WO2006072255A1

WO2006072255A1 - Digital radar system with clutter reduction

Info

Publication number: WO2006072255A1
Application number: PCT/DK2006/000014
Authority: WO
Inventors: Karsten Biehl
Original assignee: Navico Danmark A/S
Priority date: 2005-01-10
Filing date: 2006-01-10
Publication date: 2006-07-13

Abstract

The present invention relates to a digital radar system with clutter reduction. A first object of the invention is to provide a system for clutter reduction and target detection in a radar system enabling detection of small targets in signals with high amounts of clutter and simultaneously enabling detection of fast moving targets. A further object of the invention is to avoid latency when an operator changes manual adjustments of e.g. gain, clutter controls or changing display modes. This might be achieved by a radar system where the processor means store N independent scan-related images in a memory, which images are stored in a true-motion format. The processor means maintain a clutter map. The processor means generate M 2-dimentional sensitivity maps, the calculation of which may be based on the clutter map, the scan-related images and manual input from the operator. The processor means perform M scan-to-scan correlations, each using one of the sensitivity maps. The resulting M images from the scan- to-scan correlations are combined into a single image, which can be rotated and displaced to be presented to the operator, in his choice of display mode. Each scan-to- scan correlation can use data from 1 to N scan related images. By using individual sensitivity maps for each scan-to-scan correlation, the contribution of clutter from each scan-to-scan correlation can be controlled.

Description

Digital radar system with clutter reduction

FIELD OF THE INVENTION

The present invention relates to a radar system comprising a digital processor means, which processor means are connected to at least one memory, in which memory the processor means generate and store at least a first and a second digital scan-related image, which scan-related images comprise a number of independent pixels, which processor means generate at least a first correlated bitmap by a pixel-to-pixel logical scan calculation between scan-related images, which processor means generate an output image based on at least one correlated bitmap.

BACKGROUND OF THE INVENTION

A radar system is a very important tool for navigation and prevention of collisions of ships or of other objects on the sea. However, the radar image could sometimes be very noisy due to clutter from rain, snow or sea, which could lead to a misinterpretation of the echoes received, hi a worst case scenario, the clutter could conceal a ship on collision course, which could result in a catastrophic outcome. Therefore, various forms of radar systems with clutter reduction have been proposed. A simple technique for clutter reduction is based on the amplitude of the echoes received, which erases all the echoes below a certain amplitude. The pitfall of this method relates to the fact that small targets could be filtered out.

The nature of clutter is stochastic and can be described by statistical measures in a two-dimensional clutter map.

UK 1125789 describes radar equipment with clutter reduction means. The radar system here described is an analogue radar system where clutter reduction is made by allowing the passage of those echo pulses which exceed a set amplitude threshold where an adjustment of the amplitude threshold value is dependent on stored pulse numbers resulting from echo pulses arising from the transmitted radar pulse preceding, whereby the output of that second threshold device exhibits a clutter reduction.

US 6307501 describes a radar system that uses a clutter map for clutter reduction. The clutter map contains information of the measured clutter level, which is used for reducing the detected clutter. The clutter map is periodically re-centred to maintain its origin of coordinates.

It is well known to use scan-to-scan correlation between scans for clutter reduction in a true motion display mode.

hi the following, the terms scan-related image and independent are often used. The term scan-related image refers to a digital radar image obtained from digitalisation of echoes received from one scan by the radar system. The word independent is used to distinguish between information received from separate scans. Furthermore, the term bitmap refers to a two-dimensional array comprising a number of pixels which pixels each has a digital value.

Furthermore, the variables M and N, which are used in the following, are integers.

DESCRIPTION / SUMMARY OF THE INVENTION

The object of the invention is to provide a system for clutter reduction and target detection in a radar system enabling detection of small targets in signals with high amounts of clutter and simultaneously enabling detection of fast moving targets. A further object of the invention is to avoid latency when an operator changes manual adjustments of e.g. gain, clutter controls or changes display modes.

This might be achieved by a radar system where the processor means store N inde- pendent scan-related images in a memory, which images are stored in a true-motion format, where the processor means generate a sensitivity map based on input data. The processor means perform M correlations between N independent scan-related images by use of sensitivity maps during a mathematical calculation of the scan-related im- ages which processor means further process the M correlated bitmaps for generating an image for presentation on e.g. a radar screen.

The processor means maintain a clutter map. The processor means generate M 2- dimentional sensitivity maps, the calculation of which may be based on the clutter map, the scan-related images and manual input from the operator. The processor means perform M scan-to-scan correlations, each using one of the sensitivity maps.

The resulting M images from the scan-to-scan correlations are combined into a single image, which can be rotated and displaced before presented to the operator, in his choice of display mode. Each scan-to-scan correlation can use data from 1 to N scan related images. By using individual sensitivity maps for each scan-to-scan correlation, the contribution of clutter from each scan-to-scan correlation can be controlled.

In an embodiment of the invention the processor means use input data from manual adjustment means for generating the sensitivity maps. This way, the gain adjustment of the radar receiving unit can be adjusted to a high value allowing noise and clutter to pass through and to be stored in the digital memory where the filtering of the received radar images first takes place after these images have been digitally stored, and where the filtering takes place between corresponding pixels stored in independent scan- related images. By using a fast radar system and fast processor means, the M and N integers may be numbers that are higher than 10, and in future radar systems, the integers may be much higher. Li this way, clutter or noise that only appears in some scans can be filtered out during the calculation of the image presented to the operator. Hereby, an effective clutter reduction can take place in that the correlation time for clutter signals are typically significantly smaller than the scan time of the radar system.

In an embodiment of the invention the processor means use input data from manual adjustment means for generating the sensitivity maps. The manual adjustment means might be control buttons or the like. Because the present and previous scan images are preserved and stored, the manual adjustments take place immediately avoiding latency in the radar image. The sensitivity maps could be controlled and adjusted by a combination of manual adjustment by the operator and automatic adjustment by the processor means. The present invention is very flexible in the sense that many combinations of correlations of independent scan related images are possible. In one embodiment of the invention the processor means can store a first, a second and a third independent scan- related image where a logical calculation of related pixels from the first, second and third scan-related images generates pixels of a third correlated bitmap. In a preferred embodiment of the invention, the logical calculation can be carried out by a scan-to- scan correlation between the scans in order to detect the targets and at the same time remove some of the clutter. Hereby, it is achieved that three independent radar scans are used for the logical calculation of the pixels in the images. The digital analysis of the pixels could be made so that only targets that occur in all three pictures are accepted as targets. In this way, clutter might be reduced in that only if the clutter occurs in three independent scans, the clutter will be accepted. However, even rather weak targets will give even small pixel values, and if this small pixel value occurs in all three scans, this value will be accepted as a target. This means that only targets that are not present in all of the three scans will not be detected.

In another embodiment of the invention four independent scan related images are used in combination to generate the output image. This is achieved if the processor means can store the first, the second, the third and a fourth independent scan-related images where a logical calculation of related pixels from the first, second, third and fourth scan-related images are used by the processor means for generating pixels of a fourth correlated bitmap. Hereby, it is achieved that even more clutter is reduced so effectively that even the gain adjustment, which might be used before the digitalising of the images, could be adjusted to a level where most of the noise and clutter are accepted. In this way, even very weak targets will lead to positive pixel values in the resulting images. If even a very weak echo is indicated in the same pixels in four out of four scans, this could be accepted as a target and presented on a screen.

The logical calculations which can be made at corresponding pixels at the first, the second, the third and the fourth scan-related image where each pixel is correlated in three out of four corresponding pixels for generation of the pixel, where a correlated image is created on the basis of generated pixels. Hereby, it is achieved that even if a weak target is detected in only three out of four corresponding pixels, it will be accepted.

Individual sensitivity maps are used in order to suppress clutter and at the same time detect small and/or fast moving targets, clutter from each scan-to-scan correlation can in this way be controlled.

In an embodiment of the invention the radar system might comprise a first, second, third and fourth sensitivity map stored in the memory where an update of the pixel values are performed by the processor so that the first sensitivity map has a first low value, where the second sensitivity map has a lower value than the first sensitivity map, where the third sensitivity map has a lower value than the second sensitivity map, and where the fourth sensitivity map has a lower value than the third sensitivity map. Hereby, it is achieved that each sensitivity map is adjusted dependent on to which of the scan correlations it is related. Furthermore, an active filtering is achieved with reference to the actual sensitivity maps, and a filtering could be performed in a subtraction of digital values where the pixel values from the sensitivity maps are subtracted from related bitmaps.

The correlated bitmaps can be multiplied by a weighting factor, which weighting factor corresponds to the respective correlated bitmap for generating an output image. This way, the output image contains information from a number of individually weighted correlated scanned images. Furthermore, a number of scan-correlated bitmaps could be combined for creating a digital image which image contains parts of different scan correlations. This leads to a very fast reaction on the radar image of fast moving targets relatively near the radar system, but at the same time, an effective clutter reduction and an effective detection of weak targets make these targets visible on the image.

The pixels of the first scan provide the basis for an update of the corresponding pixels in the clutter map, which clutter map is based on statistical analysis of the pixels in the first scan and the corresponding pixels in the clutter map. The statistical analysis could comprise a calculation of the mean and the standard deviation of the clutter. Hereby, a dynamic clutter reduction, which is continuously self-adjusting is achieved. Additionally the sensitivity maps could be adjusted by the operator by manual controls. In this way the clutter reduction could be a combination of automatic and manual controls. If the manual parameters are changed by the user, it would have impact on the presented radar image immediately. This can be achieved because a number of previous true- motion scan-related images are stored, which images have not been filtered to remove clutter.

The output radar image is updated and processed continuously where the update of the radar output image is based on combinations of N independent scan-related images stored in the memory, which scan-related images have not been filtered and are in true-motion format. Hereby, adjustments made by an operator would have impact on the output image immediately and latency in the radar output image is avoided even though the output image is based on present and previous independent scan-related images. The adjustments could be clutter controls, sensitivity level controls, gain settings or changing of display modes from e.g. true-motion to relative-motion.

SHORT DESCRIPTION OF THE DRAWING

The invention will be explained in more detail with reference to the drawing, where fig. 1 shows a general system overview of the radar system, fig. 2 shows the calculation of the clutter and the sensitivity map, fig. 3 shows a filtering process, fig. 4 shows a filtering with a one-out-of-one and three-out-of-four situation, fig. 5 shows the pixel filtering in a one-out-of-one process, fig. 6 shows the pixel filtering in a two-out-of-two process, and fig. 7 shows the pixel filtering in a three-out-of-four process.

DETAILED DESCRIPTION / PREFERRED EMBODIMENT

Fig. 1 shows a general overview of the radar system comprising a radar antenna 140 connected to an analogue transceiver unit 160 from which the echoes are received and delivered to an analog to a digital converter 180. The A/D converter 180 is connected to processor means 100, which makes digital signal processing of the echoes possible. The radar system is connected to a memory unit 120 for storing the scan-related images, the sensitivity maps and the clutter map. The scan-related images are stored in the scan image memory 122, and the sensitivity maps are stored in the sensitivity map memory 124. The clutter map is stored in the clutter map memory 126, which is updated continuously for each scan. The radar system also comprises a display 134 for presenting a radar image and an interface 132 for manual gain and clutter adjustment etc. Further the system could comprise connection means to an external position system 130 and. an external heading measuring system 138.

Fig. 2 shows the calculation of a clutter map 264 and the processing of a first sensitivity map, which map is based on the first scan-related image 242. First, the clutter map 264 is updated by the clutter map calculation 262 followed by the processing of the first sensitivity map, which first sensitivity map is processed by the sensitivity map calculation 282. The maps are stored in the clutter map memory 260 and the sensitivity map memory 280.

The clutter map calculations are performed by statistical analysis of the clutter in the first scan-related image 242 and the values of the present clutter map 264.

The sensitivity map calculation 282 is based on a statistical description of the clutter stored in the clutter map 264. The sensitivity map calculation 282 could also be controlled by manual gain adjustment by the user. The clutter map and the first sensitivity are continuously being updated for each scan.

Fig. 3 shows the processing of four independent scan-related images. In the one-out- of-one situation 301, a first scan-related image 342 is filtered by a first sensitivity map 382, which is carried out by the target detection 390. The first sensitivity map 382 has a relatively high subtraction level compared to the following situations 303,305,307. Hereby, most of the clutter is reduced. In the next step of the process, the clutter- reduced map 383 is processed by the scan correlation 370. However, only one bitmap 383 is available in this situation where the scan correlation 370 is shown for illustra- tion purposes. Finally, the correlated image 371 is adjusted by a first weighting factor 350.

In the two-out-of-two situation 303 the first and second scan-related images 344 are filtered by a related second sensitivity map 384 which filtering is carried out by the target detection 392. The sensitivity map 384 comprises a medium high subtraction level compared to the following situations 305,307. Hereby, most of the clutter is reduced. Subsequently, the clutter-reduced maps 385 are processed by a two-out-of-two scan correlation 372 replacing each pixel value that is not present in both clutter re- duced maps. Finally, the correlated image 373 is adjusted by a second weighting factor 352.

In the three-out-of-three situation 305, the first, second and third scan-related images 346 are filtered by a related third sensitivity map 386 which filtering is carried out by the target detection 394. The sensitivity map 386 comprises a medium subtraction level compared to the following situation 307. Hereby, some of the clutter is reduced. Subsequently, the clutter-reduced maps 387 are processed by three-out-of-three scan correlations 374 replacing each pixel value that is not present in all of the three clutter- reduced maps 387. Finally, the correlated image 375 is adjusted by a third weighting factor 354.

m the four-out-of-four situation 307, the first, second, third and fourth scan-related images 348 are filtered by a related fourth sensitivity map 388 which filtering is carried out by the target detection 396. The sensitivity map 388 comprises a low subtrac- tion level compared to the previous situations 301,303,305. Hereby, some of the clutter is reduced. Subsequently, the clutter-reduced maps 389 are processed by a four- out-of-four scan correlation 376 replacing each pixel value that is not present in four out of the four clutter-reduced bitmaps 389. Finally, the correlated images 377 are adjusted by a fourth weighting factor 356.

Finally, the correlated images are combined by means 398 to generate a final radar output image 399. The weighting factors 350,352,354,356 are used to weight the scan- to-scan correlation relative to each other. Fig. 4 shows a possible combination of a one-out-of-one situation 401 together with a three-out-of-four situation 403 where the correlated bitmaps are individually weighted and combined to form an output image 499.

In the one-out-of-one situation 401, a first scan-related image 442 is filtered by a first sensitivity map 482 which is carried out by the target detection 490. The first sensitivity map 482 has a relatively high subtraction level compared to the following three- out-of-four situation 403. Hereby, most of the clutter is reduced. In the next step of the process, the clutter-reduced map 483 is processed by the scan correlation 470. Finally, the correlated bitmap 471 is adjusted by a first weighting factor 450.

In the three-out-of-four situation 403, the first, second, third and fourth scan-related images 444 are filtered by a related sensitivity map 484 which is carried out by the target detection 492. The sensitivity map 484 comprises a considerably lower subtrac- tion level compared to the previous one-out-of-one situation 401. Hereby, some of the clutter is reduced. Subsequently, the clutter-reduced maps 485 are processed by a three-out-of-four scan correlation 472 replacing each pixel value that is not present in three out of the four clutter-reduced maps 485. Finally, the correlated image 473 is adjusted by a weighting factor 452.

The correlated images are combined by means 498 to generate a final radar output image 499. The weighting factors 450,452 are used to weight the scan-to-scan correlation relative to each other.

The final radar image 499 is hereby a weighted combination of the two correlated bitmaps 471,473. As a result it is possible to detect a fast moving target in the near area of the radar and concurrently detect a weak target.

The following figures uses binary integration, but the invention is not limited to the type of integration used and could in another embodiment of the invention comprise an amplitude integration or the like. The subtraction filtering comprising of subtracting the scan related image by the sensitivity map leaves every bit behind that is greater than zero. In combination with the binary integration the pixels value of the resulting image is binary.

Fig. 5 shows the first scan-related image 542 on a pixel level. To illustrate the filtering process, a row 501 of the first scan-related image is shown. The histogram 505 of the row visualizes the relative pixel amplitudes. The pixel 503 illustrates clutter, the group of pixels 500 illustrates a target, and the pixel 515 illustrates a weak target. The clutter pixel 503 of the row 501 exceeds the value of the sensitivity 507. Hereby, the clutter pixel 512 in the resulting row 509 is preserved as a false target. IQ the resulting row 509, the weak target 511 is removed by the filter. The figure illustrates a situation where the clutter pixel 503,512 is preserved and the weak target 515,511 is erased by the filtering process which is unwanted.

Fig. 6 shows the processing of a first scan-related image 642 and a second scan-related image 644. m the first part of the process, the two scans 642,644 are filtered by a corresponding sensitivity map 607. The process is illustrated by a row of pixels 601,603 of each scan followed by a corresponding histogram 605,606 of the relative amplitude. In this two-out-of-two scan situation, the sensitivity map 607 is lower than the previous sensitivity map 507, as shown in fig. 5. Hereby, the clutter pixel 613 and the weak target 615 of the scan-related image 642 are preserved as a false target in the clutter- reduced row 616 because the result of the subtraction is greater than zero. However, the clutter pixel 619 in the second scan-related image 644 is lower than the clutter pixel 613 in the first scan-related image 642 due to the random nature of clutter. The clutter pixel 619 is not greater than the sensitivity map 607 and the clutter pixel 619 is therefore removed because the result of the subtraction process is lower than or equal to zero. The two out of two scan correlation 672 replaces each pixel that is not present in both scans. Hereby, the clutter pixel 613 is removed because it is not present in both clutter-reduced rows 616, 617. However, the weak target 615 is present in both clutter- reduced rows 616, 617 and is, hereby, preserved.

The clutter is random and do not correlate over time which is why most of the clutter can be removed by a scan correlation. The correlation process is carried out by a bi- nary integration but could in another embodiment of the invention comprise of a amplitude integration process.

Fig. 7 shows the three out of four scan situation where a row of the scan-related im- ages are shown for the first 701, second 703, third 705 and fourth 707 scan-related rows. The figure shows a more detailed description of the three-out-of-four scan situation 403 as shown in fig. 4. To illustrate the principle of the process, a single row of the scan-related images has been chosen.

The scan-related rows 701,703,705,707 are all filtered by a related sensitivity map 707. The sensitivity map 707 is set to a minimum subtraction level relative to the previous sensitivity maps 507,607 as shown in fig. 5 and 6. The pixels 730,731 illustrate clutter, the group of pixels 740 illustrates a target, and the pixels 720,722,723 illustrate a weak target. Most of the clutter is expected to pass the filtering process because of the low subtraction level. To remove most of the clutter, a three-out-of-four scan correlation 772 is performed. The correlation process makes it possible to remove most of the clutter 730,731 and at the same time detect a weak target 720,722,723 which is not present in all of the four scans. The target 740 and the weak target 720,722,723 are, hereby, preserved in the output row 780 as targets 781,782.

Claims

1. Radar system comprising an A/D converter (180) from which a digital signal is supplied to processor means (100), which processor means (100) are connected to at least one memory (120), in which memory (120) the processor means (100) generate and store at least a first and a second digital scan-related image (242,342,344,346,348,442,444,542,642,644), which scan-related images (242,342,344,346,348,442,444,542,642,644) comprise a number of independent pixels, which processor means (100) correlates the scan related images (242,342,344,346,348,442,444,542,642,644) characterized in that the processor means (100) store N independent scan-related images (242,342,344,346,348,442,444,542,642,644) in the memory (120,122), which images (242,342,344,346,348,442,444,542,642,644) are stored in a true motion format, where the processor means (100) generates at least one sensitivity map based on input data (132,242), where the processor means (100) perform M correlations between N independent scan-related images (242,342,344,346,348,442,444,542,642,644) by use of at least one sensitivity map (382,384,386,388,482,484) during a mathematical calculation of the scan-related images (242,342,344,346,348,442,444,542,642,644), which processor means (100) further process the M correlated bitmaps for generating an image (399,499) for presentation.

2. Radar system according to claim 1 characterized in that the processor means (100) use input data from manual adjustment means (132) for generating at least one sensitivity map (382,384,386,388,482,484).

3. Radar system according to claim 1 characterized in that the processor means (100) use present and/or previous scans (242,342,344,346,348,442,444,542,642,644) for generating input data for generating at least one sensitivity map (382,384,386,388,482,484).

4. Radar system according to at least one of the claims 1-3 characterized in that the processor means (100) store the first, a second and a third independent scan-related image (346) where a logical calculation (374) of related pixels from the first, second and third scan-related images (346) are used by the processor means (100) for generating pixels of a third correlated bitmap (375).

5. Radar system according to at least one of the claims 1-4 characterized in that the processor means (100) store the first, the second, the third and a fourth independent scan-related images (348), where a logical calculation (376) of related pixels from the first, second, third and fourth scan-related images (348) are used by the processor means (100) for generating pixels of a fourth correlated bitmap (377).

6. Radar system according to at least one of the claims 1-5 characterized in that the logical calculations are made at corresponding pixels at the first, the second, the third and the fourth scan-related image (444) where each pixel is correlated in three out of four corresponding pixels for generation of the pixel, where a correlated image (473) is created on the basis of generated pixels.

7. Radar system according to at least one of the claims 1-6, characterized in that the radar system comprising at least a first, second, third and fourth sensitivity map, which sensitivity map (382,384,386,388,482,488) are stored in the memory (120,124), where an update of the pixel values is performed by the processor (100) so that each sensitiv- ity map (382,384,386,388,482,488) controls the sensitivity in each generation of correlated bitmaps (371,373,375,377,471,473).

8. Radar system according to claim 7 characterized in that the radar system comprising at least a first, second, third and fourth sensitivity map, which sensitivity map (382,384,386,388,482,488) are stored in the memory (120,124), where an update of the pixel values is performed by the processor (100) so that the sensitivity maps (382,384,386,388,482,488) has a first low value where the second sensitivity map (384) has a lower value than the first sensitivity map (382), where the third sensitivity maps (386) has a lower value than the second sensitivity maps (384), where the fourth sensitivity maps (388) has a lower value than the third sensitivity maps (384).

9. Radar system according to at least one of the claims 1-8 characterized in that the pixels in the first, second, third and fourth scan-related images (242,342,344,346,348,442,444,542,642,644), are filtered by individual sensitivity maps related to each scan-to-scan correlation.

10. Radar system according to at least one of the claims 1-9 characterized in that the correlated bitmaps (371,373,375,377,471,473) are multiplied by a weighting factor

(350,352,354,356,450,452) which corresponds to the respective correlated bitmap (371,373,375,377,471,473) for generating an output image (399,499), which image (399,499) contains information from a number of independent correlated bitmaps (371, 373, 375, 377,471,473).

11. Radar system according to at least one of the claims 1-10 characterized in that the pixels of the first scan (242,342,344,346,348,442,444,542,642,644) provide the basis for an update of the corresponding pixels for a clutter map (264) which clutter map (264) is based on statistics of the pixels in the first and/or previous scans and the corresponding pixels in the clutter map (264) .