JP2940666B2 - Method and apparatus for selecting landing site of spacecraft - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is intended to automatically select an appropriate landing site without obstacles when a spacecraft lands on a celestial body such as the moon, and to finally enable a safe landing. Technical field relating to a method and a device for selecting a landing point of a simple spacecraft.

[0002]

2. Description of the Related Art Conventionally, in missions to land on the moon and Mars, images of the ground surface are acquired during orbit in orbit and transmitted to the earth, and an appropriate landing site is obtained by a control center on the earth. In some cases, if necessary, a command was given to the spacecraft in orbit to correct the previously programmed landing orbit. In this case, no correction operation is performed when the landing aircraft enters the landing operation. When imaging a surface such as the moon from an orbit, the distance is large and the resolution of the image is not sufficient.Therefore, it is not always possible to detect obstacles when actually landing without any extra small leaks, and safety is low. You only have to guess which areas are likely to be expensive. Further, since the landing trajectory actually taken by the landing aircraft may include an error from the planned trajectory, it is not always possible to accurately land at a planned position considered safe. Therefore, it is necessary to select a sufficiently wide flat land as the landing point to avoid the risk of a soft landing failure,
Sometimes you have to land a great distance from your immediate destination.

In order to avoid such uncertainties,
When the spacecraft lands on the surface of the moon or planet, approaching the landing target point, while observing the image near the target point acquired from the spacecraft, the flatness of the target point is improved by remote control from the ground or the mother ship. There is a method of guiding to a safe part and landing. If such a method is possible, a spacecraft can be guided and landed on a relatively narrow flat land if it has a minimum area enough for landing, By observing on the screen, even if an obstacle is found in a large area set in advance, it should be possible to avoid such an obstacle and land at a safe point. However, if such a remote control is used when the spacecraft and the control base are far apart, it takes time to communicate, so there is no room to make a correct decision in accordance with the operation of the spacecraft. There is a risk that the command will arrive late and land in a dangerous area. In order to avoid such telecommunications, the landing aircraft should be equipped with means to automatically detect areas without obstacles or the presence of obstacles, automatically search for suitable places and correct landing trajectories. You only need to land. However, it was difficult to actually mount such a suitable-for-landing detection device on a spacecraft because there were a problem with the weight of the device, a problem with the reliability of the judgment, and a problem with the time required for the judgment.

[0004]

SUMMARY OF THE INVENTION Therefore, the problem to be solved by the present invention is to solve the problem of obstacles near the landing point autonomously even after the spacecraft enters the landing position, without depending on the ground or the control of the mother ship. Observing and observing an obstacle at the planned landing site, search for a nearby landing site, and correct the landing trajectory so that the vehicle can land safely.

[0005]

In order to solve the above-mentioned problems, a method for selecting a landing site of a spacecraft according to the present invention acquires a multi-value image representing a distribution of brightness by imaging an area around a planned landing site.
A binary image is generated by binarizing the multi-valued image according to the brightness, and a distance conversion is performed by performing a distance conversion represented by a distance from an image boundary of the binary image to generate a distance conversion image. A safer landing point is calculated based on the above.

[0006] Further, a binarization value may be determined depending on whether the value is between two different threshold values. Note that these thresholds may be set based on a histogram obtained for brightness in a multi-valued image. Furthermore, the distance conversion is performed by recursively applying the reduction filter to the obtained binary image, and the output integrated value obtained by integrating the filter outputs each time the reduction filter is applied becomes smaller than the preset threshold value. At this time, a point corresponding to any of the remaining pixel positions may be set as a landing point. In addition, it is more preferable that, of the points corresponding to the remaining pixel positions, the point at which the cost required for correcting the trajectory with respect to the first scheduled landing point is minimized is the landing point.

In the method for selecting a landing point of a spacecraft according to the present invention, a device for detecting an obstacle or the like and selecting an appropriate landing point is mounted on the spacecraft itself, and the spacecraft lands in a landing orbit. After taking an attitude, this spacecraft-mounted device picks up a suitable site for landing based on the image obtained by imaging the area around the planned landing site and corrects the landing trajectory so as to conform to this It may be.

In order to solve the above problems, a landing point selecting apparatus for a spacecraft according to the present invention includes: an image pickup apparatus for picking up an image of a target ground surface to generate a multi-value image representing a brightness distribution of an image pickup area; A binarization circuit for inputting an image from the imaging device and binarizing the image based on a preset threshold to generate a binary image according to the brightness of the imaging region; A distance conversion circuit that calculates and stores a value based on the distance from the boundary of, and an arithmetic circuit that selects a landing point based on the distance conversion image stored in the distance conversion circuit.

The binarization circuit sets 1 when the pixel value of the acquired multi-valued image is between the two thresholds based on two different thresholds set in advance and 1 when the pixel value is not between the two thresholds. A binary image is generated by setting the value to 0, and the distance conversion circuit recursively applies the reduction filter to the generated binary image. The filter output is integrated every time the reduction filter is applied. An output integrated value is obtained by performing a distance conversion such that an image remaining when the output integrated value becomes smaller than a preset threshold value is used as a final distance conversion image. A point corresponding to any one of the pixel positions may be a landing point.

According to the method for selecting a landing point of a spacecraft according to the present invention, the distance between a binary image obtained by binarizing a multi-valued image representing the distribution of lightness obtained by imaging the periphery of the planned landing point according to the lightness is calculated. Based on the distance conversion image generated by performing the conversion, it calculates the safe landing point such as the center position of the appropriate landing area and the position at an appropriate distance from the danger area, so the amount of information is small. By performing a very simple logical operation, accurate judgment can be made. In order to realize the binarization processing, there is a method of accessing all the pixels by the CPU and performing table conversion. This method does not require resources other than the CPU and the memory, but depends on the processing speed of the CPU. There is also a method of accessing the image memory by a binarizing circuit and converting the table. This method is capable of high-speed processing, but requires the addition of a dedicated circuit, and increases the load on the spacecraft.

[0011] However, according to the above method, high-speed processing is not always possible, but high reliability can be guaranteed. If you can add a small and lightweight logic circuit, it will be possible to perform the desired operation sufficiently, and small rocks, crevasses, crevices, etc. that you overlooked when observed in orbit This can be avoided just before landing to prevent landing failure. Further, since the spacecraft itself can determine a safe place by judging the situation of the landing point just before landing, the previously specified landing candidate site does not need to be a large flat land as in the past.
Therefore, it is possible to select a proximity point that is convenient for the operation after landing.

In the case of obstacles protruding from a plane, such as rocks or mountain-shaped terrain rolling on the ground surface of the moon or Mars, the angle of incidence of sunlight on the sunny side increases the brightness, and The brightness on the flat surface or the flat surface on which the shadow of the obstacle is affected is extremely small due to the sparse atmosphere. In addition, the brightness of the crack existing in the flat surface is small. Therefore, by using two thresholds for binarization, if dark and bright areas are cut at the same time, mountains, rocks, or crevasses that may hinder the landing of the spacecraft will be removed at once. Eliminating, it is possible to easily extract only a flat surface portion that uniformly reflects. In addition, since the brightness level of the flat portion greatly changes depending on local conditions such as the altitude of the sun or the characteristics of the ground surface such as geology and grain size, it is difficult to determine an appropriate threshold value in advance. It is also possible to deal with the situation by selecting from several candidates and designating the threshold value according to the conditions at the time of landing. However, the threshold value used for binarization based on the histogram of the brightness in the image actually acquired is Is set, a threshold value suitable for the actual conditions of the actual landing is obtained, so that a suitable landing site can be determined accurately.

[0013] A binary image representing a flat land and an obstacle in the vicinity of the scheduled landing point is image-processed to determine the actual landing point, and the trajectory is corrected. Since the spacecraft is already in landing operation, it is not permissible to spend time searching for the best landing location and modifying the landing trajectory. For this reason, in the landing point selection method of the present invention, it is possible to deal with using a distance conversion method which is considered to be logically simple, can be operated at high speed by a simple logic circuit, and has no practical problem. Is also good. Although various distance conversion methods are known, as the number of pixels constituting an image increases, the processing operation time increases and the required memory also increases. For this reason, there are not many things that can be applied to a spacecraft in which light weight and small size are absolute requirements for actual mounting.

Several distance conversion methods that solve such difficulties are also known. For example, a well-known sequential distance conversion algorithm using an urban distance is first assigned a large number to a pixel having a binary number of 1 in an input binary image and to a pixel having a binary number of 0. The image memory is initialized by assigning 0, and 1 is added to the value of the immediately preceding pixel in the same column as the target pixel while scanning the pixels of the initialized image in the forward direction from the upper left to the lower right. The pixel value in the image memory is sequentially updated by replacing the pixel value in the image memory by using the smallest value of the value of the previous pixel in the same row as the value plus 1 and the value of the pixel itself. There is a method of obtaining a distance-converted image by sequentially updating pixel values in the same manner as above while scanning the updated image memory in the reverse direction from the lower right to the upper left. With this method, the distance function used in the present invention can be obtained by determining the distance function based on relatively simple logic and by one round trip scanning.

In the landing point selecting method of the present invention, a distance conversion is performed by recursively applying a reduction filter to the obtained binary image, and an output integration obtained by integrating filter outputs each time the reduction filter is applied. When the value becomes smaller than a preset threshold value, a point corresponding to any of the remaining pixel positions is set as a landing point, and high-speed image processing is performed using an extremely simple logic element. be able to. The reduction filter is a simple logical filter having a kernel size of, for example, 3 × 3. When the pixel of interest in the center of the kernel and all eight neighboring pixels around it are 1, the value is 1
It is a logical product output element that sets the value to 0 if at least one pixel is 0. The distance conversion is realized by recursively convolving the filter with the image, and the process of convolving all the pixels at a very high speed by incorporating the filter into a dedicated chip for image processing can be repeated.

When the reduction filter is operated once for all the pixels of the binary image, 1 is calculated from the contour of the pixel area having the value 1.
A new area is generated in which the value 1 is assigned to pixels inside the pixel. When a reduction filter is applied to this new image, a new area is generated in which the value 1 is assigned to pixels one pixel further inside from the outline of the previously generated area. If the same procedure is repeated, the values of all the pixels become 0 at the end, and the area left in the image immediately before that represents the central area of the flat surface, and the number of times of the reduction filter operation until then is flat. The distance from the contour line of the surface to the center corresponds to the value represented by the pixel distance on the image.

Accordingly, if the last remaining area is set as a landing point, the landing aircraft can be landed at a sufficiently safe place farthest from obstacles such as dangerous mountains, rocks, and cracks. It should be noted that there is no danger if it is far enough from the obstacle even if it does not stick to the farthest place from the obstacle. An area may be selected as a landing point. In addition, since the landing aircraft is already in a state of landing, it is necessary to determine at least one landing point in the acquired multi-value image. For this reason, every time the reduction filter is applied, the number of pixels having a result of 1 is counted to calculate the area of a region having a value of 1. When this integrated value becomes smaller than a predetermined threshold value, There is a method that makes it possible to make a safe landing by making an arbitrary point in the area remaining in the area as a landing point so that a candidate place always exists. In addition, the point corresponding to the pixel position of the distance-converted image in which the cost required for correcting the trajectory for the first scheduled landing point is minimized is set as the landing point, thereby easily switching to a new landing point. Can be operated economically.

Next, a device for detecting an obstacle or the like and selecting an appropriate landing point is mounted on the spacecraft itself, and after the spacecraft takes a landing attitude, the spacecraft is scheduled to land using the spacecraft mounted device. According to the method for selecting a landing point of a spacecraft according to the present invention, in which the landing trajectory is corrected so as to conform to a suitable landing site selected based on a captured image around the point, detection is performed by a device mounted on the spacecraft itself. There is no loss time due to the communication time with the control base outside the spacecraft, and appropriate correction operation based on terrain information that can be more accurately grasped by images enlarged as approaching the landing point can be determined. Therefore, it is possible to avoid a small rock, crevasse, crevices, etc. immediately before landing without failing to search for a large area landing light area by scrutinizing the moon surface etc. while in orbit. Prevention And it can be safely land the lander.

Further, according to the spacecraft landing point selecting device of the present invention, there is provided a binarizing circuit for processing a relatively small amount of information contained in the output of the imaging device mounted on the spacecraft at high speed according to simple logic. Since it is composed of a distance conversion circuit and an arithmetic circuit, it is possible to function sufficiently with a lightweight, compact and highly reliable logic circuit for mounting on a spacecraft that is relatively slow, and it can communicate with the control room on the earth and on the mother ship. Without intervening, it is possible to automatically calculate the optimum trajectory in the spacecraft and correct the landing trajectory. Note that a binary image is generated in which 1 is set when the pixel value of the obtained multi-valued image is between two different thresholds set in advance, and 0 otherwise.
A reduction filter is recursively applied to a generated binary image. The reduction filter is integrated when the filter output is integrated for each application of the reduction filter and remains when the output integrated value becomes smaller than a preset threshold value. Which performs a distance conversion such that the obtained image becomes a final distance conversion image, and the arithmetic circuit sets a point corresponding to any of the pixel positions of the final distance conversion image as a landing point. Since the logic used for the calculation is further simplified, the target landing candidate point can be selected with a smaller and lighter device.

[0020]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of a method and apparatus for selecting a landing point of a spacecraft according to the present invention will be described below in detail with reference to the drawings. The gist of the present invention is not to avoid by individually recognizing craters and mountain regions that become obstacles, but to determine a landing point at a position far from the obstacle in a region where no obstacle exists by simple image processing. Thus, the processing device is mounted on a spacecraft so that arithmetic processing can be performed automatically. Here, for simplicity,
A description will be given of a case in which a lunar lander that descends vertically on the lunar surface and lands, which enables relatively simple handling, is used. On the moon, sunlight is obliquely incident, and shadows appear on craters and mountains. It can be assumed that the reflectance of the lunar surface is almost uniform near the target landing site, and that the difference in brightness is mainly determined by the relationship between sunlight and the surface normal direction. Therefore, it suffices to search for and land on a surface having an appropriate area with uniform brightness.

[0021]

Embodiment 1 FIG. 1 is a block diagram showing a first embodiment of a method for selecting a landing point of a spacecraft according to the present invention, FIG. 2 is a flowchart thereof, and FIG. FIG. 4 is a conceptual diagram for explaining, FIG. 4 is an example of a binarization method used in the method of the present embodiment, and FIG. 5 is a diagram for explaining an example of a distance conversion method. In FIG. 1, reference numeral 10 denotes an imaging device such as a CCD camera, and 20 denotes an image processing device, which is mounted on a landing aircraft. The CCD camera 10 is directed vertically downward, and photographs the moon surface on which the landing aircraft lands. The image processing device 20 includes an image memory 21, a CPU 22, a filtering operation unit 23, and a pixel counting unit 24.

The image processing steps in the CCD camera 10 and the image processing device 20 are as shown in the flowchart of FIG. As shown in FIG. 3 as a model, the spacecraft 1 lowers its altitude along a scheduled landing trajectory toward a flat surface 2 on the moon. The flat land 2 to be landed was selected by observing the lunar surface while the spacecraft 1 was in an orbit away from the moon. The planned landing trajectory was optimized to land at this landing site. Orbit. However, at the landing site, there are rocks 3 and cracks 4 that can be overlooked from above. When the spacecraft 1 lands, you must avoid such obstacles.
The aircraft cannot be soft-landed safely. Even when the spacecraft 1 lands on the flat surface 2, if the spacecraft 1 lands on the boundary position of the flat surface, the support legs of the spacecraft may fall into the cracks 4 or hang on the rocks 3, and the attitude after landing may not be maintained. is there. Therefore, in order to surely land the spacecraft 1 on the flat ground surface 2, it is necessary to guide the spacecraft 1 by finding a portion as close to the center of the flat area as possible. In addition, since the spacecraft 1 is already in the landing attitude, it is preferable to be able to immediately determine a suitable place and to make the orbit correction as small as possible.

On the moon, the sun 5 is at a relatively low angle, and when the projections such as the rocks 3 are irradiated with the sun, the sun is strongly reflected and brightly illuminated, and the shade is illuminated by the atmosphere. It is dark because it is rare and light does not reach. Similarly, the lightness of the shadow 4 and the sunlit portion of the crack 4 has an extreme value. On the other hand, the portion of the flat land 2 has a substantially constant reflectance and diffuses the sunlight, has a substantially uniform brightness, and can be distinguished from a portion where an obstacle exists. The lightness of the flat land changes depending on the geology of the target area, the particle size of the material covering the surface, the distance from the sun, the altitude of the sun, and the like, and the measurement environment. Therefore, it is necessary that the brightness determined as a flat ground has a certain allowable width.

In the landing site selection method of the present invention, when the altitude of the spacecraft 1 decreases and reaches an appropriate altitude at which the terrain on the moon can be observed in more detail, the landing site selection device mounted on the spacecraft 1 is used. Start. When the processing step is started (S1), the CCD camera 10 captures an image of a spacecraft landing scheduled area on the moon surface that exists vertically below. In the captured images, rocks and mountains that look very bright on the sun side are drawn with dark shadows, and concave parts such as cracks in craters etc. show a dark shadow, so that the terrain can be read from the difference in brightness. It has become. The CCD camera 10 digitizes each pixel of the captured video based on the brightness and outputs it as a multi-valued image signal. The image memory 21 stores the CCD camera 10
Are stored in the form of a matrix (S2).

Next, the multi-valued image is binarized. The binarization is performed by setting a threshold T in advance, as is generally performed, and setting the pixel value of the corresponding mapping to 1 if the pixel value V of the multi-valued image is larger than the threshold T, and converting the pixel value to 0 otherwise. Good,
In this embodiment, the first threshold T1 and the second threshold T2 are provided in consideration of the fact that the shaded portion represents a crater or a mountain shade, and that the slope on the sunny side becomes brighter than the flat ground. When the pixel value V of the value image is between the first threshold value T1 and the second threshold value T2, the pixel value of the binarized image is set to 1, and the pixel value V is outside the first threshold value T1 and the second threshold value T2. At the same time or at the same time as these thresholds. Further, in view of the fact that the brightness of the flat ground changes depending on the measurement environment, in this embodiment, since an accurate threshold value individually corresponding to the state of the landing site is used,
The threshold is set based on the actual measurement value. The CPU 22 accesses the memory 21 and examines the appearance frequency of each pixel in the image for each lightness level to create a histogram. Using the histogram, a shaded portion that is generated near an obstacle and a bright portion on the sunny side. Are determined as a first threshold value T1 and a second threshold value T2 for determining a flat portion other than an obstacle (S3). Further, the brightness of each pixel in the multi-valued image is compared with a threshold value, and a binary image in which an area of a flat land where landing is possible is represented by 1 and an obstacle and its vicinity is represented by 0 is generated and stored in an image memory. (S4).

FIG. 4 shows an example of the binarization processing algorithm. The matrix on the left in the figure converts the lightness of the pixels in the effective portion of the light and dark image obtained by the CCD camera 10 vertically directed to the landing surface area on the moon surface into a hexadecimal number from 0 to F according to the lightness. 3 schematically shows a state in which the multi-valued image 31 is stored in an image memory. The matrix on the right side of the figure shows the brightness from 4 to A, for example, assuming that the first threshold T1 for a dark area is 3 and the second threshold T2 for an extremely bright area is B from the result of the histogram analysis of the multi-valued image 31. This represents a binary image 32 obtained as a result of performing a binarization process in which a portion having a level value is 1 and other portions are 0. This 2
In the value image 32, the lunar portion corresponding to the area where the value is 1 is a flat suitable landing site. However, in order to land safely, it is preferable to select a position as far as possible from an obstacle among the above-mentioned landing sites.

Then, next, the filtering operation means 2
3 performs distance conversion by applying a reduction filter having a kernel size of 3 × 3 to the binary image (S5). The distance conversion refers to a process of converting the value of each pixel having a value of 1 into a distance value from the contour when a boundary between 0 and 1 is defined as a contour in the binary image. The reduction filter performs an AND operation to set the value to 1 when all the pixels in the kernel are 1 and to set the value to 0 if at least one pixel is 0. This reduction filter is applied from the upper left to the lower right of the image.
When the scanning is performed once, a new value 1 area is generated by reducing the periphery of the value 1 area in the binary image by one pixel. This new region represents a set of pixels one or more pixels away from the outline of the first region inward. An image obtained by adding this new image to the first image is assigned a value of 1 to pixels that are less than one pixel away from the contour of the area of value 1 in the original image, and is assigned a distance of one pixel or more. The distance conversion image is obtained by adding a value 2 to a certain pixel.

When the same image processing operation is performed on a new region representing a set of pixels one or more pixels inward from the outline of the first region, this time represents a set of pixels two or more pixels inward from the outline. Is generated. When this is added to the previous distance conversion image, a distance conversion image is obtained in which the value 3 is assigned to the internal region further separated by two or more pixels. By repeating the reduction filter operation until all the areas of value 1 disappear, a distance conversion image expressing the distance from the obstacle by a numerical value can be obtained. FIG. 5 is a diagram illustrating an example of distance conversion based on the above logic. The left diagram in FIG. 5 shows an example of a binary image to be subjected to arithmetic processing, and the right diagram shows a distance conversion image 42 obtained by performing distance conversion on the binary image 41. In the figure, the value 1 of the binary image 41
It can be seen that the central portion of the region is located at a distance of 4 pixels from the contour.

Since the purpose of using the distance conversion in the method of the present invention is to find a position sufficiently distant from the obstacle, the distance conversion processing is repeated until the area of value 1 completely disappears. It is not enough to call. In this embodiment, the value 1 remaining by the pixel counting unit 24
Is calculated (S6), and when this area value becomes equal to or less than a certain threshold value N, the distance conversion process is terminated, thereby shortening the calculation time (S7). The above distance conversion is realized by recursively convolving this filter with the image, and it is possible to execute it by software.However, by incorporating it into a dedicated chip for image processing, extremely fast The convolution process can be repeated.

The area having the maximum value in the distance conversion image thus obtained indicates the central area of the flat part farthest from the obstacle, and is a point where the vehicle can land more safely. From these candidate sites, a site that minimizes an evaluation function in consideration of various costs required for trajectory correction, time until landing, fuel consumption, and the like is selected. Since the spacecraft travels toward the center of the image, the most basic determination method is to use the one with the shortest distance from the center of the image as the landing point. The CPU 22 accesses the distance conversion image information stored in the image memory, calculates the distance from the image center of the candidate point (S8), determines the shortest distance as the landing point (S9), and operates An instruction to correct the landing trajectory is given to the device, and a series of arithmetic processing ends (S10).
Since the number of times of the convolution process is directly used as an index representing the distance from the outline of the region, the number of processes corresponding to the distance from the obstacle required for safe landing is calculated and set in advance. However, it is also possible to set a region having a value of 1 remaining when the set number of processes is repeated as a landing candidate region.

In this embodiment, the threshold value for binarization is determined based on the histogram obtained from the actually measured value. However, in the present invention, the threshold value may be determined in advance as a fixed value. Needless to say. Further, it may be obtained by calculation from the angle of incidence of the sun and the state of the moon surface without depending on the actually measured value. In these cases, there is an effect that the amount of calculation to be performed during a busy time during landing is reduced to some extent.
Also, instead of performing binarization by comparison with a threshold,
The calculation time can be reduced by using a look-up table (LUT). In the above description, the binning process is performed in which the land suitable for landing is 1 and the obstacle is 0. However, it is needless to say that the process may be reversed or another binary value may be used.

[0032]

[Embodiment 2] FIG. 6 is a block diagram showing a second embodiment of the apparatus used to carry out the method for selecting a suitable landing site for a spacecraft according to the present invention. The point that the suitable landing site selecting apparatus of the present embodiment is different from that of the first embodiment is that the arithmetic unit is simplified by performing the advanced image processing after reducing the image information at first. By the way. The CCD camera 51 captures an image of the moon surface vertically below from the spacecraft, and
A video signal of 12 × 512 pixels is supplied. The image processing device 52 receives this video signal, and receives 2
After binarization by the binarization circuit 53, the image information reduction circuit 54 divides the 512 × 512 bit binary number matrix by 8-bit vertical and horizontal bits, reduces the information by taking the logical sum within the square, and 64 × 64. Convert to a matrix of bits. This reduced matrix is image information that has a value of 1 when an obstacle is included in the portion of the moon corresponding to the first eye, and has a value of 0 when there is no obstacle. Avoiding obstacles allows for safer maneuvering.

The lunar picked-up image reduced to 64 × 64-bit information is stored in the reduced image memory 55 and the CPU 56
Accesses this, converts it into a 64 × 64-bit distance image by the same distance conversion method as in the first embodiment, and stores it in the distance image memory 57. From the distance conversion image stored in the distance image memory 57, a point at which landing can be safely performed with the slightest trajectory correction is calculated, and a trajectory correction instruction is given to the control device of the spacecraft so that it can land at this new suitable landing site. give. With such a configuration, instead of performing a large amount of calculation based on the high resolution of the CCD camera 51, the amount of information required for safe landing is narrowed down at the beginning, thereby reducing the load on the processing unit. Can be reduced, so that a sufficiently practical calculation can be performed even for a spacecraft-mounted computer having a relatively low calculation speed instead of high reliability.

[0034]

Third Embodiment FIG. 7 is a block diagram showing a third embodiment of an apparatus used for carrying out the method for selecting a suitable landing site for a spacecraft according to the present invention. The difference between the landing site selecting apparatus of this embodiment and that of the first embodiment is that the CPU accesses the image information stored in the memory to create a histogram relating to the brightness level, and uses this histogram to determine the flat surface. Instead of calculating the threshold value, a high speed calculation is performed by a dedicated circuit without applying a load on the CPU to obtain the threshold value. It is the same as in the second embodiment that a binary image is obtained using these thresholds and then reduced to simplify the determination of a suitable landing position.

The CCD camera 61 picks up an image of the moon surface vertically below, generates an image signal of, for example, 512 × 512 pixels, and generates a first image signal.
Is stored in the image memory 62. Histogram generation circuit 6
Reference numeral 3 sequentially inputs the image signals stored in the first image memory 62 and accumulates the frequency for each brightness level to create a histogram. The histogram represents the brightness distribution shape according to environmental conditions such as rock quality and the altitude of the sun on the surface of the moon. Therefore, the sunlit portion of the highland where the brightness becomes highest, the sunlit portion of the crack, and the shade portion where the brightness becomes the lowest are removed, and the middle brightness range indicated by the flat portion can be known. The binarization parameter determination circuit 64 executes such an analysis by a logic circuit prepared in advance, and calculates and determines a first threshold value and a second threshold value for cutting out a flat portion. While reading the image signal from the first image memory 62, the binarization circuit 65 compares the image signal of 512 × 512 pixels with the threshold value determined by the binarization parameter determination circuit 64 to binarize the image signal. Generate an image. Next, the image reduction circuit 66 converts the binarized 512 × 512 pixels into 8 ×
It is determined whether or not there is an obstacle in each area, divided every eight pixels, and a new binary matrix composed of 64 × 64 areas is generated and stored in the second image memory 67. The CPU 68 accesses the second image memory 67 and
The reduced image of the 4 × 64 area is subjected to the same image processing as in the first embodiment, a safe position is selected even when the spacecraft lands, and a trajectory correction command for landing the spacecraft at its landing point is automatically set. Transmit to the control device.

When such a configuration is adopted, the computational load is reduced, so that a sufficiently practical computation can be performed even with a spacecraft CPU having a low computation speed in addition to high reliability. . In the above description, the histogram circuit 63 creates a histogram using the image signal once stored in the image memory 62.
Instead, the output of the CCD camera 61 is stored in the image memory 6.
In parallel with storing the data in 2, each pixel is directly input and the brightness of each pixel is sequentially examined, and the number of pixels for each brightness level is integrated.
It is also possible to complete the histogram almost at the same time when the image signal is stored in the image memory 62, calculate the threshold value, and immediately start the binarization operation. As is clear from the description of each of the above embodiments, the landing point selection method and apparatus of the present invention can be applied to any celestial object whose brightness changes according to the terrain illuminated by the sun. It is also effective when landing on the surface of a planet or asteroid.

[0037]

As described above, the method and apparatus for selecting a landing site for a spacecraft according to the present invention are mounted on the spacecraft itself and secured on the basis of an enlarged image of the vicinity of the planned landing site acquired from above the landing site. When selecting a landing site based on terrain information acquired from the orbit, it is not necessary to select a wide flat area in consideration of the safety factor to select a landing point. Scrutinization and selection are no longer necessary, and conditions for a suitable landing site have been eased.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of a method for selecting a landing point of a spacecraft according to the present invention.

FIG. 2 is a flowchart illustrating a landing point selection method according to the present embodiment.

FIG. 3 is a conceptual diagram illustrating a usage state of a landing point selection method according to the present embodiment.

FIG. 4 is a diagram illustrating an example of a binarization method used in the method of the present embodiment.

FIG. 5 is a diagram illustrating an example of a distance conversion method used in the method of the present embodiment.

FIG. 6 is a block diagram showing a second embodiment of a landing point selecting device for a spacecraft according to the present invention.

FIG. 7 is a block diagram showing a third embodiment of a landing point selecting apparatus for a spacecraft according to the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Spacecraft 2 Lunar flat surface 3 Rock 4 Crack 5 Sun 10 Imaging device 20 Image processing device 21 Image memory 22 CPU 23 Filtering operation means 24 Pixel counting means 31 Multi-valued image 32 Binary image 41 Binary image 42 Distance conversion image Reference Signs List 51 CCD camera 52 Image processing device 53 Binarization circuit 54 Image information reduction circuit 55 Reduced image memory 56 CPU 57 Distance image memory 61 CCD camera 62 First image memory 63 Histogram generation circuit 64 Binarization parameter determination circuit 65 Binarization Circuit 66 Image reduction circuit 67 Second image memory 68 CPU

──────────────────────────────────────────────────の Continuing on the front page (72) Inventor Toshiyuki Yoshida 1 Kawasaki-cho, Kakamigahara-shi, Gifu Kawasaki Heavy Industries, Ltd. Gifu Plant (72) Inventor Daihachi Ogawa 1, Kawasaki-cho, Kakamigahara-shi, Gifu Kawasaki Heavy Industries, Ltd. (56) References JP-A-8-181976 (JP, A) JP-A-8-287400 (JP, A) JP-A-9-280856 (JP, A) JP-A-9-301300 (JP) , A) (58) Fields investigated (Int. Cl. ⁶ , DB name) B64G 1/66 G06T 1/00 H04N 7/18 H04N 5/225 G01C 7/02 G01C 11/24 G08G 5/02 G08G 9 / 02

Claims (7)

(57) [Claims] 1. A by imaging the surrounding land planned site obtains multivalued image representing the distribution of brightness, binarizes follow the multivalued image brightness to generate a binary image, the image of the binary image A landing point selection method for generating a distance conversion image by performing distance conversion represented by a distance from a boundary, and calculating a safer landing point based on a distance value in the distance conversion image. 2. The landing point selecting method according to claim 1, wherein the value of the binarization is determined based on whether or not the value is between two different thresholds. 3. The landing point selection method according to claim 2, wherein said threshold value is determined based on a histogram obtained for brightness of said multi-valued image. 4. The distance conversion is performed by recursively applying a reduction filter to the binary image, and integrating a filter output for each application to obtain an output integrated value. 4. A landing point selection method according to claim 1, wherein a point corresponding to one of the remaining pixel positions when the pixel value becomes smaller than the set threshold value is set as a landing point. 5. The landing point selection method according to claim 1, wherein a point at which the cost required for correcting the trajectory with respect to said landing point is minimized is a landing point. 6. An imaging device for imaging a target surface to generate a multi-valued image representing a brightness distribution of an imaging region, and binarizing the multi-valued image based on a preset threshold value input from the imaging device. A binarization circuit for generating a binary image according to the brightness of a region, and a distance conversion by inputting the binary image from the binarization circuit and calculating a value for each pixel based on a distance from an image boundary;
A landing point selection device comprising: a distance conversion circuit that generates and stores an image ; and a calculation circuit that selects a landing point based on a distance value in the distance conversion image stored in the distance conversion circuit. 7. The multi-level image processing apparatus according to claim 1, wherein the binarizing circuit determines a pixel value of the multi-valued image based on two different thresholds set in advance.
Generating the binary image by distinguishing between when and between the thresholds, the distance conversion circuit,
A reduction filter is applied recursively to the binary image, and the filter output is integrated every time the reduction filter is applied once to obtain an output integrated value, and the output integrated value becomes smaller than a preset threshold value. claims residual image subjected to distance transform to the distance transform images, the arithmetic circuit is characterized in that the landing site of the point corresponding to either of the remaining pixel positions of the distance transform image when the 6. The landing site selection device according to 6 .