JP2011138310A - Method of determining number and perching position of harmful bird and apparatus therefor, and method of expelling harmful bird - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method of determining the number and perching positions of harmful birds, the processing load for which is significantly reduced by simplifying data processing by means of automatically extracting only necessary and minimum information from image data of the harmful birds in perching. <P>SOLUTION: The method of determining the number and the perching positions of the harmful birds includes: performing erasure processing for erasing moving objects from image data acquired by taking the image of the harmful birds 1 in perching for a given time period to define target data; comparing the target data with previously created reference data for identifying the postures of the harmful birds 1 in perching; and determining the number and the perching positions of the harmful birds 1 in perching by recognizing presence or absence of the postures identical with or similar to the reference data in the target data. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a method and apparatus for determining the number of parked harmful birds and a parked position, and a method for controlling harmful birds.

  In recent years, damage has been caused by eating large amounts of harmful birds, especially crows, and crops are eaten away. In residential and urban areas, pollution caused by noise and feces has become a problem.

  Furthermore, recently, the issue of avian influenza has been emphasized, and the control of harmful birds has become an urgently important issue.

  For this reason, various methods for controlling harmful birds have been proposed. Among them, for example, a monitoring device is provided in the vicinity of a catcher as in JP-A-2004-097019 to confirm the capture status of harmful birds and animals in a remote place. In order to remotely operate the trap, a control method using video data has been proposed.

  Various image processing methods for detecting the number of objects to be imaged from video data and knowing the number of objects to be imaged have been studied. For example, as disclosed in JP-A-10-063863, There have been proposed image processing methods for recognizing pedestrians passing through sidewalks, corridors or doorways, vehicles traveling on the road or passing through gates, and objects moving on a belt conveyor.

JP 2004-097019 A Japanese Patent Laid-Open No. 10-063863

  However, when the above image processing method is used to control harmful birds, for example, when using the template matching method, if the harmful birds move such as changing their posture, the consistency with the template deteriorates and stable recognition is achieved. It becomes difficult.

  In addition, when this template matching method and the background subtraction method are combined, it is possible to recognize a moving object, but harmful birds that do not move in a stationary state are recognized as a background and have a stable knowledge. It becomes difficult.

  In addition, since the inter-frame difference method or the optical flow method recognizes all moving objects, it is difficult to achieve stable recognition under the influence of temporary illumination changes and rain / snow particles.

  For this reason, an image processing method for performing dynamic background update processing that solves the above problems has been proposed.

  In this method, it is possible to stably and accurately know the number of moving objects in the image as well as stationary objects without being affected by disturbances such as rain and snow particles and lighting changes. The data processing process for that is complicated and large-scale.

  By the way, video monitoring of harmful birds is often performed outdoors, and in some cases, it may not be an environment in which power can be constantly supplied.

  Therefore, when the image processing method that performs the dynamic background update process is used for the control of harmful birds, the data processing process is complicated and large-scaled as described above. For example, if the power supply is a battery, the battery will run out quickly and the operating time will be shortened. There is a problem that is difficult.

  The present invention simplifies the data processing process by automatically extracting only the necessary and minimum information from the captured video data of the harmful birds in the stationary state, thereby greatly reducing the data processing load, thereby saving power and The present invention provides a method for determining the number of harmful birds excellent in practicality that realizes a miniaturization and a stopping position.

  The gist of the present invention will be described with reference to the accompanying drawings.

  Criteria that specify the pre-created mode of harmful birds 1 that have been created in advance by subjecting the video data obtained by photographing the harmful birds 1 in the stationary state for a predetermined period of time to the target data by erasing moving objects Comparing the data and the target data, and confirming the presence or absence of the same or approximate form as the reference data in the target data, and determining the number and the stop position of the harmful birds 1 in the stopped state, This is related to the method for determining the number of harmful birds and the stopping position.

  Further, in the method of determining the number of harmful birds and the stopping position according to claim 1, the video data is obtained by photographing the harmful birds 1 stopped on the linear body 2 such as an electric wire or an overhead wire for a predetermined time. And the reference data is also a method for determining the number of harmful birds and the position of the harmful birds, characterized by specifying aspects of the harmful birds 1 that are stopped on the linear body 2 such as an electric wire or an overhead wire. It is concerned.

  Further, in the method for determining the number of harmful birds and the stopping position according to any one of claims 1 and 2, position data is recorded in the target data and the reference data. This relates to the method for determining the number and the stop position.

  Further, in the method for determining the number of harmful birds and the stopping position according to any one of claims 1 to 3, unnecessary information other than the harmful birds 1 in the stopped state in the target data subjected to the erasure process. The present invention relates to a method for determining the number of harmful birds and a stopping position, characterized by performing disturbance suppression processing for erasing or reducing data.

  Further, in the method for determining the number of harmful birds and the stop position according to any one of claims 1 to 4, the erasure processing is performed by reducing the video data obtained by photographing the stop harmful bird 1 for a predetermined time. The present invention relates to a method for determining the number of harmful birds and a stopping position, which is a low-pass filtering process for extracting a frequency component.

  Further, in the method for determining the number of harmful birds and the stop position according to any one of claims 4 and 5, the disturbance suppression process is a low-frequency component that extracts a low-frequency component from the target data subjected to the erasure process. The number and retention of harmful birds, characterized by any one of pass filtering, band-pass filtering that extracts a predetermined frequency component, or high-pass filtering that extracts a high frequency component, or a combination of these processes This is related to the position determination method.

  Further, in the method for determining the number of harmful birds and the stop position according to any one of claims 1 to 6, the brightness of the video data before the erasure process or the target data after the erasure process is manipulated, The present invention relates to a method for determining the number of harmful birds and a stationary position, characterized in that a luminance conversion process is performed to clarify the stationary harmful birds 1 in the video data or the target data.

  Further, in the method for determining the number of harmful birds and the stop position according to any one of claims 1 to 7, the preliminarily created state of the harmful bird 1 in the stopped state is the photographing of the harmful bird 1 in the stopped state. The method relates to a method for determining the number of harmful birds and the stopping position, characterized in that the direction is front or back and at least one type, and when the shooting direction is diagonal, the type is at least two. is there.

  Further, in the method for determining the number of harmful birds and the stop position according to any one of claims 1 to 8, the harmful bird 1 is a crow. It is concerned.

  Moreover, it is a method of controlling the harmful birds 1, and the number of harmful birds according to any one of claims 1 to 9 and the number of the harmful birds 1 in the stopped state determined using the method for determining the stopping position. The present invention relates to a method for controlling harmful birds, characterized in that it is determined whether or not to attack the harmful bird 1 in the stationary state, and the attack is executed.

  The method for controlling harmful birds according to claim 10, wherein an optimum attacking means is selected when executing the attack.

  Further, in the method for controlling harmful birds according to any one of claims 10 and 11, the number of harmful birds 1 in the same position after the attack is determined by the same method, and the effect of the attack is determined. This relates to a method for controlling harmful birds characterized by confirmation.

  The method for controlling harmful birds according to claim 12, wherein after the effect of the attack is confirmed, if it is determined that the effect is insufficient, the attack is performed again. .

  In addition, an imaging unit 3 that captures the harmful birds 1 in a stationary state for a predetermined time, an erasing processing unit 4 that erases moving objects from video data obtained by imaging with the imaging unit 3, and an erasing processing unit 4 The mode of the stationary harmful bird 1 in the processed target data is compared with the reference data that specifies the mode of the stationary harmful bird 1 prepared in advance, and is the same as the reference data in the target data or It is related to a device for determining the number of harmful birds and a stop position, comprising a determination unit 5 for determining the number of stop harmful birds 1 and the stop position by knowing whether there is an approximate mode. is there.

  Since the present invention deletes moving objects such as flying harmful birds and rain / snow particles from the video data obtained by capturing the harmful harmful birds in the stopped state, and only uses the minimum necessary information. In addition, it uses a very simple means to determine the number of stationary harmful birds and their location by collating with reference data created in advance. This is a method for determining the number of harmful birds and the stopping position that can be realized with this method.

It is explanatory summary figure which shows a present Example. It is explanatory summary figure which shows a present Example.

  An embodiment of the present invention which is considered to be suitable will be briefly described with reference to the drawings showing the operation of the present invention.

  Since the present invention deletes unnecessary data such as flying harmful birds 1 and rain snow particles from the video data obtained by photographing the stationary harmful birds 1 for a predetermined time, the target data is also obtained. Since the harmful birds 1 that are stopped are harmful birds 1 that are stopped on a linear body 2 such as an electric wire or an overhead wire, the captured video data is attacked even under a poor shooting environment. Since it becomes the minimum necessary stable target data for determining the necessity of this, and this target data is the minimum necessary data, the process of ascertaining the number and the location of the harmful birds 1 that are in the stationary state, This is performed by an extremely simple operation of collating the mode of the stationary harmful bird 1 in the target data with the mode of the stationary harmful bird 1 prepared in advance.

  That is, the present invention is necessary and sufficient for judging the necessity of an attack on video data in which harmful birds 1 stopped on the linear body 2 and flying harmful birds 1 are mixed. Since the number of harmful birds 1 is determined based on the narrowed-down data and the number of harmful birds 1 that are parked on the linear body 2, it is stable even by simple means of collation with reference data. In addition, accurate information can be obtained.

  For example, when shooting harmful birds 1 that are parked on the linear body 2 from an oblique direction, the two types of modes in which the harmful birds 1 that are stationary in the video are facing right or left are shown. Therefore, a simple process in which the two types of harmful birds 1 are prepared in advance as reference data and the reference data is compared with the stopped harmful birds 1 in the target data. Thus, it is possible to determine the number and the stopping position of the harmful birds 1 in the stopped state by recognizing the presence or absence of the same or approximate mode as the reference data. That is, the stop position of the harmful birds 1 can be ascertained from the coordinates in which the same or approximate form exists in the target data, and the number of the harmful birds 1 can be ascertained from the number of the coordinates, thereby simplifying the video data processing process. As a result, the processing load is greatly reduced.

  Since the conventional method tries to calculate the number of harmful birds 1 accurately regardless of the state of the harmful birds 1, the scale of arithmetic processing becomes large and complicated, and the calculation becomes inaccurate when occlusion occurs.

  In this respect, since the present invention targets the harmful birds 1 that are stopped on the linear body 2, no occlusion occurs, and even if occlusion occurs in the flying harmful birds 1, it moves. It will not cause any trouble because it erases the things that are on the floor.

  A specific embodiment of the present invention will be described with reference to FIGS.

  The present embodiment is a case where the present invention is used in a crow control apparatus as a harmful bird 1, and includes digital video data obtained by photographing a stationary crow (including the case where the video data is digitally processed after photographing). By inputting, the number of stopped crows and the individual stopping positions are determined without being affected by disturbance even under poor lighting conditions. It is a pest control device for harmful birds that determines whether or not to attack a crow based on the number of, and selects an optimal attack means when the attack is determined.

  Further, in the present embodiment, as shown in FIG. 1, an imaging unit 3 that captures a crow in a stationary state for a predetermined period of time, and rain and snow particles and moving crows are captured from video data obtained by imaging with the imaging unit 3. The erasure processing unit 4 to be erased is collated with the crow state in the stationary state in the target data processed by the erasure processing unit 4 and the reference data specifying the state of the crow in the stationary state created in advance. Confirming the presence or absence of an aspect that is the same as or similar to the reference data in the center, and confirming the number of stationary crows and the stopping position of each stationary crow. An output unit for output), an attack determination unit 6 for determining whether or not an attack is necessary based on the determination result output from the determination unit 5, and an optimal attack when it is determined that an attack is necessary Attack means selection unit 7 for selecting In is configured.

  Specifically, the imaging unit 3 is a video camera, and the video camera is disposed in a fixed state at a predetermined position, for example, a crow that is stopped at a place where noise or feces are frequently generated, , A video image is obtained by photographing a crow stopped on a linear body 2 (including trees, narrow concrete blocks, etc.) such as an electric wire or an overhead wire for a predetermined time on a fixed screen.

  The video camera may be a digital video camera or an analog video camera. However, when an analog video camera is used, the video data is converted into a digital signal by AD conversion.

  In addition, when shooting a dark bird such as the crow of this embodiment, a video camera is arranged so that the background is as bright as possible so that the difference between the crow and the background becomes as clear as possible ( By acquiring the video (so that the crow is emphasized as much as possible), the accuracy of the data processing in the erasure processing unit 4 and the recognition processing in the determination unit 5, which will be described later, is improved, and the more accurate number and position of the crows in the stationary state Can be determined.

  When shooting is not possible in the above situation, for example, in the evening, or when the background building is dark and the space in the video is dark overall, for example, among the pixels in the video data, Contrast stretching, which converts the lowest luminance to black, and conversely the highest luminance to white to emphasize the difference between light and dark, and the color data input signal and the actual output image signal By relatively adjusting gamma correction, and when the video data is grayscale, by flattening the histogram of the brightness of the original image, and applying brightness conversion processing such as histogram flattening to make the image clearer, More stable video data can be obtained.

  In this luminance conversion process, for example, when the information to be extracted is a crow, optimal luminance value amplification is appropriately performed, for example, the luminance value is doubled to four times.

  Further, the maximum value of the luminance value is limited to 255, and when the maximum value is exceeded by the amplification of the luminance value, this luminance value is set to the maximum to brighten the background in the video data. In this case, since the black crow does not change so much, the difference in brightness between the crow and the background becomes large, so that the accuracy of identifying the crow in the determination unit 5 is improved.

  Further, for example, after increasing the difference between light and darkness of the crow and the background as described above, it is possible to further improve the accuracy of recognition by further reversing the black and white.

  Since this luminance conversion processing has a large load of calculation processing of video data, it takes a very long time to process the video data captured by the video camera as it is. Therefore, it is preferable to perform the processing after the video data is processed by the erasure processing unit 4.

  In addition, when the obtained video data is composed of color signals such as red, green, and blue, for example, only the luminance signal may be used, and this color signal may be represented by, for example, YUV, CMY. , HSV may be used for color conversion processing.

  If all the crows in the video data acquired as described above or the video data acquired as described above are subjected to luminance conversion processing or color conversion processing are in a stationary state and do not move. It is possible to easily identify the crow that is stationary, but depending on the video data, there may be a crow in flight, or there may be a crow that is stationary but with movement that causes its wings to flutter. As a result, the accuracy of recognition decreases.

  For this reason, an erasure processing unit 4 is provided for erasing such moving objects from the video data.

  The video data photographed by the video camera of this embodiment is, for example, a collection of 15 to 30 still images per second, and the still images are composed of pixels arranged in the vertical direction and the horizontal direction (position in the spatial direction). It is an aggregate.

  The erasure processing unit 4 multiplies (weights) each of the plurality of consecutive still images with a different filter coefficient, adds all the still images multiplied with the different filter coefficients, and then adds the same. A process of dividing a still image by adding the pixel values at spatial positions into a single still image, that is, a process of erasing moving objects by performing a temporal low-pass filtering process is performed. .

  Note that the filter coefficient used here is appropriately determined so that information data to be detected in the video data is emphasized and other disturbances are suppressed.

  This low-pass filtering process is expressed by a mathematical expression, where t is the number of still images, x (t, n, m) is a pixel value in position (n, m) or x rows and y columns, and w (t) is a filter coefficient. Then, the following equation is obtained.

  If the filter coefficient is 1 / T in the above equation, this low-pass filtering process is a synchronous addition process in which each still image is simply added and divided by the added number of still images.

  By performing temporal low-pass filtering on the video data acquired by the erasure processing unit 4 in this way, flying crows, rain / snow particles in the video data, or flapping wings of stationary crows, etc. , It is possible to make a single still image from which moving objects are erased.

  In addition, in order to avoid the overflow of the addition result when the above-described still images are added, the pixel values of the still images may be added in advance after being divided by the number of still images.

  In addition, in order to further improve the accuracy when recognizing a stationary crow from target data obtained by performing such processing (one still image from which moving objects are deleted), the target obtained It is preferable to perform disturbance suppression processing to suppress disturbances other than moving objects from the data (for example, background data and noise) or to emphasize more crows.

  Specifically, this disturbance suppression processing is low for, for example, the target data, that is, the vertical direction and the horizontal direction of pixels arranged in the vertical direction and the horizontal direction (position in the spatial direction), or both directions. A process that blurs disturbances such as background noise by applying band-pass filtering, a process that emphasizes the outline of the image by applying band-pass filtering, and a high-pass filtering and binarization process. For example, the process of enhancing the boundary between dark and light parts, that is, the outline part, and especially high-pass filtering and binary to enhance the outline of dark parts such as crows. It is suitable to apply a crystallization treatment.

  By providing a disturbance suppression processing unit that performs such disturbance processing, the spatial portion of the video data is suppressed, and only the stationary crow is emphasized. It is possible to improve the accuracy of knowing the number of stopped crows and the stopping position, and the configuration of the disturbance suppression processing unit is any of the above-described low-pass filtering process, band-pass filtering process, and high-pass filtering process. Alternatively, a configuration in which these are combined may be appropriately determined depending on the background environment, the illuminance state, and the like in which the crow in the stationary state is photographed.

  As described above, the video data captured by the video camera is processed by the erasure processing unit 4 (the luminance conversion process, the color conversion process, the disturbance suppression process, or a combination of these processes is also performed as necessary) Get the data.

  The target data obtained in this way has two-dimensional spatial coordinates.

  This target data is the minimum necessary information for ascertaining and outputting the number of crows in the stopped state and the stopping position.

  In addition, the target data may be subjected to a binarization process that further simplifies the confirmation process in the determination unit 5, and when the target data includes a lot of noise, the above-described two processes are performed to suppress the noise. A morphological filter, a median filter, or the like, such as dilation processing or degeneration processing, may be applied to the pixel values that have been digitized, or a combination processing, labeling processing, or the like may be combined.

  In this way, it is desirable to improve the accuracy of the recognition processing by simplifying the target data or removing noise.

  Note that a linear discriminating method or the like is generally used as a threshold value when the above binarization processing is performed.

  In addition, the present embodiment includes a determination unit 5 that knows the number of stopped crows and the stop position from target data and outputs a determination result.

  The determining unit 5 collates the reference data of the crow mode in the stationary state prepared in advance with the crow mode in the target data obtained by performing the above-described processing, and is the same as or approximate to the reference data. The presence / absence is confirmed, the number of stopped crows and the stop position are determined, and the determination result is output.

  Specifically, when a stationary crow is photographed from the position where the video camera is placed, it is confirmed in advance how the stationary crow appears, and the crow having the same aspect as this is captured. Create reference data that identifies

  In other words, when the crow stops on the electric wire, it stops either in the front direction or in the back direction. For example, if the video camera's shooting direction is oblique to the stopped crow, The crow in the stopped state of the video is either right-facing or left-facing. Accordingly, two types of reference data are prepared: a right-handed crow and a left-handed crow.

  In addition, for example, when the shooting direction of the video camera is from the front or the back with respect to the crow that is in the stationary state, the crow in the stationary state of the captured image is reflected only in the front direction or the back direction, and this Since the front-facing or rear-facing aspect appears almost the same two-dimensionally, it is only necessary to prepare only one type of reference data for the front-facing crow.

  In this way, the reference data is created in advance, and all the spatial coordinates in the target data are collated with the reference data, and the stationary state is obtained from the coordinates in which the same or approximate form exists in the target data. The stop position of the crow can be ascertained, and the number of stop crows can be ascertained by counting the coordinates of the stop positions of the known crows. And the stop position can be confirmed, and the confirmed result is output.

  If there is any doubt about the number of crows that are parked and the final result of the parked position (for example, if there is no effect or no attack is seen), the function to send the target data is In addition, the target data may be monitored at a remote place.

  As in this embodiment, when the number of reference data is small, the stability of clustering is increased, the accuracy is further improved, and the number of stopped crows and the stop position can be ascertained more accurately. .

  Also, in this collation process, it is determined that the similarity is equal to or greater than a preset threshold value and is counted as one bird.

  As a method for determining this threshold value, for example, an automatic threshold value determining method for determining the threshold value from the gray value of the image or the shape of the histogram such as a mode method, a differential histogram method, or a discriminant analysis method may be used. The similarity between the target data and the reference data is calculated.

  As a calculation method of the similarity, for example, a general mutual calculation method, an absolute value difference method, a phase-only correlation method, or the like is used.

  In addition, background difference processing may be performed before the above-described similarity calculation. When this background disposal processing is performed, it is desirable to dynamically update the background appropriately.

  In this way, the video data obtained by photographing the stationary crow with the video camera is erased from the moving object by the erasure processing unit 4, and is subjected to the color conversion processing, the luminance conversion processing, or the disturbance reduction processing as necessary, and is stopped. The accuracy of collation with reference data in the determination unit 5 is improved, and the number of stopped crows and their stopping positions are determined more accurately and stably. Is output to the attack determination unit 6 to determine whether or not there is a need for an attack for crow control.

  The attack determination unit 6 performs a process of determining whether or not the crow stop state of the photographed video is an attacking situation. If the number of crows in the stop state is a predetermined number or more, an attack is performed. Judgment is made to determine whether or not the attack is possible by setting a threshold value such that if the predetermined number is not reached, it is determined that the attack is not performed.

  Specifically, for example, when the threshold value in the attack determination unit 6 is set to 10, when the number of stationary crows output from the determination unit 5 is eight, it is determined that no attack is performed, and the determination unit 5 When the number of stationary crows outputted by is 10 or more, a process of determining to attack is performed.

  The threshold value can be dynamically updated with reference to past attack means and a history of crow's reaction when this attack is made.

  In this way, when the result of determining whether or not to attack by the attack determination unit 6 is sent to the attack means selection unit 7 and the result of attacking the attack means selection unit 7 is sent The attack means selection unit 7 appropriately performs processing for selecting effective attack means.

  Specifically, the attack means selection unit 7 refers to a history of crow reaction during past attacks accumulated as a database and performs a process of selecting an optimum attack means.

  The above reaction history is, for example, in what kind of environment (weather, time, ambient illumination, etc.), what kind of attack was effective, or what kind of attack , A database that records the situation / results of past attacks, such as how many seconds after the attack the number of crows stopped or how many seconds later it increased It is.

  Also, simply comparing the number of parked crows before and after the attack, it is not possible to determine whether the number of crows after the attack was originally a crow or a newly parked crow. Therefore, an accurate effect cannot be judged.

  In this regard, in this embodiment, the information sent from the determination unit 5 is not only the number of stopped crows, but also the information on the stopping positions of individual crows. Record the confirmed result once the position is confirmed, and after the attack, acquire the video again and apply the same processing to the acquired video as before the attack to determine the number of stopped crows and the stop position after the attack. By obtaining a confirmed result and comparing it with information on the crow's stopping position before and after the attack, it is possible to see how many crows in the stopped state before the attack flew away by the attack and how many newly stopped after the attack. It is possible to discriminate and record a more accurate effect in the database.

  Therefore, the present optimal attack means can be selected from the history of this reaction. For example, sound and sound waves are more effective when the surroundings are bright in the information when a crow was attacked in the past, and dark. In some cases, the laser beam was effective, and the most effective attack method was selected depending on the shooting environment at the time of the attack.For example, the crow became accustomed to the attack because the laser beam attack continued as an attack method. If there is a history of reactions that can be judged to have occurred (a situation in which the number of crows that flew gradually has decreased), it is determined that the crows are not yet used rather than selecting an attack method using a laser beam again. The attack means selection unit 7 selects an optimum attack means as appropriate based on the current situation and beta-based information, such as selecting an attack means using a sound wave to be performed. Select Oite, and configured to control more effectively crow.

  Further, for example, the attack means may be configured to select an attack means having directivity such as a laser beam or a highly directional sound wave generator. That is, it is effective if the crow is attacked pinpointed based on the stop position information of each crow obtained by the collation processing.

  FIG. 2 shows a specific configuration in which the number of stationary crows and the stationary position can be determined more accurately and stably by performing the above-described color conversion processing and luminance conversion.

  In the case of FIG. 2, the video data from the video camera is subjected to color conversion processing, and the color conversion processed video data is subjected to erasure processing in the erasure processing unit 4 to erase moving objects in the video data. The target data obtained by performing the erasure process is further subjected to a luminance conversion process and a disturbance suppression process, the target data is subjected to a binarization process, and a morphology filter process such as an expansion process or a degeneration process. And then subjecting it to median filter processing and labeling, the target data in which the stationary crows are emphasized is obtained, and this target data is compared with the reference data to determine the number of stationary crows and the stationary position. It will be determined accurately.

  Since the present embodiment is configured as described above, it has the following operational effects.

  A stationary crow is photographed with a video camera, the moving object is deleted from the photographed image to obtain target data, and reference data specifying the state of the stationary crow is created in advance. Simply by comparing the state of crows in the target data with the crow state in the target data, the number of stop crows and the stop position are determined from the target data, and the determination result is extracted. However, the processing load can be greatly reduced.

  In the present embodiment, video data in which crows stopped on the linear body 2 and flying crows are mixed is necessary on the linear body 2 that is necessary and sufficient to determine the necessity of the attack. Since the number of crows is narrowed down to the number of crows and the number of crows is ascertained based on the narrowed data, stable and accurate information can be obtained even by simple means of collation with reference data. .

  In addition, since the conventional method tries to calculate the number of crows accurately regardless of the state of the crow, the scale of the arithmetic processing becomes large and complicated, and when occlusion occurs, the calculation is inaccurate.

  In this respect, since the present embodiment is intended for a crow stopped on the linear body 2, no occlusion occurs, and even if an occlusion occurs in a flying crow, the moving object This will not cause any trouble.

  As described above, this embodiment can improve and stabilize the number of parked crows and the accuracy of the parked position, so that the control apparatus can be realized with a power-saving circuit.

  Therefore, the overall configuration of the control device can be reduced in size, and can be easily installed regardless of the installation location of the device. If it is installed in various crows and roosts to control crows, residential areas In addition, it is possible to reduce noise and excrement damage in urban areas, and to provide a groundbreaking harmful bird control apparatus with excellent practicality that can reduce crop damage on farmland.

  Note that the present invention is not limited to this embodiment, and the specific configuration of each component can be designed as appropriate.

DESCRIPTION OF SYMBOLS 1 Harmful birds 2 Linear body 3 Imaging part 4 Erasing process part 5 Determination part

Claims (14)

  Reference data specifying pre-created types of stationary harmful birds that have been created by subjecting video data obtained by photographing stationary harmful birds for a predetermined period of time to target data by erasing moving objects Harmful birds characterized by collating with the target data and determining the number and positions of the harmful birds in the stationary state by recognizing the presence or absence of the same or similar mode as the reference data in the target data How to determine the number of stops and the stop position.   The method of determining the number of harmful birds and the stopping position according to claim 1, wherein the video data is obtained by photographing a harmful bird stopped on a linear object such as an electric wire or an overhead wire for a predetermined time, and The method of determining the number of harmful birds and the position of the retention, characterized in that the data also specify aspects of harmful birds that are stopped on a linear object such as an electric wire or an overhead wire.   The number of harmful birds according to any one of claims 1 and 2, wherein position data is recorded in the target data and the reference data. How to determine the stop position.   The number of harmful birds according to any one of claims 1 to 3, and a method for determining a stopping position, wherein unnecessary information data other than the harmful birds in the stationary state are erased in the target data subjected to the erasing process. Alternatively, a method for determining the number of harmful birds and a stop position, wherein a disturbance suppression process is performed.   5. The method for determining the number of harmful birds and the stop position according to claim 1, wherein the erasing process includes a low-frequency component from the video data obtained by photographing the stopped harmful birds for a predetermined time. A method for determining the number of harmful birds and a stop position, characterized by low-pass filtering to be extracted.   6. The method of determining the number of harmful birds and the stop position according to claim 4, wherein the disturbance suppression processing is a low-pass filtering that extracts a low-frequency component from the target data subjected to the erasure processing. The number of harmful birds and the position of the stop, characterized by any one of processing, band-pass filtering processing for extracting a predetermined frequency component, high-pass filtering processing for extracting a high frequency component, or a combination of these processing Confirmation method.   7. The method of determining the number of harmful birds and a stop position according to claim 1, wherein the video data is operated by operating a brightness of the video data before the erasure process or the target data after the erasure process. Alternatively, a method for determining the number of harmful birds and a stationary position, wherein brightness conversion processing is performed to clarify the stationary harmful birds in the target data.   The number of harmful birds according to any one of claims 1 to 7 and a method for determining a stop position, wherein the pre-prepared state of the harmful birds is a front view or a shooting direction of the stopped harmful birds is front or A method for determining the number of harmful birds and a stopping position, characterized in that at least one type of mode is used for the back side and at least two types of mode is used when the shooting direction is oblique.   9. The method for determining the number of harmful birds and the stopping position according to any one of claims 1 to 8, wherein the harmful birds are crows.   A method for controlling harmful birds, wherein the number of harmful birds according to any one of claims 1 to 9 and the number of harmful birds in the stationary state determined using the method for determining a stationary position, A method for controlling harmful birds, characterized by determining whether or not to attack the harmful birds and executing the attack.   11. The method for controlling harmful birds according to claim 10, wherein an optimal attack means is selected when executing the attack.   12. The method for controlling harmful birds according to any one of claims 10 and 11, wherein the number of harmful birds in the same position after the attack is determined by the same method, and the effect of the attack is confirmed. A method for controlling harmful birds characterized by   13. The method for controlling harmful birds according to claim 12, wherein after the effect of the attack is confirmed, if it is determined that the effect is insufficient, the attack is executed again.   An imaging unit that photographs a stationary harmful bird for a predetermined time, an erasing processing unit that erases moving objects from video data obtained by photographing with the imaging unit, and the target data in the target data processed by the erasing processing unit Matching the state of the harmful bird in the stationary state with the reference data specifying the state of the harmful bird in the stationary state prepared in advance, and confirming whether there is an aspect that is the same or similar to the reference data in the target data An apparatus for determining the number of harmful birds and a stop position, comprising: a determination unit that determines the number of stop harmful birds and a stop position.

Images