JP2000113175A - Sea state chart preparation system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prepare an effective sea state chart without a defective area such as a cloudy area while reducing an error without performing preprocessing to an original satellite image. SOLUTION: A synthetic image 12 is generated by selecting the maximum value of pixel values at the same pixel position out of plural satellite images 11 through an image synthesizing means 2, a defective area removing means 3 judges the area, where the deviation between the pixel values of the synthetic image 12 and each satellite image 11 at the same pixel position is higher than a threshold, as the defective area of cloud or noise and removes it, the average value of respective remaining pixel values is set as the pixel value of the same pixel position, a defective area removed image 13 is generated, an interpolated image 14 is generated by interpolating the respective pixel values of defective area removed images 13 through an interpolating means 14, and an isoterm image 15 generated from this interpolated image 14 is synthesized to this image and a sea state chart 17 is prepared.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an oceanographic charting system, and more particularly, to the use of a satellite image taken from an artificial satellite to create an oceanographic chart showing the water temperature distribution on the sea surface displayed on an isotherm without any loss due to a cloud area. The present invention relates to a marine chart creating system.

[0002]

2. Description of the Related Art In general, an oceanographic chart showing the water temperature distribution on the sea surface has been recognized as being useful as fishery information, and is used to determine operating points in accordance with the water temperature distribution. To create such a sea chart, we request observations from thousands of fishing boats, including government research vessels existing in the range of the chart, and accumulate several days' worth of the data to collect water temperature distribution maps, or sea charts. Has been created.

[0003] However, in this sea chart drawing method, several days of information are collected and organized in order to draw a single sea chart,
Based on this information, humans gather in "can",
Requires expert skills. For sea areas where vessel data is not available, empirical predictions from past data are included, which raises questions about the objectivity of the created sea charts.

[0004] Furthermore, in the offshore waters, which is the main part of the figure, it seems that there is no particular problem from the viewpoint of the type of fishery, but the situation is different for coastal fisheries. In other words, for the coastal fishing boats, the daily fine movement of the front is a factor that determines the abundance of fishing.

[0005] In view of the above, as a method of tracking daily marine micro phenomena and creating information useful for coastal fisheries, that is, a sea chart, considering the immediacy, synchronism, and wide area, artificial It is considered that a method of creating an oceanographic map based on a satellite image taken from a satellite is very effective. However, since a satellite image taken from an artificial satellite, for example, NOAA / AVHRR image data is image data taken in a wavelength band of an infrared region or a visible region, in most cases, a cloud region exists in the image, The information on the sea surface below the cloud area cannot be obtained.

Conventionally, when obtaining sea surface temperature distribution information from a satellite image, a plurality of satellite images photographed during a predetermined period are synthesized based on the assumption that the sea surface temperature changes at a slower rate than the movement of the cloud area. Accordingly, a method of temporally interpolating a missing area such as a cloud area included in each satellite image has been proposed. FIG. 18 is a flowchart showing a conventional satellite image synthesizing method.

First, image data of a plurality of satellite images taken during a predetermined period is read (step 181), and a representative value at the same pixel (pixel position) is calculated for each pixel among the image data (step 182). ). As a method of calculating the representative value, there is a method of setting an average value, a maximum value, a deviation value, or the like of the same pixel as a representative value.

Then, one new composite image data is generated from the obtained representative value for each pixel (step 183). Thus, even if a missing area such as a cloud area exists in any of the satellite images, it is interpolated by another satellite image having a time difference that does not affect the change in sea surface temperature, and the water temperature distribution information with a certain degree of accuracy is obtained. Was obtained.

[0009]

However, when a sea chart is created by using such a satellite image synthesizing method, a representative value is simply calculated for each pixel from image data of a plurality of satellite images captured during a predetermined period. Is used to generate a composite image, and in order to obtain an accurate oceanographic chart, cloud area removal processing and noise removal processing are required for each original image data, increasing the processing load In addition, when these processes are incomplete, there is a problem that the accuracy of the obtained sea condition map is reduced.

For example, when the average value of the same pixel is obtained between each image data and is used as the representative value of the pixel, or the representative value of the pixel is set between the image data based on the deviation value of the same pixel. If the cloud area and noise are completely removed by pre-processing, the most average composite image during the period when each image data was obtained can be obtained, Image data can be synthesized. However, it is impossible to completely remove the cloud area and noise, and as a result, many errors are included.

When the maximum value of the same pixel is obtained between the image data and used as the representative value of the pixel, noise data is generally selected because the sea surface temperature tends to be higher than that of the cloud area or other parts. It is the least likely. However, when the temperature change is large, the information on the low temperature side is ignored, and as a result, many errors are included.

The present invention has been made to solve such a problem, and it is possible to create a useful sea chart with a small error and no missing area such as a cloud area without performing preprocessing on an original satellite image. The purpose is to provide a method for creating sea condition maps.

[0013]

In order to achieve the above object, a sea chart drawing system according to the present invention extracts pixel values at the same pixel position from a plurality of satellite images taken within a predetermined period. Image synthesizing means for generating a composite image by setting the maximum value to the pixel value at the same pixel position of the composite image, and calculating the deviation between the pixel value of the composite image at the same pixel position and the pixel value of each satellite image. Calculate and remove those pixel values of these satellite images whose deviation is larger than a predetermined threshold value as a missing area such as cloud or noise,
A missing area removing unit that generates an missing area removing image by setting an average value of the remaining pixel values to a pixel value at the same pixel position of the missing area removing image.

Further, an interpolation means for generating an interpolated image by interpolating each pixel value of the missing area removal image from pixel values around the pixel value, and an isothermal line by extracting an isothermal line from the interpolated image An isotherm extracting means for generating an image,
A sea chart output means for generating a sea chart by synthesizing the interpolation image and the isotherm image.

Therefore, according to the present invention, a plurality of satellite images photographed within a predetermined period are synthesized using the maximum value at each pixel position, and each pixel is obtained from the obtained synthesized image and each satellite image. Temporal interpolation processing is performed by removing missing areas such as clouds or noise based on the deviation. And
A spatial interpolation process is performed by interpolating each pixel value of the image from which the missing measurement area has been removed with surrounding pixel values, and an oceanographic map showing the temperature distribution of the sea surface is created using the obtained interpolated image.

[0016]

Next, the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a sea chart drawing system according to an embodiment of the present invention, and FIG. 2 is a flowchart showing a processing procedure of the sea chart drawing system. In addition,
Actually, these processes are executed by an arithmetic processing system such as a computer, and each process is realized by a program.

First, a plurality of satellite image data 10 photographed over a predetermined period are input to the satellite image input means 1, the image data is subjected to format conversion, atmospheric influence correction, etc., and output as a satellite image 11. (Step 21). As the satellite image data 10, image data in which the water temperature is expressed as the gradation value of each pixel, for example, AV
HRR / NOAA data is used. This data is based on the National Oceanic and
It was taken from a NOAA satellite managed by Atmospheric Administration.

The NOAA satellite has an AVHRR sensor (Advanced Very High Resolutio) which is a visible infrared radiation thermometer.
n Radiometer), and the temperature distribution on the ground surface and the sea surface is measured by a total of 4 to 5 channels of visible channel 2 and infrared channel 2 to 3. In the following description, the AVHRR / NOAA data obtained by the “infrared channel 4” that targets the temperature distribution on the sea surface will be referred to as satellite image data 1
0, and the satellite image input means 1 generates and uses a satellite image 11 as shown in FIGS.

In the image synthesizing means 2, these satellite images 11
Are synthesized using the maximum value at each pixel position to obtain a synthesized image 1
2 (FIG. 2: step 22). FIG. 3 is a flowchart showing a satellite image combining process, and FIG. 4 is an explanatory diagram showing a satellite image combining process. Each satellite image 11A ~
When synthesizing 11N, first, an unprocessed pixel position is selected (step 31), and the pixel value of the selected pixel position (that is, the gradation value indicating the water temperature) is extracted from each of the satellite images 11A to 11N (step 32). ).

Then, the maximum value (highest water temperature) is selected from the extracted pixel values (step 33), and the selected maximum value is set as the pixel value at the selected pixel position of the composite image 12 (step 34). These series of processes (steps 31 to 34) are repeatedly executed for all pixels (step 35), and the process ends.

Normally, since the temperature of the cloud area is lower than the sea surface temperature, even if a cloud area exists in any of the satellite images 11A to 11N at an arbitrary pixel position, the normal sea surface temperature is also indicated. If the pixel value is included, the maximum value (highest water temperature) of each of the satellite images 11A to 11N is selected, so that the cloud area is replaced with the normal sea surface temperature, and the cloud area and the similar low temperature are used. The side noise is removed, and a composite image 12 as shown in FIG. 13 is generated.

The composite image 12 thus generated
Is input to the missing area removing means 3 as an image for reference,
Based on this, the missing area of each satellite image 11 is removed, and the missing area removed image 13 is generated (FIG. 2: Step 23).
FIG. 5 is a flowchart showing the missing area detection processing of the composite image, and FIG. 6 is an explanatory diagram showing the missing area removal processing of the composite image.

When removing the missing area from each of the satellite images 11A to 11N, first, an unprocessed pixel position is selected (step 51), and the pixel value of the composite image 12 at the selected pixel position (that is, the floor indicating the water temperature). Key value) and each satellite image 11A
Calculate the deviation from the pixel value of ~ 11N (step 5
2). Then, each satellite image 1 at the selected pixel position
Among the pixel values of 1A to 11N, those whose absolute value of the deviation is equal to or smaller than a predetermined threshold are extracted (step 53).

Thereafter, the average value of the extracted pixel values is calculated (step 54), and the obtained average value is set as the pixel value at the selected pixel position of the missing area removal image 13 (step 55). These series of processes (steps 51 to 51)
55) is repeatedly executed for all pixels (step 5).
6), end the process.

As a result, of the satellite images 11A to 11N, a pixel value whose deviation from the pixel value (maximum value) of the composite image is larger than a predetermined threshold value is a missing area such as a cloud area or noise. It is determined that the possibility is high and removed, and a representative pixel value at the selected pixel position is calculated and set from the pixel values indicating the remaining normal sea surface temperature, for example, by averaging these pixel values. As shown in 14,
A missing area removal image 13 is generated. If the pixel values of each of the satellite images 11A to 11N at the same pixel position are all larger than the threshold value, a missing area is set.

Therefore, all the satellite images 11A to 11A
By comparing the average value or the deviation value of each pixel at the same pixel position of N as the representative value, for the pixel position where the normal sea surface temperature is detected in any of the satellite images 11A to 11N, There is no possibility that defective data such as a cloud area or noise is included, and errors due to these defective data can be reduced.

As the threshold value, an empirically derived value may be used, but each of the satellite images 11A to 11N may be used.
By using the difference between the maximum water temperature and the minimum water temperature actually observed by a ship or the like in the sea area to be processed during the period in which the is captured, the missing measurement area can be determined more accurately.

The missing area removal image 13 generated in this way is input to the interpolation means 4, and the missing area removal image 13
The missing area included in is interpolated and the interpolated image 14 (water temperature diagram)
Is generated (FIG. 2: Step 24). FIG. 7 is a flowchart showing the interpolation processing of the missing area removal image, and FIG. 8 is an explanatory diagram showing the interpolation processing of the missing area removal image.

When interpolating the missing area included in the missing area removal image, first, an unprocessed pixel position is selected (step 7).
1) Extract pixel values of pixel positions P1 to P8 within a predetermined influence range R from the selected pixel position P0 (step 7).
2), and calculate the average value of these (step 73). Subsequently, the average value is set as the estimated value of the pixel position P0 (step 74), and the estimated value is compared with each pixel value of the pixel positions P1 to P8, and those having a large difference are excluded (step 74). 75).

The remaining pixel values are weighted and averaged based on the distance from the pixel position P0 to calculate an analysis value at the pixel position P0 (step 76). Set to value (Step 7
7). These series of processes (steps 71 to 77) are repeatedly executed for all pixels (step 78), and the process ends.

As a result, the pixel values of the missing area included in the missing area removal image 13 can be accurately interpolated from the surrounding pixel values, and the pixel values at all the pixel positions, that is, the gradation values indicating the sea surface water temperature are obtained. The calculated interpolation image 14 (water temperature chart) as shown in FIG. 15 is generated. As the estimated value, as described above, the average pixel value of each pixel within the influence range may be used, but the sea surface temperature actually observed by a ship or the like in the sea area corresponding to the influence range is used. This makes it possible to suppress an interpolation error even at a pixel position inside a relatively wide missing measurement area.

If the influence range R is increased, the error included in the interpolation image 14 can be reduced, and if the influence range R is reduced, the time required for the interpolation processing can be reduced. Therefore, by appropriately selecting the influence range R, a sea chart with a desired error level can be obtained at a desired time. Note that the same influence range R may be used in all sea areas, but may be set for each predetermined sea area.

For example, in a sea area where a warm current and a cold current are mixed, an error is likely to occur due to a large and complicated change in water temperature. In such a case, the error can be reduced by increasing the influence range R. Further, in a sea area where the water temperature change is small, such as the central part of the Kuroshio, the interpolation processing time can be shortened by reducing the influence range R. Therefore, the range of influence R depends on the sea area where the magnitude of the water temperature change differs.
Is set, it is possible to obtain a sea chart with a small error in a relatively short processing time.

The interpolated image 14 thus generated
Is input to the isotherm extraction means 5 to extract an isotherm of the sea surface temperature and generate an isotherm image 15 (FIG. 2: Step 25). FIG. 9 is a flowchart showing the isotherm extraction process, and FIG. 10 is an explanatory diagram showing the isotherm extraction process.

When extracting isotherms from the interpolated image 14,
First, the water temperature change in the interpolated image 14 is frequency-analyzed by two-dimensional FFT (Fast Fourier Transform) (step 91). In this case, the magnitude of the change in water temperature between unit pixels, for example, between adjacent pixels is regarded as a frequency, and the level of each frequency spectrum is calculated. Subsequently, a cutoff frequency used for the low-pass filter processing is selected from the analysis result (step 92).

The method of determining the cutoff frequency differs depending on how to draw the isotherm. For example, if a relatively low cut-off frequency is selected, a smooth isotherm can be drawn, as is commonly used in fisheries, and selecting a high frequency will result in a level of complexity that can be used for marine research. You can draw a simple isotherm. Therefore, by selecting an arbitrary cutoff frequency, a sea state chart of an isotherm according to the purpose of use can be created. Experience shows that 0.8
About 1.0 [Hz / degree] is appropriate.

A low-pass filter process is performed on the interpolated image 14 using the cutoff frequency selected in this way to generate a smoothed image (step 93). And
An isotherm is extracted from the obtained smoothed image to generate an isotherm image (isotherm diagram) 15 as shown in FIG. 16 (step 94), and the process ends. As a result, after generating a smoothed image by averaging the surrounding pixel values, an easy-to-view isotherm image 15 corresponding to the water temperature distribution of each interpolated image 14 can be shortened as compared with a case where an isotherm is extracted. Can be generated in time.

The isotherm image 15 (isotherm diagram) generated by the isotherm extracting means 5 and the interpolation means 4
The interpolated image (water temperature chart) generated by the above is synthesized by the sea chart output means 6, and a sea chart as shown in FIG. 17 is output (FIG. 2: step 26).

[0039]

As described above, according to the present invention, a plurality of satellite images shot within a predetermined period are synthesized using the maximum value at each pixel position, and the obtained synthesized image and each satellite image are synthesized. Then, a temporal interpolation process is performed by removing missing areas such as clouds or noise based on the deviation of each pixel. Then, a spatial interpolation process is performed by interpolating each pixel value of the image from which the missing area has been removed with its surrounding pixel values, and using the obtained interpolated image, a sea state map showing the temperature distribution of the sea surface is created. It is like that.

Therefore, compared with the case where a synthetic image is generated by simply obtaining a representative value for each pixel from the image data of a plurality of satellite images, the cloud area elimination processing and the noise elimination processing for each original satellite image data are performed. This eliminates the need for processing, greatly reducing the processing load, and making it possible to create a useful oceanographic map with small errors and no missing areas such as cloud areas.

[Brief description of the drawings]

FIG. 1 is a block diagram of an oceanographic chart creation system according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a processing procedure of the sea condition chart creating system.

FIG. 3 is a flowchart illustrating a process of synthesizing a satellite image.

FIG. 4 is an explanatory diagram showing a process of synthesizing a satellite image.

FIG. 5 is a flowchart showing missing area removal processing of a composite image.

FIG. 6 is an explanatory view showing a missing area removal process of a composite image.

FIG. 7 is a flowchart showing interpolation processing of a missing area removal image.

FIG. 8 is an explanatory diagram showing an interpolation process of a missing area removal image.

FIG. 9 is a flowchart showing an isotherm extraction process.

FIG. 10 is an explanatory diagram showing an isotherm extraction process.

FIG. 11 is an explanatory diagram showing an example of a satellite image.

FIG. 12 is an explanatory diagram showing an example of a satellite image.

FIG. 13 is an explanatory diagram illustrating an example of a composite image.

FIG. 14 is an explanatory diagram illustrating an example of a missing area removal image.

FIG. 15 is an explanatory diagram illustrating an example of an interpolation image.

FIG. 16 is an explanatory diagram showing an example of an isotherm image.

FIG. 17 is an explanatory diagram showing an example of a sea condition diagram.

FIG. 18 shows a conventional image combining process.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Satellite image input means, 2 ... Image synthesis means, 3 ... Missing area removal means, 4 ... Interpolation means, 5 ... Isothermal line extraction means, 6 ... Sea state map output means, 10 ... Satellite image data, 11 ... Satellite images , 12: Composite image, 13: Missing area removal image, 14: Interpolated image (water temperature diagram), 15: Isotherm image (Isotherm diagram), 1
6 ... sea condition map.

Claims (5)

[Claims] 1. A sea chart creation system for creating a sea chart showing a water temperature distribution on the sea surface using a satellite image taken from an artificial satellite, comprising: a plurality of satellite images taken within a predetermined period; Image synthesizing means for generating a synthesized image by extracting a value and setting the maximum value to a pixel value at the same pixel position of the synthesized image; a pixel value of the synthesized image at the same pixel position and a pixel value of each satellite image Of the pixel values of each of these satellite images, the one whose deviation is larger than a predetermined threshold value is determined as a missing area such as a cloud or noise and removed, and the average of the remaining pixel values is calculated. A missing area removing unit that generates a missing area removing image by setting a value to a pixel value at the same pixel position of the missing area removing image; By interpolating from the pixel values of Interpolation means for generating an interpolation image, isotherm extraction means for generating an isotherm image by extracting an isotherm from the interpolation image, and a sea chart for generating a sea chart by synthesizing the interpolation image and the isotherm image. An oceanographic chart creation system comprising an output unit. 2. A marine chart creating system according to claim 1, wherein said missing area removing means is actually observed in a sea area to be processed during a period in which each satellite image used in said image synthesizing means is photographed. An oceanographic chart creation system characterized by using a difference between a maximum water temperature and a minimum water temperature as a threshold value. 3. The marine chart creating system according to claim 1, wherein the interpolation means calculates, for each pixel of the missing area removal image, an average value of each surrounding pixel included within a predetermined influence range from the pixel position. It is characterized in that, as an estimated value of a position, a peripheral pixel having a large difference from the estimated value is excluded, and a weighted average of the remaining peripheral pixels is set to a pixel value of the pixel position in the interpolation image. Oceanographic charting system. 4. The sea condition chart creating system according to claim 1, wherein the interpolating means sets, for each pixel of the missing area elimination image, an estimated value of the water temperature actually observed at the pixel position, and determines a predetermined value from the pixel position. Among the surrounding pixels included in the influence range, those having a large difference from the estimated value are excluded, and the remaining surrounding pixels are set to the pixel value at the pixel position in the interpolation image by weighted averaging. Marine charting system. 5. The sea condition chart creating system according to claim 1, wherein the isotherm extracting means frequency-analyzes a change in water temperature in the interpolated image by two-dimensional FFT, selects a cutoff frequency, and uses the selected cutoff frequency. An oceanographic chart creation system, wherein an isotherm is extracted after smoothing an interpolation image by low-pass filter processing.