JP2007292645A - Device and method for discrimination of bat species - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device and method for discriminating the species of bats from the ultrasonic wave signals by receiving the ultrasonic wave signals of complicated patterns of frequency band width which is emitted by the bat. <P>SOLUTION: The device comprises the ultrasonic wave receiver 2 for receiving the ultrasonic wave emitted by the bat 1, the ultrasonic wave signal analyzer 3 for analyzing the signal from the ultrasonic wave receiver 2 and discriminating the bat species from the analysis result. The ultrasonic wave signal analyzer 3 is characterized by discriminating the bat species by using information at least one of an upper limit frequency, a lower limit frequency and an energy peak frequency of the ultrasonic frequency emitted by the bat. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an apparatus and a method for discriminating bat species by analyzing ultrasonic waves emitted by bats. By distinguishing bat species, it can be used for environmental assessment.

  In recent years, environmental destruction such as forests has become serious. For environmental surveys of forests, etc., surveys of animals and plants are usually conducted. However, investigators must be in the mountains and it is difficult to conduct regular surveys. In particular, research on animals is extremely difficult because they run away when an investigator approaches. Although it is conceivable to perform fixed-point observations using a television camera or the like, the presence or type of animals cannot be determined unless the investigator visually recognizes them, and the burden is large. Furthermore, the field of view of television cameras is limited, and a wide range of investigations are not possible. A method that can automatically and quantitatively observe the type and distribution of animals is desired, but at present there is no effective method.

  Bats are known to emit ultrasound to measure the position of obstacles and prey. The inventor of the present invention has found that there are characteristics in the frequency and generation pattern of the emitted ultrasonic wave for each bat species through bat research for many years. There are about 30 types of small bats in Japan, but bats prefer different environments, so if you can monitor the ultrasonic waves generated by bats and identify bat species, you can understand the environmental conditions of the place to some extent. can do. If monitoring is performed for a long period of time, it is also possible to grasp environmental changes, which can be used as environmental monitoring. Unlike a TV camera, ultrasonic observation has fewer restrictions on the field of view, and the observed ultrasonic waves can be analyzed automatically, enabling quantitative automatic observation. By installing devices at multiple locations in the forest and transmitting the observed information via radio waves, etc., even if the investigator does not regularly enter the back of the mountain, it automatically changes the forest environment over a long period of time. Monitoring is possible.

Patent Document 1 suggests that the frequency change pattern and period of the ultrasonic wave are different for each type of bat. However, Patent Document 1 does not describe a specific configuration for discriminating bat species. It is only described that the frequency changes, and it is neither described nor suggested that the bat emits a complex pattern of ultrasound having a frequency bandwidth.
JP 2003-304797 A

  The present inventor has found through long years of research that bats do not simply emit ultrasonic waves of a specific frequency, but emit ultrasonic waves of a complex pattern with a frequency bandwidth. Therefore, it is difficult to distinguish bat species by simply observing the ultrasonic frequency.

  An object of the present invention is to provide an apparatus and method for receiving an ultrasonic signal of a complicated pattern having a frequency bandwidth, which is emitted from a bat, and discriminating a bat species from the ultrasonic signal.

  In order to solve the above problems, the present invention has the following configuration.

  A bat species discriminating apparatus comprising: an ultrasonic receiver that receives an ultrasonic wave emitted by a bat; and an ultrasonic signal analyzer that analyzes a signal from the ultrasonic receiver and discriminates a bat species from an analysis result. The ultrasonic signal analyzing apparatus is to determine a bat species using bat ultrasonic information including at least one of an upper limit frequency, a lower limit frequency, and an energy peak frequency of an ultrasonic wave emitted by the bat.

  An ultrasonic reception step of receiving ultrasonic waves emitted by a bat, and an ultrasonic signal analysis step of analyzing the received ultrasonic signal and discriminating a bat species from an analysis result, wherein the bat species discrimination method comprises: The ultrasonic signal analysis step is to determine a bat species using bat ultrasonic information including at least one of an upper limit frequency, a lower limit frequency, and an energy peak frequency of an ultrasonic wave emitted by the bat.

  Moreover, it has the following embodiments.

  The ultrasonic signal analyzing apparatus or process is to determine a bat species using a search period echo location call emitted by a bat.

  The bat ultrasound information further includes at least one of a sonar pattern and a frequency change pattern, and is to determine a bat species from a combination of a plurality of information.

  The sonar pattern includes at least one of a pulse duration, a pulse number, and a pulse interval of an ultrasonic wave emitted by a bat.

  The ultrasonic signal analyzing apparatus has a bat ultrasonic information table storing a plurality of types of known bat ultrasonic information, and the ultrasonic bat ultrasonic information generated by the bat received by the ultrasonic receiver It is to discriminate bat species by comparing with the bat ultrasound information table.

  A recording / reproducing device connectable to the ultrasonic receiver and the ultrasonic signal analyzing device, wherein the recording / reproducing device records an ultrasonic signal from the ultrasonic receiver, and at the time of ultrasonic signal analysis, The recorded ultrasonic signal is reproduced by connecting to an ultrasonic signal analyzer.

  Bats emit a complex pattern of ultrasonic waves with a frequency bandwidth, but it is important to note that among the ultrasonic waves with that frequency bandwidth, the upper limit frequency, lower limit frequency, and energy peak frequency differ for each bat species. The inventor has discovered through years of research. Here, the energy peak frequency is a frequency at which energy becomes maximum in an ultrasonic wave having a frequency bandwidth. Not only simply detecting the frequency of the ultrasonic wave, but also analyzing the ultrasonic wave with the frequency bandwidth, calculating the upper limit frequency, the lower limit frequency, and the energy peak frequency, and using these parameters to distinguish the bat species The bat species can be identified more reliably. In addition, since the sonar pattern (pulse duration, number of pulses, pulse interval) and frequency change pattern are different for each bat species, the bat species can be more reliably discriminated by taking these factors into account. By installing this bat species discriminating apparatus in the forest, it is possible to monitor the surrounding bat species and the frequency of appearance of each bat, which can be used for environmental monitoring.

  FIG. 1 is a conceptual diagram of an embodiment of the present invention. The ultrasonic wave emitted by the bat 1 is received by the ultrasonic wave receiver 2. The ultrasonic signal received by the ultrasonic receiver 2 is sent to the ultrasonic signal analyzer 3. A recording / reproducing apparatus 4 that can be connected to the ultrasonic receiver 2 and the ultrasonic signal analyzing apparatus 3 may be provided. In this case, the ultrasonic signal can be once recorded in the recording / reproducing apparatus 4, and the recorded ultrasonic signal can be reproduced later for signal analysis. The ultrasonic signal analysis device 3 discriminates bat species based on the received ultrasonic signal. The bat species discrimination information analyzed by the ultrasonic signal analysis device 3 is displayed on the display device 5 or transmitted to the external device via the external transmission device 6.

  FIG. 2 is a block diagram of the ultrasonic signal analyzing apparatus 3. The ultrasonic signal received by the ultrasonic receiver 2 is sent to the frequency / sonar pattern analyzing means 31. The frequency / sonar pattern analyzing means 31 calculates an upper limit frequency, a lower limit frequency, an energy peak frequency, a frequency change pattern, a pulse duration, a pulse interval, a pulse number, and the like of the ultrasonic signal. The frequency / sonar pattern analysis data calculated by the frequency / sonar pattern analysis means 31 is sent to the bat species discrimination means 32. The bat ultrasound information table 33 stores ultrasonic frequency / sonar pattern analysis data of a plurality of known bats. The bat species discriminating means 32 compares the frequency / sonar pattern analysis data sent from the frequency / sonar pattern analysis means 31 with the known bat frequency / sonar pattern analysis data stored in the bat ultrasound information table 33. The bat species is discriminated and the discrimination result is output.

The experimental results are shown below.
Introduction Voice signals of bats are roughly classified into echo location calls and social calls based on their functions. The echo location call is mainly issued intermittently during flight, and is a means of detecting objects (such as food) and their traits in space and obtaining information such as the position of other individuals. Depending on the species of bats, there are a variety of species that use frequency bands around 11 kHz (Zbinden & Zingg, 1986; Fullard & Dawson, 1997) to voice pulses for echo location calls (Fenton & Bell, 1981). . There have been many reports on the functions of echo location calls and social calls, but little has been done on acoustic properties. The echo location call is classified into an FM (Frequency Modulation) pulse and a CF (Constant Frequency) pulse according to acoustic characteristics, and can be further classified into a search period, a tracking period, and an end period according to the behavior when the sound is emitted. The exploration phase is considered to be an ideal audio signal for studies that distinguish species acoustically (Murray et al, 2001).

  Bats communicate primarily through olfaction, touch, and hearing (Fenton, 1985). Information between individuals when individuals are separated is considered to be exchanged by social calls (Pfalzer & Kusch, 2003). Social calls are issued under various conditions, such as during flight and roosting, and their functions can be used for distress calls, warning calls, isolation calls, direction calls, Categorized as mating calls (Pfalzer & Kusch, 2003). It is impossible to strictly differentiate between these two types of audio signals because echo-location calls may be used for communication within the same species (Fenton, 1985). The difference between the social call and the echo location call is that the social call itself mainly has a communication function. Bats tend to make the social call frequency band lower than the echo location call to avoid tuning the frequency band of the echo location call and social call (Pfalzer & Kusch, 2003), but the social call pulse structure varies. Yes, it can be an index for distinguishing species.

  In this experiment, for the purpose of distinguishing the species of bats by audio signals, the echolocation call of 7 genus of 11 Japanese bats, 1 genus of 1 Korean bat and 4 genus of Japanese bats We collected 5 types of social calls from 1 genus of Korean bats and analyzed the acoustic characteristics of the audio signals.

Audio signal recording Audio signals were recorded at known roosts and foraging sites of bats in Hiroshima, Yamaguchi, and Saishu Island, Korea. The breakdown of the genus is Species of the genus Eurynia genus (Miniopterus fuscus), Greater bats (Myotis macrodactylus), Abra bats (Pipistrellus abramus), Greater bats (Pipistrellus endoi), (Rhinolophus cornutus), chrysanthemum bat (Rhinolophus ferrumequinum), toad bat genus (Tadarida insignis), exploration echo location call of the bat genus bat (Vespertilio superans), follow-up echo location call of the bat , Social calls of the long-browed bat, white-headed bat, bat and chick. It also uses haze nets in the vicinity of known roosts, such as the long-tailed bat, the genus bat genus (Murina hilgendorfi), the bat bat (Murina silvatica), the giant bat genus bat (Myotis daubentonii), and the noren bat (Myotis nattereri), catching the bat, indoor Proboscis bats, Proboscis bats, Doventon bats, Noren bats exploration echo location calls, long-tailed bat follow-up echo location calls, final echo location calls, , Recorded social calls of Dobenton Bats, Noren Bats, Abra Bats. For recording, an ultrasonic detector (time expansion type, frequency range 10 kHz-150 kHz) was used, the time expansion factor was increased 10 times, and the audio signal was recorded on a digital audio tape with a digital tape recorder.

  FIG. 3 is a diagram showing the frequency range and duration of the search period echo location call for each bat species. The horizontal axis in FIG. 3 represents time, but the bat species are arranged side by side so that the pulse durations of the bat species can be easily compared. One scale on the horizontal axis indicates 0.1 second.

  FIG. 4 is a diagram showing the sonar pattern of the echo location call in the search period, the follow-up period, and the final period in the long-tailed bat and the bat. The horizontal axis indicates time, and one scale indicates 0.2 seconds.

Audio signal analysis Of the recorded bat audio signals, a total of 263 echo location calls of 7 genera and 12 species and 224 social calls of 4 genera and 6 species are digitally imported into a personal computer for acoustic analysis. Analysis was performed using the soft BatSound Pro (Pettersson, Sweden).

  The upper limit frequency (F-max), lower limit frequency (F-min), energy peak frequency (E-max), and pulse duration (duration) of the frequency band of audio signals, and the number of pulses that make up the signal only for social calls ( pulses) was the analysis item. Hereinafter, it is set as F-max, F-min, E-max, duration, and pulses. Use spectrogram (FFT size 512, Hanning window) to analyze F-max, F-min, and pulses, use power spectrum (FFT size 1024, Hanning window) to analyze E-max, and analyze duration. The oscillogram was used. Also, referring to the literature on the structure of social call pulses (Pfalzer & Kusch, 2003), the structure of each social call pulse is squawk type (A type), repeat type (B type), curved type (C type), complex type. Classified into (D type). Hereinafter referred to as A type, B type, C type, and D type.

Statistical analysis
Statistical analysis was performed using a total of 253 echolocation calls of 6 species and 10 species of bats emitting FM pulses and a total of 224 social calls of 4 species and 6 species. For analysis, discriminant analysis was performed using the software Excel Statistics 2002 (SSRI) for multivariate analysis.

BatSound
F-max, F-min, E-max, duration, and pulses of bat audio signals analyzed by Pro, and echo call by interdisciplinary, intergeneric, intergeneric, interspecies, social call 4 For each type (A type, B type, C type, D type), the discriminant predictive value and standard discriminant coefficient between different species and between different species in the same genus were investigated.

Results Acoustic properties of echolocation calls Table 1 shows the acoustic properties of bat echolocation calls that could be recorded in this experiment.

  The bat, bat, bat, dobenton bat, white bat, noren bat, bat, morea bat, giant bat, bat can be categorized into species that emit echo location calls of the FM pulse, The horseshoe bat can be classified into species that emit echo location calls of CF pulses. Echo location calls ranged from giant bats using a minimum frequency band of 9.7 kHz to proboscis bats using a maximum frequency band of 110.0 kHz, and the pulse structure was similar within the genus. The only thing that was directly heard by the human ear was the Echo Location Call of the Toad Bat.

  Focusing on species that emit FM pulses in the search-time echo location call, the echo location call of the toad bat and chick bat has characteristics that the duration is longer and the frequency band to be used is lower than other species, The frequency bands used by the bat, the doventon bat and the white-tailed bat, and the long-tailed bat and the long-tailed bat were close to each other. The Doventon bat and the Noren bat sometimes emitted a noise-type pulse consisting of two FM pulses. With the exception of giant bats and chick bats, the long-tailed bat duration was short (average 2.3 ms) and the long-tailed bat duration was long (average 9.3 ms).

  In addition, the long-tailed bats had higher F-max and steep pulses than other genera. As for the acoustic characteristics of the search echo location call in bats that emit FM pulses, individual differences in F-max were conspicuous, and there were individual differences in duration for species with lower frequency bands. The echolocation call of the white-winged bat and the white-winged bat has a few FM pulses before and after the CF pulse, the structures of the emitted pulses are similar to each other, and there is a clear species difference in the frequency value of the CF pulse.

  The long-tailed bat and the bat have expanded the frequency band to be used by increasing the F-max and lowering the F-min when the behavior when the pulse is emitted enters the tracking period from the search period. At this time, the duration was shortened to increase the repetition rate of the pulse, which was 150 to 170 per second. At the end, the frequency band to be used was lowered, the duration was shortened to increase the pulse repetition rate, and it became 190-210 per second. Echo location calls in the exploration and follow-up periods of the long-tailed bat have a higher frequency band than that of the bat, but at the end, the frequency band used by the bat was higher overall.

Table 2 shows the acoustic characteristics of social calls of bats that could be recorded in this experiment.

  The social call of the long-tailed bat that was able to be recorded is A type, B1, B2, C1, C2, C1, D1, D2, D3, D4 frequency band from 10.4kHz to 88.1kHz In the Doventon bat, the frequency band used by the B type is 11.0 kHz to 46.1 kHz, and in the white bat, the frequency band used by the A type, B type, D1 type, and D2 type is 10.1 kHz to 56.8 kHz, the Noren bat In A type and B type, the frequency band used from 19.0kHz to 55.4kHz, and in the case of the bat, the frequency band used by A type, C1, C2, D1, D2, D3 type is 11.8kHz to 109.3kHz, chick In the bat, the frequency band used for the C type was 6.5kHz to 29.7kHz.

  A type social call is a noisy voice signal, and the structure of the pulse is similar, and compared to other types of social calls, the frequency band used by all bats that make type A social calls is It was low and could be heard by human ears. Compared with other species, the Norenbat A-type social call has a higher frequency band, and the frequency elements (F-max, F-min, E-max) and duration are small. On the other hand, the type A social call of the long-tailed bat, the white-tailed bat, and the bat bat had large individual differences in both frequency element and duration. The great white bat made an A-type social call when the author invaded the roost, while the noren bat and the bat were fired when they were captured, and the long-tailed bat could be emitted in either case.

  Type B social calls are voice signals that emit very short pulses continuously for a long time compared to other types of social calls, and the structure of the pulses is similar for all species. Doventon bat type B social calls tend to have lower F-mins for all signals compared to other species, while Noren bats tend to have higher F-max for all signals compared to other species. It was. The frequency component of the type B social call between the long-tailed bat and the white-winged bat is large among individuals, the doventon bat, the white-headed bat, and the noren bat duration are small, and the long-tailed bat, the white-headed bat and the white-tailed bat, the doventon Bats and Noren bats were close to each other in pulses. The long-tailed bat and the white-winged bat made a B-type social call when the author invaded the roost, while the doventon bat and the noren bat could shoot when captured.

  The C-type social call has a simpler pulse structure compared to another type of social call, and the bat C-type social call resembles an echo-location call with a gentle slope FM pulse. The C-type social call of bats uses a lower frequency band than other species, and the frequency component of the C-type social call in the long-tailed bat and the bat is large among individuals, but it is smaller in the chick bat. It was. The long-tailed bat made a C-type social call when the author invaded the roost, while the chick bat was sometimes heard during the flight and was always heard by the human ear. The bat made a C1 type social call during the flight, and when the mother and the sub-animal faced each other, the sub-animal sometimes made a C2 type social call.

  The D-type social call has a wide frequency band (12.0 kHz to 109.3 kHz) used by all the bats in this experiment, and the structure of the pulse is diverse. The D-type social call of the long-tailed bat, the white-tailed bat, and the bat bat had less individual differences in frequency components than other types of social calls. The structure of the D-type social call pulse in the bat was very undulating, and distinct species differences were observed compared to other species. The number of pulses was higher than that of other species, and the pulses of the long-tailed bat and white-headed bat were close to each other. The long-tailed bat and the white-tailed bat made a D-type social call when the author invaded the roosting mating season, while the bats could be emitted during the mating season.

Discrimination of echo location calls by statistical analysis Table 3 shows the inter-species, inter-generic, intra-generic inter-species, and follow-up and final echo locations in the search period echo location call for bats that emit FM pulses by discriminant analysis. Shows the discriminatory probability between calls of different types.

  The discriminant predictive value between the different types of search location echolocation calls was 83.4%, and the discriminant predictive value of the long-necked bat, the doventon bat, the noren bat, the morea-bra bat, the giant bat and the chick bat was 100%. The bat echo location call is a bit more likely to be the Echo location call, and the discriminatory probability is 70.7%. The white bat echo location call was sometimes identified as the doventon bat echo location call, but not the other way around. Similar results were also found among the long-necked bat, the long-necked bat, the white-headed bat and the noren bat. The index of discrimination between different species was F-min (Table 4).

  The discriminant probability between different genera of the search period echolocation call was 85.8%, and the discriminant probability between the genus Aphidium and the chick genus was 100%. The genus Bat genus, Proboscis bat genus and Giant bat genus also each had a discriminatory median of 95% or more, and were clearly distinguishable. The echolocation call of the genus Brassica is somewhat discriminated as the echolocation call of the genus Bat bat, with a discriminative predictive value of 71.4%. The standard discriminant coefficient approximated all analysis items, but the index for discrimination between different genera was E-max (Table 4).

  The discriminant predictive value between the species of the genus bat genus in the search period echolocation call was 90.5%, the difference between the species of the genus bat genus was 100%, and the difference between the species of the genus bat genus was 100%. Proboscis bat echo location calls were sometimes identified as proboscis bat echo location calls, but not the other way around. The index for discriminating between different species of the genus Bat genus was duration, F-max, F-min and duration were found between species of the genus Giant bat, and F-min was found between species of the genus Bat bat (Table 4).

  The discriminative probability of follow-up echolocation calls between different species was 85.4%, and the ubiquitous bat and the bat were sometimes identified as one type of echolocation call. The indicators for discrimination between different species were F-min and E-max (Table 4). On the other hand, the discriminatory probability between different species of the final echolocation call was 97.6%, and it was possible to distinguish the species clearly. The bat echolocation call was sometimes identified as the ubiquitous bat echolocation call, but not the other way around. The index for discrimination between different species was E-max (Table 4).

Discrimination of social call by statistical analysis Table 5 shows the discriminant predictability between different species and the same genera of bats that make a social call by discriminant analysis.

  The discriminatory probability of type A social calls between different species was 70.4%, and each species of the long-tailed bat, white-headed bat, noren bat, and bat was sometimes distinguished from other species. None of the species showed high hit rates and could not be clearly distinguished. The type A social call of the bat was identified as the type A social call of the long-tailed bat, but not the other way around. The index for discrimination between different species was E-max (Table 6).

  The discriminatory probability of type B social calls between different species was 71.6%, and each species of the long-tailed bat, the doventon bat, the white-headed bat, and the noren bat was sometimes distinguished from other species. The type B social call of the white-tailed bat may be identified as a type B social call of all other species, with a low discriminatory probability of 57.1%. The B-type social call of the long-tailed bat was sometimes identified as a doventon bat, but not the other way around. Indicators for discrimination between different species were E-max and duration (Table 6). The discriminant predictive value between the species of the genus Giant bat was 87.7%, and the discriminant predictive value of the species Dawbenton bat and Noren bat was 100%. However, the B-type social call of the white-headed bat was sometimes identified as the B-type social call of the Dobenton bat and the Noren bat. Indicators for discrimination between different species were duration and pulses (Table 6).

  The discriminant probability of different types of C-type social calls was 90.0%, and the discriminant probability of chick bats was 100%. The C-type social call of the long-tailed bat and the bat was sometimes distinguished from each other, and the C-type social call of the bat was sometimes identified as the chick bat, but not the other way around. The index for discrimination between different species was E-max (Table 6).

  The discriminatory probability of D-type social calls between different species was 85.2%, and each species of the long-tailed bat, white-headed bat, and bat bat could be distinguished from other species. The D-type social call of the long-browed bat was slightly discriminated as the D-type social call of the white-tailed bat. The indicators for discrimination between different species were F-max and E-max (Table 6).

Consideration Acoustic characteristics of echolocation calls
Some bats that emit FM pulses contain a small number of CF pulses at the rear of the emitted pulse, and there is a view that this is FM / CF type (Vaughan, et al, 1997). In this study, bats that produce FM / CF are classified into the genus Bat genus, Brassica genus, Capricorn bat genus, and Cypress genus. In general, however, bat-generated pulses are broadly divided into FM and CF types. In this study, all genus other than the genus B. bats were analyzed as bats emitting FM pulses. However, as a result, the genus emitting FM type pulses is identified as the genus emitting only FM type pulses, and the genus emitting FM / CF type pulses is identified as the genus emitting only FM / CF type pulses. The mid rate was 100%. As a result, the classification of FM pulses into FM type and FM / CF type is supported, and further acoustical classification is expected.

  It is speculated that some of the high-band frequencies of FM pulses emitted by bats are severely attenuated by the atmosphere (Fenton & Bell, 1981). Therefore, as an acoustic characteristic of echo location calls in bats that emit FM pulses, F-max has individual differences compared to F-min and E-max, resulting in a low percentage of species identification indicators. It is thought that it became. Also, due to the Doppler effect, the frequency of the recorded audio signal is modulated by the direction of the ultrasonic detector and the bat that emits the audio signal. According to Jones & Parijs, 1993, the frequency of signals emitted by bats affected by the Doppler effect modulates up to 2 kHz. The echolocation calls of the Moria bats, the toad bats, and the bats that could be recorded outdoors in this study were recorded without considering the Doppler effect. As a result of re-discriminant analysis assuming that all of these echo location calls were affected by the Doppler effect, there was no fluctuation in the discriminant predictive value. In this study, the species of bats that make echo location calls It can be considered that the effect of the Doppler effect was small.

  Most bats are messy places, for example, expanding the frequency band of pulses (Murray et al, 2001). Probably, the characteristics of the pulse are finely mutated in order to obtain detailed information such as the size, shape, and nature of the object from echoes of the cluttered forest floor and leaf surface with many irregular reflections (Matsumura, 1999). Pulses with a wide frequency band can clearly analyze terrain and object details and are considered suitable for foraging in messy places and on the surface of the water (Vaughan, et al, 1997). In this study, the echo location call was recorded in the open space, but the frequency band of the long-tailed bat, long-tailed bat, and long-nosed bat pulse is wider than that of other species, and the characteristics of forest (messy place) FM pulses Appeared. These species do not expand the frequency band of the pulse when moving to a messy place, but are thought to inherit genetically a wide frequency band suitable for forests.

  In addition, the frequency band of pulses of the Dobenton bat and the great white bat inhabiting the water surface was relatively wide and approximated. The standard discrimination coefficient between different species of the genus Panthera is high in the ratio of F-max, F-min, duration, and this is mainly because the frequency band of the pulse (F-max-F-min) and duration are the indicators of discrimination. It means to become. The habitat of the genus Giant bat, which was the subject of this study, is in the vicinity of the forest and the water surface, and it can be said that the difference in habitat has become one of the indicators of discrimination. In the same genus and the same habitat, it can be said that duration was an index for distinguishing species in the same habitat (long-tailed bat and long-tailed bat, doventon bat and white-headed bat).

  Furthermore, the index of discrimination between different species was F-min. It was confirmed that bat echo location calls do not synchronize the frequency band to be completely used with other species, and have various unique frequency bands. As described above, since the high-band frequency is attenuated by the atmosphere and there are individual differences in F-max, it can be said that various unique frequency bands are determined by F-min. The discriminant hit rate between different species showed high hit rate for most species, and F-min could be said to be an indicator of discrimination. However, because of the small number of samples, some species may have shown high hit rates, and further investigation is needed.

  As mentioned in the results, the tracking period and final echo location call of the long-tailed bat and the bat are changing the frequency band and duration used, respectively, and increasing the pulse repetition rate compared to the search period echo location call. . When bats follow the bait, they expand the frequency band to gain more object information and clearly increase the repetition of pulses (Parsons et al, 1997). Most bats detect food and enter the follow-up and end stages, shortening the duration to avoid duplication of self-generated pulses and echoed pulses (Kalko, 1995).

  Therefore, the individual difference between the duration of the follow-up period and the end period was smaller than that of the search period. In the long-tailed bat and the long-tailed bat, it was confirmed that the CF portion disappeared in all the FM pulses of the follow-up and final echolocation calls, as described above, in order to obtain more information on the object. This is to expand the frequency band of the pulse. The discriminatory probability between the bat and the long-tailed bat in the exploratory echolocation call was 75.0%, the follow-up echolocation call was 85.4%, and the final echolocation call was 95.8%. It was confirmed that the signal species difference became clear by the action when the voice was emitted. On the other hand, the final echo location call of bats has various frequency band modulations from the follow-up period depending on the size of the bait and the feeding location. Species can be clearly discriminated in the long-tailed bat and the bat, and E-max can be said to be an indicator of discrimination between different species in the final echolocation call.

  In this experiment, the Echo Location Calls produced by the White Bat and the White Bat are similar in structure to each other due to the characteristics of the CF pulse, but the frequency of the emitted pulse does not fluctuate. It is easy to identify the species only by the CF part. Therefore, it can be said that bats that emit CF pulses serve as an index for species discrimination of bats unless the frequency band of the emitted pulses is synchronized with other species.

Acoustic characteristics of social calls The social calls of bats that could be recorded in this experiment could be classified into four types according to the structure of the pulse, and their sonagrams were also diverse. Bats tend to make the social call frequency band lower than the echo location call to avoid tuning the frequency band of the echo location call and social call (Pfalzer & Kusch, 2003). The higher frequency social calls were only 3 out of a total of 224 signals that could be recorded.

  Type A social calls are signals that reach relatively long distances (Lawrence & Simmons, 1982). In this experiment, this social call has been heard far away from the colony, and since it was issued when the roost invaded or captured the bats, the warning signal emitted when the bats felt stressed Conceivable. As for the acoustic characteristics of type A social calls, the pulse structure is similar for all species, but even in the same individual, there is a difference in frequency elements (F-max, F-min, E-max) and duration values. there were. The duration of the pulse may vary depending on the duration of bat capture, and it is thought that the characteristics of the pulse are probably changed depending on how the stress is felt and the duration. For this reason, it may be synchronized with the frequency band of other types of A-type social calls, and the discriminatory probability is considered to have decreased. Many examples have been reported in birds and the like that the signal is unsuitable as an indicator for discriminating species, but that the warning sound itself does not interfere with other species.

  A B-type social call is an audio signal that emits very short pulses continuously for a long time, as described in the results. The situation when a B-type social call is made in this experiment is the same as the situation when an A-type social call is made, when a roost is invaded or a bat is captured. However, there are two types of B-type social calls in the bat, which may have different meanings and functions. As a B-type social call acoustic characteristic, a single signal is formed by multiple pulses. Even in the same individual, the pulse and duration may vary depending on the bat capture duration and the same capture duration, and the signal characteristics are similar to those of type A social calls. As a result of discriminant analysis, the ratio of discriminant is high by E-max and duration, but considering the characteristics of this signal, duration is unsuitable. When only F-max, F-min, and E-max of B-type social call of each bat are discriminated and analyzed, the discriminatory predictive value between different species is 59.5%, and between different species of the giant bat genus is 73.7%, Both decreased by more than 10%. From the above, it is an inappropriate signal as an index for clearly discriminating species.

C-type social call is Phyllostomus
discolor) (Esser & Schmidt, 1989) and the Mexican bat (Tadarida brasiliensis) (Balcombe & McCracken, 1992), which were also emitted during parent-child communication, these signals have the function of isolation calls and direction calls, and the structure of the pulse Is a curve (Fenton, 1985). Bats learn how to feed by accompanying their mothers when the cubs are ready to fly, depending on the species (Brown, 1976). There were six types of C-type social calls that could be recorded in this experiment, one was a bat's parent and child, and the other five were signals from bats flying or from a colony. It is. However, it was September when we were able to record together, and this was the limit of the time when the parent-child communication mentioned above was performed. Since the pulse structure of the C-type social call in this experiment is a curve, these signals may mean parent-child communication.

  A C-type social call is a one-pulse signal, but the structure of the pulse is diverse. The bat C1 social call is similar to the structure of the pulse of a similar echo location call, with a bat that emits in the foraging area near the larvae's roost, with a frequency band approximately 10 kHz lower than the echo location call. It was a signal. According to Russo & Jones (1999), the cool bats (Pipistrellus kuhlii) and the European winged bat (Pipistrellus pipistrellus) of the genus Brassicae have a pulse similar to the C1 type social call of the bat that could be recorded in this experiment. A signal composed of 4 may be emitted while flying at the foraging site, and this signal is said to have a meaning to protect the territory of the foraging site from other individuals.

  Only four signals were able to record the C1 social call of the bat, all of which were one-pulse signals, and could not record a signal consisting of 3-4 pulses. Assuming this to be species specific, the C1 social call of the bat may have a meaning to protect the territory of the feeding area from other individuals. In this experiment, it was confirmed that the pulse structure of the echo location call was similar within the genus. Moreover, although it was not possible to record a C-type social call from two or more species of bats in the same genus, there is a possibility that the structures of pulses in the genus are similar to each other. Brassica bats may emit pulses similar to echolocation calls to protect the territory of the feeding area from other individuals.

  In this experiment, the discriminatory predictability between different types of C-type social calls was 90%, indicating a high predictive value. In these species, E-max can be said to be an index for discriminating species.

  A social call of type D is thought to mean a courtship signal for bats during the mating period (Barlow & Jones, 1997a). In this experiment as well, this type of signal was emitted from a colony of colonies and flying bats only at the time of bat mating. Each type of bat has multiple types of D-type signals, and it is unclear whether each of the multiple types has a different meaning.

The structure of the D-type social call pulse is unique to each type of bat, and it was easy to discriminate between different species using the spectrogram, but the discriminant mid-rate by discriminant analysis was 85.2%, which could be clearly discriminated. The result was not good. This is probably because discriminant analysis was performed without considering the undulation of the pulse. The audio signal of bats can explain the acoustic characteristics with F-max, F-min, E-max, duration, and pulses, but the pulse structure is complicated only for D-type social calls, The acoustic characteristics cannot be explained without taking into account. However, it is very difficult to analyze the undulation of the pulse.
In this study, the discriminatory probability of different types of D-type social calls was 85.2%. It can be said that this signal serves as an index for discriminating species among the long-winged bat, the white-winged bat, and the bat.

The conceptual diagram of the Example of this invention. The block diagram of an ultrasonic signal analyzer. Search location echo location call for each bat species. Echo location calls for the exploration, follow-up, and final stages of the long-tailed bat and the bat.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Bat 2 Ultrasonic receiver 3 Ultrasonic signal analyzer 4 Recording / reproducing apparatus 5 Display apparatus 6 External transmitter 31 Frequency and sonar pattern analysis means 32 Bat kind discrimination means 33 Bat ultrasonic information table

Claims (10)

An ultrasonic receiver for receiving ultrasonic waves emitted by the bats;
An ultrasonic signal analyzer for analyzing the signal from the ultrasonic receiver and discriminating a bat species from the analysis result;
A bat species discrimination device having
The ultrasonic signal analyzing apparatus is characterized in that bat species are determined using bat ultrasonic information including at least one of an upper limit frequency, a lower limit frequency, and an energy peak frequency of an ultrasonic wave emitted by a bat,
Bat species discrimination device. The ultrasonic signal analysis device is characterized by determining a bat species using a search period echo location call emitted by a bat,
The apparatus for distinguishing bat species according to claim 1. The bat ultrasound information further includes at least one of a sonar pattern and a frequency change pattern, wherein the bat species is determined from a combination of a plurality of information.
The bat species discrimination device according to claim 1. The sonar pattern includes at least one of a pulse duration, a pulse interval, and a pulse number of an ultrasonic wave emitted by a bat.
The bat species discrimination device according to claim 3. The ultrasonic signal analyzing apparatus has a bat ultrasonic information table storing a plurality of types of known bat ultrasonic information,
The bat ultrasonic information generated by the bat received by the ultrasonic receiver is compared with the bat ultrasonic information table to determine the bat species.
The apparatus for distinguishing bat species according to claim 1. A recording / reproducing apparatus connectable to the ultrasonic receiver and the ultrasonic signal analyzer;
The recording / reproducing apparatus records an ultrasonic signal from the ultrasonic receiver, and at the time of analyzing the ultrasonic signal, connects to the ultrasonic signal analyzing apparatus to reproduce the recorded ultrasonic signal. Characterized by the
The bat species discrimination device according to claim 1. An ultrasonic wave receiving process for receiving ultrasonic waves emitted by the bats;
Analyzing the received ultrasonic signal and determining the bat species from the analysis results,
A method for distinguishing bat species having
The ultrasonic signal analyzing step is characterized in that bat species are determined using bat ultrasonic information including at least one of an upper limit frequency, a lower limit frequency, and an energy peak frequency of an ultrasonic wave emitted by a bat.
How to distinguish bat species. The ultrasonic signal analysis step is characterized in that a bat species is determined using a search period echo location call emitted by a bat.
The method for distinguishing bat species according to claim 7. The bat ultrasound information further includes at least one of a sonar pattern and a frequency change pattern, wherein the bat species is determined from a combination of a plurality of information.
The method for distinguishing bat species according to claim 7. The sonar pattern includes at least one of a pulse duration, a pulse interval, and a pulse number of an ultrasonic wave emitted by a bat.
The method for distinguishing bat species according to claim 9.

Images