during the decomposition process.These findings were in agreement with Butters et al. [16] that a drone-mounted FLIR camera could detect thermal signatures from both small body fragments and whole pig carcasses.Despite the small carcass size, the thermal drone is able to capture the thermal signatures even on days 0, 5, and 7; pig carcass fragments were clearly observable from day 3 until day 16 [16].The rapid decomposition rate in this study could be attributed to the small size of the carcasses.It was found that larger carcasses attracted more insects and prolonged the decomposition stages especially the active decay [29]; thus, thermal detection was longer than in the present study.
Therefore, thermal imaging is a useful tool in documenting the thermal signatures within a small carcass which mimics a human body dismemberment, and this study could set the minimum size of detectability during search and rescue operations.
The process of decay was articulated in five successive stages (i.e., fresh, bloated, active decay, advanced decay, and remains) as previously described [23].These results are in line with Singh & Bala, [30] where insect activity increased from the bloated stage, F I G U R E 8 Thermal images of both trials taken from an altitude of 15 m.Days 0, 5, and 7 recorded the greatest intensity of thermal images on rabbit carcasses.Thermal signatures were clearly observed for both clothed and unclothed rabbit carcasses during the peak of insect activity, that is, day 5 for both trials and slightly reduced on day 7.It was clearly observed that clothed rabbit carcasses produced higher intensity than unclothed rabbit carcasses even though they had no significant difference (p > 0.05).The clothed rabbit carcass of trial 1 still can be detected on day 14 which was in its dry and remains stage.
greatest at active decay, and declined during the remains stage due to the scarcity of food sources.The result of our study showed that there was no difference in the rate of decay between clothed and unclothed carcasses throughout the experiments.Our results coincided with the results of [31,32] that clothing had no effect on the decay rate or insect succession.In terms of thermal images, clothed carcasses were found to possess higher thermal image intensity compared to the unclothed carcasses on day 5 (Figure 8) although it was not statistically significant.Despite apparent differences in thermal intensity from visual inspection, the statistical analysis revealed no significant variance between treatments.This discrepancy may arise from the statistical focus on specific temperature differences, which, given the minimal variance (below 1°C) between treatments, was deemed negligible.Our results accord with a previous report that clothing did not have any noticeable effect on the temperatures generated by the fly larval masses [9].Taken together, these results suggest that clothed and unclothed carcasses do not have a significant difference in the surface thermal temperature.
In this study, rabbit carcasses showed similar internal temperatures with live rabbits on days 0, 5, and 7. Thermal images captured after 2 h of death showed the greatest contrast with the background on day 0 due to the rise in carcass temperature.The elevated temperature of the carcasses after death could be due to the continuation of internal consumption and enteric bacterial activity [14,33].
Similar findings were reported on days 5 and 7 revealing that the fly larval mass temperature produced was similar to the live rabbit's body temperature.The greatest thermal signature detected by thermal drone was reported previously during the active decay stage when the fly larval aggregations were most active; day 5 [10].This stage lasted until day 7 and provided clear thermal images.A strong correlation was established between thermal image prominence and with amount of heat emitted from the larval masses [34].
The thermal images obtained clearly illustrate the significant differences in temperature of live rabbits compared to clothed and unclothed carcasses on days 1, 2, and 14.The findings were correlated with the stage of decomposition whereby on days 1 and 2, both carcasses were in the bloated stage.No heat signatures were detected on the deceased carcasses on day 14; the remains stage except on clothed carcasses in trial 1.A few maggots were collected for clarification and confirmed that the Chrysomya megacephala (Diptera: Calliphoridae) larvae found on the carcasses were in their third instar of development.It is well reported that the heat produced by larval masses is influenced by the larval instar [7].This observation aligns with the findings of Matuszewski et al. [29], reporting the presence of Piophilidae after the advanced decay stage.We postulated that the thermal signatures were still produced because older larvae generated more heat through aggregation and could regulate their temperature for survival, provided there was enough food to sustain their activity [24,35].A noteworthy observation occurred on day 9 of trial 2, where the thermal signatures were higher than those on day 7.Despite the carcasses being in the remains stage, the larval aggregations were more evident on day 9 likely due to the precipitation that occurred on day 7.We suggested that the moisture provided by precipitation might create a more favorable environment for maggots development, leading to an increase in their activities, and thus higher thermal emissions were recorded.
As the carcasses entered advanced decay without larval masses, rabbit carcasses remained detectable though less prominently, despite a small mean temperature difference of 0.1°C from the ambient temperature.These results enhance the detection limit of the previous research, which suggested a detectable temperature difference of 8°C [9].Moreover, in this tropical region, larval aggregations can exhibit a maximum temperature exceeding 5.5°C above the ambient temperature.This significant contrast results in clear visibility of the carcasses in thermal images (Figure 8).Our data also suggested the thermal drone's detectability is up to 7 days post-mortem in this tropical region.The warm ambient temperature and high humidity accelerate fly larval feeding on the carcass and skeletonizing the carcass within 2 weeks compared to temperate countries in Europe and America [9,15,36].Therefore, the present study improves the lower limit of thermal drones' detectability (5-7 days) under tropical climates.
Our study enhances the detectability of the thermal drone's altitude proposed by [16] where they used DJI Zenmuse XT thermal imaging camera and concluded that the detectability of the carcasses was greatest below 30 m.In comparison, our study employing the DJI Zenmuse H20T showed the thermal signatures were allowed up until 60 m with a good resolution where the carcasses can be differentiated from the surrounding environment (Figure 7A-C).This advanced thermal drone has better pixel pitch giving a higher resolution than the XT and incorporating a more advanced wide and zoom camera, and a laser rangefinder making the H20T the best thermal camera to be used during rescue operations [37].For altitudes above 70 m until 100 m, the low resolution of thermal images and greater masking prevented detection although a greater area could be monitored.However, we observed that detectability was still achievable until 100 m for carcasses that showed the greatest contrast to the surrounding environment (Figure 7H).This was only possible when the carcass produced the greatest temperature difference to the surrounding environment and during the peak feeding period.
The current and previous studies have demonstrated the potential of a more effective search technique utilizing a thermal drone and its capability to detect the heat generation produced by fly larval aggregations.Different factors including drone height, larval aggregations, and ambient temperature were shown to limit the capability of detecting heat sources.In the current study, all observations and thermal images were possessed during the early day between 9 and 11 a.m.Although thermal drones can work properly during the day, some of the images capture radiant heat from the surrounding areas, thus creating a masking effect on the carcasses.The direct sunlight in the surrounding area makes it difficult to effectively measure the temperature and decreases the visibility of the carcasses.
The thermal drone is a heat-sensitive device; therefore, in order to successfully utilize the device in search missions, a few parameters can be adjusted to optimize the success of the device.[9] demonstrated the fly larval aggregation activity was more successful during 1.47 software by measuring the intensity of the heat signature produced by the carcasses.Statistical analysis was conducted using IBM SPSS version 27.The normality of the data used for statistical inference was tested using the Kolmogorov-Smirnov and Shapiro-Wilk tests prior to other statistical tests.Comparison of means and standard deviations (SDs) were calculated for all the parameters evaluated for each decomposition day.A Two-way ANOVA was performed to analyze the differences in carcass temperatures between clothed, unclothed, and live rabbits in trials 1 and 2 of each observation day.

F I G U R E 2
Each rabbit carcass was placed in 0.381 × 0.381 m PVC coated wire mesh and fixed to the ground with 8 pcs of metal pegs; unclothed rabbit carcass (left) and clothed rabbit carcass (right).
value (I) compared to other altitudes.The contrast was measured using ImageJ software by measuring the intensity differences of heat signatures produced by the carcasses and the background.A RM-ANOVA with Greenhouse-Geisser correction showed that the means of thermal image contrasts differed significantly between drone heights (F(2.798,11.193) = 1173.574,p < 0.0005).The Fvalue (1173.574) is a ratio of the variance between groups to the F I G U R E 3 Representative photographic images (rabbit #3, clothed) during rabbit carcass decomposition (left) and corresponding thermal images (right) over days post-mortem during Trial 1. (A-F) showed rabbit carcasses from fresh until dry and remains stage.A great intensity was observed during the active decay stage; from day 5 to 7. Heat signatures can still be detected on day 14, dry and remains stage.variance within groups while (2.798, 11.193) are the degrees of freedom used in the analysis.Post-hoc analysis with a Bonferroni adjustment was conducted after significant differences in the means of thermal image contrasts across drone altitudes were determined which revealed specific statistical differences, with associated mean values, confidence intervals, and p-values for each altitude comparison.The analysis revealed that 15 m (GSD 1.333 cm/pixel), 30 m (GSD 2.667 cm/ pixel), and 60 m (GSD 5.33 cm/pixel) of drone altitudes were statistically significant with the highest contrast (156 [95% CI, 150.11-161.89],p < 0.0005) for 15 m, (105.84 [95% CI, 103.198-108.482],p < 0.005) for 30 m, and (76.94 [95% CI, 73.493-80.387],p < 0.05) Trial 2 recorded lower temperatures than trial 1; 21 ± 2.9°C for unclothed and 21.1 ± 2.1°C for F I G U R E 5 Average temperatures of unclothed rabbit carcasses, clothed rabbit carcasses, and live rabbits were recorded by thermal images of Trial 1 while ambient temperatures were recorded by a data logger.No thermal images were recorded on day 9 for trial 1 due to data corruption during image capturing.The average ambient temperature varied between 27.8 and 30.8°C.There was a significant difference in temperatures between live and unclothed or clothed rabbit carcasses on days 1, 2, and 14.Higher temperatures of rabbit carcasses were recorded on days 0, 5, and 7 of decomposition day.F I G U R E 6 Average temperatures of unclothed rabbit carcasses, clothed rabbit carcasses, and live rabbits were recorded by thermal images of Trial 2 while ambient temperatures were recorded by a data logger.The average ambient temperature varied between 26.2 and 28.8°C.There was a significant difference in temperatures between live and unclothed or clothed rabbit carcasses on days 1, 2, and 14.Higher temperatures of rabbit carcasses were recorded on days 0, 5, and 7 decomposition days.clothed (F(1) = 8.189, p < 0.05) carcasses.Simple main effects analysis showed that live rabbits did show a significant difference in temperatures between clothed and unclothed carcasses (p < 0.05).