JP2018054455A - Cloud position estimation device, cloud position estimation method and cloud position estimation program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suitably estimate a cloud position from an image obtained by an optical sensor.SOLUTION: An aircraft 1 comprises: an optical sensor 14 which is directed in a sensor direction crossing a vertical direction, for obtaining an image including cloud CL; and a control unit 18. The control unit 18 calculates a range of the cloud CL in the respective horizontal and vertical directions in the image obtained by the optical sensor 14, estimates presence probability distribution of the cloud CL with respect to a separation distance from the optical sensor 14, using a relation between altitude and likelihood of occurrence, on the basis of the calculated range of the cloud CL in the vertical direction in the image, altitude AL of the aircraft and altitude R of the possible presence of cloud, and create a cloud distribution map M in a flat surface including the horizontal direction and the sensor direction in the image on the basis of the range of the cloud CL in the horizontal direction and the presence probability distribution of the cloud CL with respect to a separation distance from the optical sensor 14.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a technique for estimating a cloud position using an image acquired by an optical sensor.

  When an image of a predetermined target is acquired using an optical sensor installed at a high place, such as when monitoring with an optical sensor mounted on an aircraft, a cloud positioned between the target and the optical sensor Is an obstacle to image acquisition. For this reason, it is necessary to grasp the position of the cloud as accurately as possible so that the image can be acquired favorably while avoiding the cloud.

  It is common to use a weather radar to detect the position of the cloud, but it is difficult to mount on a small aircraft, etc. in terms of weight and power consumption, and the position of the weather radar is used to emit radio waves. There is a problem that it may be recognized by others.

  On the other hand, for example, in the technique described in Patent Document 1, a cloud is photographed together with a reference object such as a mountain or a building, and the position of the cloud is estimated from the height and position of the reference object on the obtained image. It is possible to grasp the cloud position without using weather radar.

JP, 2015-11014, A

  However, since the technique described in Patent Document 1 requires the presence of a reference object for estimating the position of the cloud, it cannot be applied, for example, when acquiring an image on the ocean or at a high altitude where only the imaging target and the cloud exist. .

  The present invention has been made in view of the above problems, and an object of the present invention is to suitably estimate a cloud position from an image acquired by an optical sensor without requiring a reference object for estimating the cloud position. To do.

In order to achieve the above object, the invention described in claim 1 is a cloud position estimation device that estimates the position of a cloud using an optical sensor installed at a high place,
The optical sensor is oriented in a sensor direction that intersects the vertical direction and acquires an image including a cloud;
Cloud range calculation means for calculating a cloud range in each of horizontal and vertical directions on the image based on the image acquired by the optical sensor;
Based on the vertical range of the cloud on the image calculated by the cloud range calculation means, the altitude of the optical sensor, and the possible altitude of the cloud, using the relationship between altitude and the likelihood of cloud generation, Cloud distribution estimation means for estimating a cloud existence probability distribution with respect to a separation distance from the optical sensor;
Based on the horizontal range of the cloud on the image calculated by the cloud range calculation means and the cloud existence probability distribution with respect to the separation distance from the optical sensor estimated by the cloud distribution estimation means. A cloud distribution generation means for generating a cloud existence probability distribution in a plane including the horizontal direction above and the sensor direction;
It is characterized by providing.

The invention according to claim 2 is the cloud position estimation apparatus according to claim 1,
The optical sensor is mounted on a moving body, acquires a plurality of images having different acquisition times and including the same cloud,
The cloud range calculation means, the cloud distribution estimation means, and the cloud distribution generation means generate cloud existence probability distributions in the plurality of planes corresponding to the plurality of images,
The present invention is characterized by comprising cloud distribution integration means for weighting and integrating the generated existence probability distributions of the plurality of clouds so that the weight coefficients corresponding to newer images become larger.

The invention described in claim 3 is the cloud position estimation apparatus according to claim 1 or 2,
The cloud distribution estimation means includes
Based on the elevation angle of the optical sensor, the vertical range of the cloud on the image, and the altitude of the optical sensor, a straight line connecting the optical sensor and the highest altitude point of the cloud is calculated,
As a range in which a cloud can exist in a vertical section including the straight line, a cloud existence range surrounded by the straight line and the altitude where the cloud can exist in consideration of the roundness of the earth is calculated,
Calculating the cloud existence probability distribution with respect to the separation distance from the optical sensor in the cloud existence range, assuming that the lower the altitude of the position on the straight line in the cloud existence range, the higher the cloud existence probability. It is characterized by.

The invention according to claim 4 is the cloud position estimation apparatus according to any one of claims 1 to 3,
The cloud position estimation device is characterized in that, when an image of a predetermined target is acquired by the optical sensor, the position of a cloud included in the image is estimated.

Invention of Claim 5 is the cloud position estimation apparatus as described in any one of Claims 1-4.
The optical sensor is mounted on an aircraft.

  The invention described in claim 6 and claim 7 is a cloud position estimation method and a cloud position estimation program having the same features as the cloud position estimation apparatus according to claim 1.

According to the first aspect of the present invention, the cloud range in each of the horizontal and vertical directions on the image acquired by the optical sensor is calculated, the vertical range of the cloud on the calculated image, the optical sensor Based on the altitude and the altitude at which clouds can exist, the existence probability distribution of the clouds with respect to the separation distance from the optical sensor is estimated using the relationship between the altitude and the likelihood of cloud generation. Based on the cloud existence probability distribution with respect to the separation distance from the optical sensor and the horizontal range of the cloud on the image, the presence of the cloud in a plane including the horizontal direction and the sensor direction on the image. A probability distribution is generated.
Therefore, unlike the conventional case where the presence of a reference object for estimating the position of the cloud is essential, the position of the cloud is preferably estimated from the image acquired by the optical sensor without the need for such a reference object. be able to.

According to the second aspect of the present invention, a plurality of images having different acquisition times and including the same cloud are acquired by the optical sensor that moves together with the moving body, and the plane is corresponding to the plurality of images. A plurality of cloud existence probability distributions are generated. Then, the generated existence probability distributions of the plurality of clouds are integrated while being weighted so that the weighting coefficient becomes larger as it corresponds to a newer image.
Therefore, since the cloud existence probability distribution based on new images acquired from a different direction from the past is added to the past as more accurate information, the cloud position can be estimated more accurately. it can.

It is a block diagram which shows schematic structure of the aircraft in this embodiment. It is a flowchart which shows the flow of a cloud position estimation process. It is a figure for demonstrating a cloud position estimation process. It is a figure for demonstrating a cloud position estimation process.

  Hereinafter, an embodiment when the cloud position estimation apparatus according to the present invention is applied to an aircraft 1 will be described with reference to the drawings.

[Constitution]
First, the configuration of the aircraft 1 in the present embodiment will be described with reference to FIG.
FIG. 1 is a block diagram showing a schematic configuration of the aircraft 1.
The aircraft 1 uses a mounted optical sensor 14 to monitor a predetermined monitoring target. More specifically, the aircraft 1 includes the cloud position estimation device according to the present invention, thereby preventing the acquisition of an image of the monitoring target. It is configured to be able to monitor the monitoring object suitably while estimating the position of the obtained cloud and avoiding this.
Specifically, as shown in FIG. 1, the aircraft 1 includes a flight mechanism 11, an operation unit 12, an airframe sensor 13, an optical sensor 14, a storage unit 16, and a control unit 18. ing.

Among these, the flight mechanism 11 is a mechanism for flying the aircraft 1 and includes an internal combustion engine (for example, a jet engine) that generates a propulsive force, an actuator for driving a control surface, and the like.
The operation unit 12 includes a control stick operated by a pilot, various operation keys, and the like, and outputs signals corresponding to operation states of the control stick and various operation keys to the control unit 18.

  The body sensor 13 is a variety of sensors for detecting the position and flight state of the aircraft 1, and includes a radar, a gyro sensor, a speed sensor, a GPS (Global Positioning System), and the like. These body sensors 13 obtain various information based on a control command from the control unit 18 and output the signals to the control unit 18.

  The optical sensor 14 is for monitoring a monitoring target, and is provided on the body of the aircraft 1 so as to be mounted on a gimbal (not shown) and to change the direction. The optical sensor 14 acquires an image outside the aircraft 1 and outputs the acquired image information to the control unit 18.

The storage unit 16 is a memory that stores programs and data for realizing various functions of the aircraft 1 and also functions as a work area. In the present embodiment, the storage unit 16 stores a cloud position estimation program 160.
The cloud position estimation program 160 is a program for causing the control unit 18 to execute a later-described cloud position estimation process.

  The control unit 18 centrally controls each unit of the aircraft 1. Specifically, the control unit 18 controls the flight of the flight mechanism 11 to control the flight of the aircraft 1, controls the operations of the airframe sensor 13 and the optical sensor 14, and is stored in the storage unit 16. The program is expanded, and various processes are executed in cooperation with the expanded program.

[Operation]
Next, the operation of the aircraft 1 when executing the cloud position estimation process will be described.
FIG. 2 is a flowchart showing the flow of the cloud position estimation process, and FIGS. 3 and 4 are diagrams for explaining the cloud position estimation process.

In the cloud position estimation process, when the monitoring target is monitored by the optical sensor 14, the cloud is interposed between the own aircraft (aircraft 1) and the monitoring target so that image acquisition of the monitoring target is not hindered. It is a process which estimates the position of. This cloud position estimation process is executed when the control unit 18 reads out and expands the cloud position estimation program 160 from the storage unit 16 when, for example, an instruction to perform the cloud position estimation process is input by a pilot operation or the like. The
Here, it is assumed that the aircraft 1 is flying and the monitoring target T, which is a flying object, is monitored from a distance in the horizontal direction (see FIG. 3A).

As shown in FIG. 2, when the cloud position estimation process is executed, the control unit 18 first acquires an image of the monitoring target T by the optical sensor 14 (step S1).
Specifically, the control unit 18 captures the monitoring target T on the sensor visual axis Ax of the optical sensor 14 oriented substantially in the horizontal direction (see FIG. 4A), and the monitoring target T is detected by the optical sensor 14. Get the image.
Hereinafter, a direction along the direction of the optical sensor 14 (substantially horizontal direction in the present embodiment) is referred to as a “sensor direction”.

Next, the control unit 18 determines whether or not there is a cloud that may hinder the acquisition of the image of the monitoring target T on the image acquired in step S1 (step S2). S2; No), the process proceeds to the above-described step S1, and the image acquisition of the monitoring target T is continued.
Note that the condition of “clouds that may hinder the acquisition of the image of the monitoring target T” is not particularly limited, and may be, for example, a cloud located at the same altitude as the monitoring target T, or all the clouds on the image, etc. Also good.

In Step S2, when it is determined that there is a cloud that may hinder the acquisition of the image of the monitoring target T on the image acquired in Step S1 (Step S2; Yes), the control unit 18 performs the following operation based on this image. The cloud range in the horizontal and vertical directions on the image is calculated (step S3).
Specifically, as illustrated in FIG. 3A, the control unit 18 displays a cloud CL in the vicinity of the horizontal line HL (above the horizontal line HL in FIG. 3A) from the acquired image. Extraction is performed as an obstacle to acquisition, and the horizontal range and vertical range of the cloud CL on the image are calculated.

Next, the control unit 18 estimates a cloud existence probability distribution with respect to the separation distance from the optical sensor 14 (step S4). In this step, the distance from the aircraft 1 to the cloud CL, which cannot be obtained directly from the image, is estimated as the cloud existence probability using the relationship between the altitude and the likelihood of cloud generation.
Specifically, as shown in FIG. 3 (b), the control unit 18 controls the vertical range (cloud height) of the cloud CL on the image calculated in step S <b> 3 and the own sensor obtained from the body sensor 13. Based on the altitude AL, the altitude (range) R at which the cloud can exist, the existence probability distribution of the cloud CL in a vertical section including the own aircraft and the highest altitude point of the cloud CL is estimated.
More specifically, the control unit 18 first determines the direction of the optical sensor 14 in the vertical section (that is, the elevation angle of the optical sensor 14), the cloud height of the cloud CL on the image, and the altitude AL of the own device. A straight line L connecting the aircraft and the highest altitude point of the cloud CL is obtained. Next, the control unit 18, as a range where the cloud CL can exist in the vertical cross section, is a cloud existence range surrounded by the straight line L and a possible cloud existence altitude R in consideration of the roundness of the ground surface G (Earth). EA is calculated. In general, darker clouds are more likely to be generated at low altitudes. Therefore, the control unit 18 has a higher probability of the presence of the cloud CL at a lower altitude position on the straight line L that is the upper edge of the cloud existence range EA. As a result, the existence probability distribution of the cloud CL with respect to the separation distance from the optical sensor 14 in the cloud existence range EA is estimated.
In FIG. 3B and FIGS. 4A and 4B described later, the level of existence probability of the cloud CL is indicated by gradation of gradation. In addition, the cloud CL indicated by a two-dot chain line in these drawings is an actual position of the cloud CL shown as a reference, and this position is not known in the cloud position estimation process of the present embodiment. Absent.

Next, the control unit 18 synthesizes the horizontal range of the cloud CL calculated in step S3 and the existence probability distribution of the cloud CL with respect to the separation distance from the optical sensor 14 estimated in step S4. As shown to (a), the cloud distribution map M showing the existence probability distribution of the cloud CL in the plane (substantially horizontal plane in this embodiment) containing the horizontal direction and sensor direction on an image is produced | generated (step S5). Then, the control unit 18 stores the generated cloud distribution map M in the storage unit 16.
In this way, a cloud distribution map M corresponding to the image acquired in step S1 is generated.

  Next, the control unit 18 stores in the storage unit 16 a cloud distribution map M (or an integrated cloud distribution map MI described later) of the cloud CL corresponding to the image acquired in the past than the image acquired in step S1. If it is determined that the image is not stored (step S6; No), the process proceeds to the above-described step S1 and the image acquisition of the monitoring target T is continued.

If it is determined in step S6 that the cloud distribution map M of the cloud CL corresponding to the image acquired in the past than the image acquired in step S1 is stored in the storage unit 16, the control unit 18 A plurality of cloud distribution maps M are integrated while weighting so that the newer one has a higher existence probability (step S7).
Specifically, as illustrated in FIG. 4B, the control unit 18 takes into account the movement of the own device based on the own device position information from the body sensor 13, and clouds based on images acquired at each position. The cloud distribution map M of CL is weighted and integrated so that the weight coefficient becomes larger as it corresponds to a newer image. Then, the control unit 18 causes the storage unit 16 to store the integrated cloud distribution map M as the integrated cloud distribution map MI.
In this way, a plurality of cloud distribution maps M having different acquisition times are integrated while being weighted by time, thereby enabling more accurate cloud CL position estimation.

  Next, the control unit 18 determines whether or not to end the cloud position estimation process by inputting an end instruction from the pilot (step S8). When the control unit 18 determines not to end the cloud position estimation process (step S8; No), the process proceeds to step S1 described above, and continues to acquire the image of the monitoring target T and ends it. When it determines (step S8; Yes), a cloud position estimation process is complete | finished.

[effect]
As described above, according to the present embodiment, the range of the cloud CL in the horizontal and vertical directions on the image acquired by the optical sensor 14 is calculated, and the cloud height of the cloud CL on the calculated image is calculated. Based on the altitude AL of the aircraft 1 (optical sensor 14) and the possible altitude R of the cloud, the existence probability distribution of the cloud CL with respect to the separation distance from the optical sensor 14 using the relationship between the altitude and the likelihood of cloud generation Is estimated. Then, based on the existence probability distribution of the cloud CL with respect to the separation distance from the optical sensor 14 and the horizontal range of the cloud CL on the image, a plane including the horizontal direction and the sensor direction on the image (the book) In the embodiment, a cloud distribution map M in a substantially horizontal plane) is generated.
Therefore, unlike the conventional case where the presence of a reference object for estimating the position of the cloud is essential, the position of the cloud CL is preferably determined from the image acquired by the optical sensor 14 without requiring such a reference object. Can be estimated.
In addition, since the position of the cloud CL can be estimated without the need for a weather radar, the weight and power consumption can be reduced, and the possibility of detecting the position of the aircraft by the weather radar emitting radio waves should be reduced. Can do.

Further, the optical sensor 14 that moves with the aircraft 1 acquires a plurality of images having different acquisition times and including the same cloud CL, and generates a plurality of cloud distribution maps M corresponding to the plurality of images. Then, the plurality of cloud distribution maps M generated are integrated while being weighted so that the weight coefficient becomes larger as it corresponds to a newer image, and an integrated cloud distribution map MI is generated.
Therefore, since the existence probability distribution of the cloud CL based on a new image acquired from a direction different from the past is added to the past as more accurate information, the position of the cloud CL is estimated more accurately. be able to.

[Modification]
The embodiments to which the present invention can be applied are not limited to the above-described embodiments, and can be appropriately changed without departing from the spirit of the present invention.

  For example, in the above embodiment, the sensor direction of the optical sensor 14 is substantially horizontal, but the sensor direction is not particularly limited as long as the sensor direction intersects the vertical direction.

  Further, when the integrated cloud distribution map MI is generated by integrating a plurality of cloud distribution maps M, the latest cloud distribution map M is sequentially added to the immediately preceding cloud distribution map M or the integrated cloud distribution map MI and the integration is performed. The cloud distribution map MI may be updated, or three or more cloud distribution maps M may be generated and then integrated simultaneously.

Further, although the case where the cloud position estimation apparatus according to the present invention is applied to the aircraft 1 has been described as an example, the present invention is widely applied to a technique for estimating the position of a cloud using an optical sensor installed at a high place. Is possible.
Therefore, for example, it is possible to estimate the position of the clouds with ground equipment using image information transmitted from an optical sensor mounted on an unmanned aircraft, or to install the optical sensor on an upper part of a high building or the like. Also good.

DESCRIPTION OF SYMBOLS 1 Aircraft 13 Airframe sensor 14 Optical sensor 16 Memory | storage part 160 Cloud position estimation program 18 Control part AL Self-machine altitude CL Cloud EA Cloud existence range L Straight line M Cloud distribution map MI Integrated cloud distribution map R The altitude where a cloud can exist

Claims (7)

A cloud position estimation device for estimating the position of a cloud using an optical sensor installed at a high place,
The optical sensor is oriented in a sensor direction that intersects the vertical direction and acquires an image including a cloud;
Cloud range calculation means for calculating a cloud range in each of horizontal and vertical directions on the image based on the image acquired by the optical sensor;
Based on the vertical range of the cloud on the image calculated by the cloud range calculation means, the altitude of the optical sensor, and the possible altitude of the cloud, using the relationship between altitude and the likelihood of cloud generation, Cloud distribution estimation means for estimating a cloud existence probability distribution with respect to a separation distance from the optical sensor;
Based on the horizontal range of the cloud on the image calculated by the cloud range calculation means and the cloud existence probability distribution with respect to the separation distance from the optical sensor estimated by the cloud distribution estimation means. A cloud distribution generation means for generating a cloud existence probability distribution in a plane including the horizontal direction above and the sensor direction;
A cloud position estimation device comprising: The optical sensor is mounted on a moving body, acquires a plurality of images having different acquisition times and including the same cloud,
The cloud range calculation means, the cloud distribution estimation means, and the cloud distribution generation means generate cloud existence probability distributions in the plurality of planes corresponding to the plurality of images,
2. The cloud distribution integration unit according to claim 1, further comprising a cloud distribution integration unit configured to weight and integrate the generated existence probability distributions of the plurality of clouds so that a weight coefficient increases as the image corresponds to a newer image. Cloud position estimation device. The cloud distribution estimation means includes
Based on the elevation angle of the optical sensor, the vertical range of the cloud on the image, and the altitude of the optical sensor, a straight line connecting the optical sensor and the highest altitude point of the cloud is calculated,
As a range in which a cloud can exist in a vertical section including the straight line, a cloud existence range surrounded by the straight line and the altitude where the cloud can exist in consideration of the roundness of the earth is calculated,
Calculating the cloud existence probability distribution with respect to the separation distance from the optical sensor in the cloud existence range, assuming that the lower the altitude of the position on the straight line in the cloud existence range, the higher the cloud existence probability. The cloud position estimation apparatus according to claim 1, wherein:   The cloud position estimation device estimates the position of a cloud included in the image when an image of a predetermined target is acquired by the optical sensor. The cloud position estimation apparatus according to claim 1.   The cloud position estimation apparatus according to claim 1, wherein the optical sensor is mounted on an aircraft. A cloud position estimation method for estimating a cloud position using an optical sensor installed at a high place,
As the optical sensor, one that is directed in a sensor direction that intersects the vertical direction and acquires an image including a cloud,
The cloud position estimation device
Based on the image acquired by the optical sensor, a cloud range calculation step for calculating a cloud range in each of the horizontal and vertical directions on the image;
Based on the vertical range of the cloud on the image calculated in the cloud range calculation step, the altitude of the optical sensor, and the possible altitude of the cloud, using the relationship between the altitude and the likelihood of cloud generation, A cloud distribution estimation step of estimating a cloud existence probability distribution with respect to a separation distance from the optical sensor;
Based on the horizontal range of the cloud on the image calculated in the cloud range calculation step and the cloud existence probability distribution with respect to the separation distance from the optical sensor estimated in the cloud distribution estimation step. A cloud distribution generation step of generating a cloud existence probability distribution in a plane including the horizontal direction above and the sensor direction;
The cloud position estimation method characterized by performing. A cloud position estimation program for estimating the position of a cloud using an image acquired by an optical sensor installed at a high place,
The optical sensor is directed in a sensor direction intersecting the vertical direction, and acquires an image including a cloud,
Cloud position estimation device
Cloud range calculation means for calculating a cloud range in each of horizontal and vertical directions on the image based on the image acquired by the optical sensor;
Based on the vertical range of the cloud on the image calculated by the cloud range calculation means, the altitude of the optical sensor, and the possible altitude of the cloud, using the relationship between altitude and the likelihood of cloud generation, A cloud distribution estimating means for estimating a cloud existence probability distribution with respect to a separation distance from the optical sensor;
Based on the horizontal range of the cloud on the image calculated by the cloud range calculation means and the cloud existence probability distribution with respect to the separation distance from the optical sensor estimated by the cloud distribution estimation means. A cloud distribution generation means for generating a cloud existence probability distribution in a plane including the horizontal direction and the sensor direction;
Cloud position estimation program characterized by functioning as