JP2001266117A - Picture processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To effectively remove only a noise component given on an observed picture. SOLUTION: A concentration adjusting part 42 adjusts the dispersion value/ average value of an observation picture signal at present time, which is obtained by an optical sensor 1, and an observes picture signal showing the same observation object, which is held by a delay part 41 and which is previous by prescribed time. A registration part 43 adjusts the positions of still pictures on both observation pictures and a difference operation part 44 performs a difference operation. A binarization part 45 extracts the signal component of not less than a prescribed threshold from an obtained differential picture signal and generates a binary picture signal. A picture degeneration part 47 degenerates a differential picture component from the obtained binary picture signal. A picture inflation part 48 restores it to an original size and the picture signal constituted only of a picture change component from which a noise component is removed is generated. A masking part 3 masks the original observation picture signal stored in a storage part 2.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for processing an observed image signal photographed by, for example, a monitoring camera mounted on a satellite or an ITV (industrial television) camera.

[0002]

2. Description of the Related Art Generally, in an image processing apparatus for processing an observation image signal photographed by, for example, a monitoring camera mounted on a satellite or an ITV camera, an image acquired at each time is set as an image at a time earlier than an acquisition time or as a reference. The difference calculation between the images has been basically used to grasp how the images have been changed.

[0003] However, the attitude of the satellite is changing every moment.
Due to the orbital fluctuation, the captured images in the same area also cause slight mutual distortion. Also, even in a surveillance system or an inspection system using an ITV camera or the like, a slight change in light / illuminance, a change in the attitude of the camera, and the like cause noise in the difference image substantially in accordance with the contour of the target screen. However, it is extremely difficult to automatically discriminate the part where the change has actually occurred from the noise part.

[0004]

As described above, in the conventional image processing apparatus, when noise on an observed image is given due to various factors such as satellite attitude / orbit variation, camera attitude variation, and the like, image change occurs. It is difficult to distinguish between a noise portion and a noise portion, and there is no effective means for removing only noise.

An object of the present invention is to provide an image processing apparatus capable of effectively removing only a noise component given on an observation image.

[0006]

According to the present invention, there is provided an image processing apparatus comprising: an observation sensor for photographing an object to be observed; storage means for storing an observation image signal photographed by the observation sensor; Characteristic adjustment means for making the same predetermined characteristics of the observation target between the observation image signal or the observation image signal before storage and the observation image signal obtained by photographing the same observation target a predetermined time ago, Difference calculating means for performing a difference calculation between the plurality of image signals whose characteristics have been adjusted by the characteristic adjusting means, and extracting a signal component having a predetermined threshold value or more from the difference image component obtained by the difference calculating means Binarizing means for forming a binary image by performing
A noise removing unit for removing a noise component from the binary image signal obtained by the binarizing unit; and masking the observed image signal stored in the storage unit based on the image signal obtained by the noise removing unit. And a mask means. In particular, the characteristic adjusting means includes:
The method is characterized in that, for a plurality of observed image signals, one of the average value and the variance value of the observation target is the same between the plurality of observed image signals. And a positioning means for positioning a specific image. Further, the noise removing means includes image reducing means for reducing the output of the binarizing means, and image expanding means for expanding the difference image component reduced by the image reducing means to restore the original size. It is characterized by.

That is, in the present invention, only the image change component is expressed by taking the difference between the current observation image signal obtained by the observation sensor and the observation image signal obtained by photographing the same observation target a predetermined time ago. Focusing on creating an image, the variance value / average value of each image signal is adjusted to make both of the multiple image signals for which the difference is to be the same density level, After adjusting the position of the specific image, a difference operation is performed, and a signal component having a predetermined threshold value or more is extracted from the obtained difference image signal to generate a binary image signal. And
In order to distinguish between a noise component and a valid image change component from the obtained binary image, the difference image component is degenerated,
Thereafter, by restoring to the original size, an image signal of only the image change component from which the noise component has been removed is created, and the original observed image signal is masked based on this image signal.

Therefore, according to the present invention, the original observation image signal is masked by the image signal created by degenerating the difference image component obtained by the difference operation and returning it to the original size. It is possible to obtain an image in which only the noise component is removed from the obtained observed image signal and only the image change component is extracted.

Further, in the above configuration, the noise removing means is provided with a labeling means for writing an identifier into the area of code 1 of the binary image binarized by the binarizing means, so that the noise is eliminated. Since a binarized pattern can be generated for each image component, and degeneration and expansion can be performed separately for each pattern, the processing load in the noise removal processing can be significantly reduced. Further, the probability of deleting valid information can be reduced.

[0010]

Embodiments of the present invention will be described below in detail with reference to the drawings.

(First Embodiment) FIG. 1 is a block diagram showing an embodiment of an image processing apparatus according to the present invention. The image processing apparatus illustrated in FIG. 1 captures (scans) an observation target such as a mountain area or an urban area at a time interval set in advance by the optical sensor 1, and acquires an observation image signal.
The observed image signal is converted into a digital value by an A / D (analog / digital) converter 11 and stored in the storage unit 3. The observation image signal stored in the storage unit 3 is input to the mask unit 3 and the mask image generation unit 4. The mask image generation unit 4 includes a delay unit 41, a density adjustment unit 42, a registration unit 43, a difference calculation unit 44, a binarization unit 45, a labeling unit 46, an image reduction unit 47, an image expansion And a unit 48.

The observation image signal input to the mask image generation unit 4 is supplied to a delay unit 41 and a density adjustment unit 42, respectively. The delay unit 41 delays the observation image signal by a predetermined time using, for example, a RAM or the like, and the delayed output is supplied to the density adjustment unit 42. The density adjusting section 42 receives the input signal and the output signal of the delay section 41 and adjusts them so that their densities are the same. here,
The two adjusted image signals are registered in a registration unit 43 based on, for example, a still image, and a difference operation is performed in a difference operation unit 44 to generate a difference image signal. It is supplied to the unit 45.

The binarizing unit 45 generates a binary image signal having logical codes 1 and 0 by extracting a signal component having a predetermined threshold value or more from the input differential image signal, and sends the binary image signal to the labeling unit 46. Output. The labeling unit 46 adds different labels as identifiers to the area indicating the logical code 1 to the input binary image signal. The binarized image signal to which the label has been added is decompressed by the image decompression unit 47 in the region to which the label has been added, and expanded by the image expansion unit 48 to return to the original size, thereby removing noise components. The generated mask image signal is supplied to the mask unit 3.

The mask unit 3 performs masking on the observation image signal stored in the storage unit 2 based on the mask image signal generated by the mask image generation unit 4,
The noise component is removed, and an image signal representing only this image change component is displayed on the display unit 5.

Next, the operation of the above configuration will be described with reference to FIGS. Note that FIG.
FIG. 3 is a diagram illustrating an example of an image generated by the operation. That is, the density adjustment unit 42 includes the observation image signal stored in the storage unit 3 as shown in FIG. 2B and the past observation image signal output from the delay unit 41 as shown in FIG. Is input. These two image signals are combined by the density adjustment unit 42 with the variance value / average value of each image signal in order to make them the same density level.
After the positions of the mountains, which are still images, on both observation images are adjusted in 3, a difference operation is performed by the difference operation unit 44 to generate a difference image signal as shown in FIG. You. In the difference image signal, a signal component equal to or larger than a predetermined threshold value is extracted by a binarization unit 45, and is generated as a binary image signal represented by logical codes 0 and 1 as shown in FIG. . In this binary image signal, an area indicating a logical code 1 indicates an image change component and a noise component, and an area indicating a logical code 0 indicates a still image component. Note that some noise components have been removed from this binary image signal.

The binarized image signal is supplied to a labeling section 46.
After the label is added to the area indicating the logical code 1 in FIG.
As shown in FIG. 3C, the image degenerating unit 47 performs degeneration for each label, and the image expanding unit 48 restores the original size and removes the noise component as shown in FIG. And is supplied to the mask unit 3. Then
The mask unit 3 can identify a noise component included in the observation image signal stored in the storage unit 2 from the mask image signal, and generates an image signal representing only the image change component by masking the noise component. can do.

As described above, in the first embodiment, the current observation image signal obtained by the optical sensor 1 and the delay unit 4
Focusing on creating an image representing only the image change component by taking the difference from the observed image signal representing the same observation object a predetermined time ago held by 1 In order to obtain the same density level, the variance value / average value of the two image signals is adjusted by the density adjustment unit 42, and the still image positions on both observation images are adjusted by the registration unit 43. Above, a difference operation is performed by the difference operation unit 44, and a signal component having a predetermined threshold value or more is extracted from the obtained difference image signal by the binarization unit 45 to generate a binary image signal. Then, in order to identify a noise component and a valid image change component from the obtained binary image, the difference image component is reduced by the image reduction unit 47, and then restored to the original size by the image expansion unit 48. By doing so, an image signal of only the image change component from which the noise component has been removed is created, and based on this image signal, the masking unit 3 masks the original observed image signal stored in the storage unit 2.

Accordingly, by masking the original observed image signal with the image signal created by degenerating the difference image component obtained by the difference operation and returning to the original size, the observed image obtained by the optical sensor 1 is obtained. It is possible to obtain an image in which only the noise component is removed from the signal and only the image change component is extracted.

Further, the first embodiment is characterized in that the labeling section 46 adds different labels to the area of the logical code 1 of the binary image, so that the binarization is performed for each difference image component. Since the reduced pattern can be generated and the reduction and expansion can be performed separately for each pattern, the processing load in the noise removal processing can be remarkably reduced. Further, the probability of deleting valid information can be reduced.

In the first embodiment, the storage unit 3
Although the example in which the mask image signal is generated from the observation image signal stored in the optical sensor 1 has been described, the mask image signal may be directly generated from the output of the optical sensor 1.

(Second Embodiment) FIG. 4 shows a second embodiment of the image processing apparatus. In FIG. 4, the same parts as those in FIG. 1 are denoted by the same reference numerals, and description thereof will be omitted.

In this image processing apparatus, the output of the storage unit 2 is input to the mask image generation unit 4 in the first embodiment, whereas the output of the mask unit 3 is input to the mask image generation unit 4. I have to. That is, in the configuration of the second embodiment, the noise component is removed by a feedback loop, whereas the noise component is removed by a feedforward loop in the first embodiment. For this reason, the configuration of the first embodiment is advantageous in terms of response speed, whereas the configuration of the second embodiment is controlled in terms of output results, which is advantageous in terms of accuracy.

(Third Embodiment) FIG. 5 shows a third embodiment of the image processing apparatus. In the third embodiment, the density adjustment process, the registration process, the difference calculation process, the binarization process, the labeling process, the image reduction process, and the image expansion process are collectively performed by one signal processing unit functioning as a microcomputer. This is done by doing.

An observation image signal obtained by photographing the observation target by the optical sensor 100 is converted into an A / D converter 1
01 is converted into a digital value in the storage units 102 and 10
3 is stored. For the observation image signal stored in the storage unit 102, the signal processing unit 104 performs density adjustment processing,
Registration processing, difference calculation processing, binarization processing,
After performing the labeling process, the image reduction process, and the image expansion process, the mask image signal from which the noise component has been removed is output to the mask unit 105. The mask unit 105 removes noise components by masking the observation image signal stored in the storage unit 102 based on the mask image signal, and outputs the result to the display unit 106. The storage unit 10
The observation image signal stored in No. 3 is held longer than the storage unit 102 by a predetermined time.

Next, the operation of the signal processing unit 104 will be described with reference to the flowchart of FIG.

That is, the signal processing unit 104 stores in the storage unit 1
After adjusting the average value and the variance value of the two observation image signals read from 02 and 103 (step S1), adjusting the position of the still image on both observation images by registration processing (step S2), The difference calculation between the observed image signals is performed (step S3). From the difference image signal obtained by this difference operation, a signal component having a predetermined threshold value or more is extracted to generate a binary image signal represented by logical codes 1 and 0 (step S4).

Next, the signal processing unit 104 adds different labels to the regions indicating the logical code 1 of the binarized image signal (step S5), and performs image degeneration for each label (step S6). After the reduced image is restored to its original size to generate a mask image signal (step S7), the mask image signal is output to the mask unit 105.

In the above processing, the signal processing unit 10
4 generates a noise correction image signal representing only noise components by subtracting the mask image signal generated in step S7 from the binary image signal generated in step S4, and outputs the noise correction image signal to the mask unit 105. (Step S8).

As described above, the present invention can also be implemented by the third embodiment. In particular, since each process can be realized by software, the configuration is simplified,
This can contribute to downsizing.

(Other Embodiments) The present invention is not limited to the above embodiments. For example, a synthetic aperture radar may be used instead of the optical sensor. Also, without being limited to sensors mounted on satellites, for example, industrial C
It may be a CD camera and an entrance / exit to a building or an inspection object on a production line.

The present invention is also applicable to, for example, earth observation satellites and spacecraft for planetary exploration.
It is also applicable to the V system. Further, in the above-described embodiment, an algorithm such as a labeling process, an image degeneration process, and an image dilation process is applied to remove noise. An algorithm that creates an image in which only the change component is extracted is also applicable.

Further, according to the present invention, as shown in FIG.
The present invention can also be applied to a system having an additional function such as outputting an alarm based on an image change extraction result and controlling other devices. In the system of FIG. 7, the artificial satellite S1 captures an image of an observation target using an optical sensor, and transmits an observation image signal from which noise components have been removed to the ground station as mission data. At the ground station, the incoming mission data is stored in the antenna 20
1 and demodulated by the receiving / demodulating / recording unit 202 and temporarily recorded, and then received by the radiometric correction unit 2
03 and the geometric correction unit 204 remove the noise component and the wraparound component given during transmission of the mission data, and supply the result to the change extraction processing unit 205.
The change extracting unit 205 converts the observation image included in the mission data into a photograph, and when an image change component exists in the observation image, notifies the alarm display unit 206 of the image change component and causes the alarm display unit 206 to display an alarm. This alarm-displayed observation image signal is supplied to the satellite operation planning device 207, and transmitted to the artificial satellite S1 via the antenna 201 as a detailed imaging instruction for the artificial satellite S1 and an imaging instruction signal for the next orbit. Become.

Therefore, according to such a system, target detection and the like can be reliably and accurately performed.

The present invention also relates to, for example, a stacked CT
3D or pseudo 3D image processing such as images, and additional processing such as filtering before and after the main elements of the image processing algorithm to improve the processing accuracy,
A method for comprehensively judging the results by using a plurality of algorithms for extracting image changes, and a method for proceeding with a man-machine system while confirming whether there are many errors in noise removal and whether effective information has been deleted. It is needless to say that the present invention can be carried out as appropriate.

[0035]

As described in detail above, according to the present invention,
It is possible to provide an image processing device capable of effectively removing only a noise component given on an observation image.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of an image processing apparatus according to the present invention.

FIG. 2 is a view showing an example of both images before and after a delay by a delay unit of the embodiment.

FIG. 3 is an exemplary view showing an example of an image generated by image processing in the embodiment.

FIG. 4 is a block diagram showing a configuration of an image processing apparatus according to a second embodiment of the present invention.

FIG. 5 is a block diagram showing a configuration of an image processing apparatus according to a third embodiment of the present invention.

FIG. 6 is a flowchart showing an operation procedure and operation contents by a signal processing unit of the third embodiment.

FIG. 7 is a block diagram showing a configuration of a system according to another embodiment of the present invention.

[Explanation of symbols]

1,100: Optical sensor, 2, 102, 103: Storage unit, 3, 105: Mask unit, 4: Mask image generation unit, 5, 106: Display unit, 11, 101: A / D (analog / digital) converter 41, a delay unit, 42, a density adjusting unit, 43, a registration unit, 44, a difference calculating unit, 45, a binarizing unit, 46, a labeling unit, 47, an image degenerating unit 48, an image expanding unit, 104, a signal Processing unit, 201: antenna, 202: reception / demodulation / recording unit, 203: radiometric correction unit, 204: geometric correction unit, 205: change extraction processing unit, 206: alarm display unit, 207: satellite operation planning device.

Claims (9)

[Claims] 1. An observation sensor for photographing an observation target, a storage means for storing an observation image signal photographed by the observation sensor, and the same observation target photographed a predetermined time before the observation image signal stored in the storage means Characteristic adjustment means for making the predetermined characteristic of the observation target the same between the observation image signal obtained as described above, and a plurality of image signals characteristic-adjusted by the characteristic adjustment means. Difference calculating means for performing a difference calculation; binarizing means for forming a binary image by extracting a signal component having a predetermined threshold value or more from the difference image component obtained by the difference calculating means; Noise removing means for removing a noise component from the binary image signal obtained by the means; and masking the observed image signal stored in the storage means based on the image signal obtained by the noise removing means. An image processing apparatus comprising: a mask unit. 2. An observation sensor for photographing an object to be observed, storage means for storing an observation image signal photographed by the observation sensor, and an observation image signal which is stored a predetermined time before the observation image signal before being stored in the storage means. Characteristic adjustment means for making the predetermined characteristic of the observation target the same as an observation image signal obtained by photographing the image signal; and a plurality of image signals characteristic-adjusted by the characteristic adjustment means. A difference operation means for performing a difference operation between the two, a binarization means for forming a binary image by extracting a signal component having a predetermined threshold value or more from the difference image component obtained by the difference operation means; A noise removing unit that removes a noise component from the binary image signal obtained by the binarizing unit; and masking the observed image signal stored in the storage unit based on the image signal obtained by the noise removing unit. An image processing apparatus comprising: a mask unit for performing 3. An observation sensor for observing an observation target, a storage means for storing an observation image signal photographed by the observation sensor, and an observation image signal stored in the storage means based on a given mask image component. On the other hand, the masking means for performing masking and the predetermined characteristic of the observation target are made the same between an output of the masking means and an observation image signal obtained by photographing the same observation target a predetermined time ago. Characteristic adjustment means; difference operation means for performing a difference operation between the plurality of image signals whose characteristics have been adjusted by the characteristic adjustment means; and a predetermined threshold value or more based on the difference image component obtained by the difference operation means Binarization means for forming a binary image by extracting the signal components of the above, and removing noise components from the binary image signal obtained by the binarization means to obtain a mask for the mask means. An image processing apparatus comprising: a noise removing unit that generates a black image component. 4. The apparatus according to claim 1, wherein the characteristic adjustment unit sets one of an average value and a variance value of the observation target to be the same among a plurality of observation image signals. An image processing device according to any one of the above. 5. The image processing apparatus according to claim 1, further comprising a positioning unit that positions a specific image between the plurality of image signals whose characteristics have been adjusted by the characteristic adjusting unit. An image processing device according to any one of the above. 6. The image reduction means for reducing the output of the binarization means, and the image expansion means for expanding the difference image component reduced by the image reduction means to return to the original size. The image processing apparatus according to claim 1, further comprising: 7. The noise removing means includes: a labeling means for adding an identifier to an area indicating a logical code 1 to an output of the binarizing means; and an image degenerating means for degenerating an area to which the identifier has been added by the labeling means. 4. The image processing apparatus according to claim 1, further comprising: means for expanding the difference image component reduced by the image reducing means and returning the difference image component to the original size. 8. The apparatus according to claim 1, wherein said noise removing means further comprises a subtracting means for subtracting an output of said image expanding means from an output of said binarizing means and applying the difference image component to said mask means. The image processing apparatus according to any one of 1, 2, 3, 6, and 7. 9. The observation sensor according to claim 1, wherein the observation sensor is an optical sensor or a synthetic aperture radar.
The image processing device according to any one of the above.