DE3018607A1 - Aircraft navigation system - uses simple correlation of surface images with several comparison images with different grey stale thresholds - Google Patents

Abstract

The aircraft navigation system has two memories for two-dimensionial reference and measured images of the earth's surface in digital form. The position of the aircraft w.r.t. the surface is determined by correlation of the reference and measurement images. A simple correlation procedure is used. For each individual reference image point only the information black or white is stored. Several comparison images are derived from the sensor output. The points of these images are also stored as black or white. The different comparison images have different grey scale threshold levels to distinguish between black and white points. At least one comparison image is correlated with the reference. The correlation function used is that for which the main to side lobe ratio is a maximum.

Description

Missile navigation device

PRIOR ART The invention is based on a navigation device for missiles as specified in the preamble of claim 1. Such a navigation device is from IEEE Transactions on Aerospace and Electronic Systems, Volume AES-15, No.1, January 1979, pages 11-19.

A comparison is made for on-board autonomous navigation of missiles (Correlation) of reference images carried along with those recorded during the flight Sensor images. The accuracy of the position determination increases with the number of correlated pixels. To add a 5 pixel to the entire gray tone information several bits per pixel are necessary. This has the consequence that the Computing effort and computing time become very large.

If you reduce the number of gray tones, information is lost.

To continue from the correlation function the exact deviation of the spatial position of the location of the sensor image from the location of the location of the reference image To be able to determine, the sensor image and the reference image should be the same Environmental conditions are included, but this is only very rarely the case.

Problem It is the object of the invention to provide a navigation device for missiles indicate in a simple way a correlation between reference and sensor image is carried out.

Solution The solution to this problem takes place with those specified in claim 1 Means. Advantageous further developments can be found in the subclaims.

Advantages Because there is only one black / white information item for each pixel is saved, only a relatively small amount of memory is required.

It is not necessary to normalize the reference and sensor images.

The circuit for realizing the correlation is very simple and the correlations are carried out very quickly.

The reference images used can be adapted to the direction of flight and turn it slightly to the flight altitude and distort it linearly.

The reference images used can be synthesized from maps and generate satellite images, which is possible when using multiple shades of gray per Pixel would not be possible.

Description The invention is illustrated by way of example with reference to the drawings explained in more detail. It shows: FIG. 1 a block diagram of the navigation device, FIG. 2 Details of the block diagram according to FIG. 1, FIG. 3 a representation of the one another pixels to be compared, and FIG. 4 further details of the block diagram according to Fig.1.

The navigation device contains a sensor 1 from that of the missile takes pictures over flown terrain. A camera (Charge Coupled Device camera) with residual light amplifier (wavelength range from 800 mm to 1.1 m) be used. The captured image is divided into individual pixels that form columns and rows are arranged, divided. Each pixel can have different shades of gray exhibit.

The digitalization of the takes place in an analog / digital converter 2 individual pixels. The digital values are stored in a sensor image memory 3. From this memory 3, the pixels of the recorded sensor image can be read out as often as required. The full gray value information is available.

The memory 3 is followed by a controllable threshold value circuit 4, which carries out a dualization of the individual pixels. Under dualization-will understood the following: there is a threshold value, which is a certain shade of gray is equivalent to. If a pixel is darker than this shade of gray, then this pixel becomes processed as "black"; if it is lighter, it is processed further as "white". Instead of several shades of gray, only the information is black for each pixel or assigned white. The image obtained in this way is referred to as the comparison image It is easy to see that depending on the choice of the threshold value, different Receives comparison images. The comparison images are stored in a comparison image memory 5 saved.

The reference image is stored in a reference image memory 7 and also i-n dualized form. Like the comparison images, it becomes a correlator 6 supplied. In it, the cross-correlation function KKF between the reference image and comparison image formed. From the position of the main tip of the cross-correlation function can be in an evaluation device 9, the deviation of the spatial position of the Determine the location of the comparison image from the location of the reference image. This deviation X in turn corresponds to the deviation of the missile from its prescribed position in Direction of flight and orthogonal to it. A clock generator 10 is also provided, whose clock pulses the comparison image memory 5 and a control unit 8, whose task is explained in more detail below, are supplied to the generation of the dualized reference image, a threshold value is used that only contains the significant and extracted image elements invariant to weather influences. It is possible, before storing the reference image with several threshold values dualizations perform and the autocorrelation functions of the reference images generated in this way to build. That dualized reference image is then stored in memory 7, where the ratio of the main to the secondary lobe of the autocorrection function is greatest.

As will become apparent in the course of the description, it is advantageous to if in the navigation device, besides the dualized reference image, the dualized reference image the threshold value used for the reference image is saved.

The dualized reference image has 64x64 pixels. The dualized The comparison image consists of 256x256 pixels.

The dualized reference image can be read into the correlator rotated and / or linearly according to the yaw angle and altitude of the missile be distorted. In this way, an optimal adaptation is achieved in an advantageous manner to the comparison picture.

The number of stored reference images depends, among other things, on the length the flight path and the drift of the inertial system that the navigation device described is used.

To form the cross-correlation function between the dualized reference image and a dualized comparison image, the reference image is gradually displayed in line and column direction shifted over the comparison image.

For the sake of simplicity, it is assumed that the comparison image from the Image points arranged in columns and rows S11 S12 S13 S14 S21 S22 S23 S24 S31 S32 S33 S34 541 S42 S43 S44 and the reference image from the pixels R11 R12 R21 R22 exists.

With the pixels shown above, there are nine settings for correlation possible and necessary.

1 .. Reference image is on S33 S34 S43 S44 2. Reference image is on S32 S33 S42 S43 3rd reference image is on S31 S32 S41 S42 4th reference image is on on S23 S24 S33 S34 9. Reference image is on S11 S12 S21 S22.

For each of the nine settings, the four pixels of the reference image must multiplied by the image points of the comparison image and the results added up will. With setting "1" Z - R11-S33 + R12tS34 + R21-543 + R22jS44 must be calculated will. Since the pixels are either black or white, they are limited Multiplication on an EXCLUSIVE OR link.

The correlator shown in FIG. 2 therefore contains four EXCLUSIVE-OR circuits 28, the output signals of which are fed to an operational amplifier 27 for summation will. The time sequence of the output signals of the operational amplifier is Cross-correlation function KKF.

The reference pixels are in shift registers 21, 22 and the comparison pixels stored in shift registers 23, 24, 25 and 26.

In Figure 2, the state is shown in which the entire first Comparison image with the image points S stored in the shift register is. A line of pixels is contained in each shift register.

In this state, the four EXCLUSIVE-OR circuits 28 are as follows Storage locations of the shift registers connected: First EXCLUSIVE-OR circuit with R11 and S33 Second EXCLUSIVE-OR circuit with R12 and S34 Third EXCLUSIVE-OR circuit with R21 and S43 Fourth EXCLUSIVE-OR circuit with R22 and S44.

With this setting, the output signal of the operational amplifier has 27 the value Z.

The respective content of the comparison image memory is shown in FIG shown in boxes. The individual lines in the boxes indicate the content of the individual shift registers 23 to 26. The box at the top left indicates the state in which all image points S of the first comparison image are currently stored. This memory state is also assumed in FIG.

The contents of the shift register are increased by one with each shift register cycle Position pushed further, whereby each pixel of the first comparison image is deleted and gradually the pixels T of the second comparison image are stored will. The individual comparison images are displayed one after the other without any spacing the shift register clocked through. The box at the bottom right indicates the state in which only the second comparison image is still contained in the shift registers is.

With the correlator according to FIG. 2, the image points of the comparison images are always compared with the reference image, which are indicated by the small boxes.

From FIG. 3 it can be seen that the -nine Settings result automatically when the shift registers are clocked through. Here, however, there are also some comparisons that -n-are not evaluated to be allowed to. However, this can be eliminated by a -simple- algorithm. at The large boxes in FIG. 3 do not contain the undesired settings small box.

It is assumed that the comparison image from K # L and the reference image consists of k-l pixels.

For correlating a k.l reference image with a K, L comparison image must be connected to the serial, isochronous comparison image memory 5 K # L shift clocks are given. This results in K # L settings (correlation patterns) at the correlator. However, the k, 1 reference image can only be used within the K, L comparison image in (K-k + 1) - (L-1 + 1) different positions can be brought. The serial shifting of the comparison image the comparison image memory results in KL - (K-k + 1) # (L-l + 1)] inadmissible correlation patterns, during which the cross-correlation function must not be evaluated ~. The inadmissible ones Correlation patterns can be easily identified by counting the shift clock for the comparison image memory. Counting starts when a ^ comparison picture - is completely in the comparison image memory. This is in Figure 3 at the box the case at the top left.

The first (K-k + 1) shift cycles deliver permissible, then k-1 follow impermissible correlation pattern (counter ZL - 1). This is repeated until ZL = (-L-1 + 1-). From this counter onwards, there are only impermissible ones up to and including ZL = L Template.

The control for the separation between permissible and impermissible Patterns is explained with the aid of FIG. The control device contains two counters ZK, ZL and two comparators 41, 42, the counter positions with the two constants Compare K-k and L-l. Counter ZK is on Modulo K counter. To it sends a counting pulse to counter ZL every K shift clocks. Both counters will set to 0 if a comparison image is complete in the comparison image memory is stored.

Inadmissible correlation patterns occur when ZK SK-k or ZL L-l is. In order to prevent further processing of impermissible samples, the Output signals of the comparators 41, 42 are fed to an OR circuit 43 which, if there is an output, passing the correlation product from the Correlator 6 to evaluation device 9 blocks.

If there is a permissible correlation pattern, there is a maximum in the cross-correlation function then ZK and ZL simultaneously represent the shift coordinates of the reference image in row and column direction (and thus in flight direction and orthogonal to it) represents. -zK indicates - by how many pixels the reference image in the line direction, zL by how much the reference image is shifted in the column direction.

Leave the shift coordinates, i.e. the mentioned deviation are then determined particularly precisely in an advantageous manner when the cross-correlation function that comparison image is evaluated in which the relationship between main to side lobes is greatest.

This comparison image is obtained particularly quickly if one is from Threshold values that are in the range of the threshold value that is used to generate the reference image was used Blank page

Claims (4)

Claims 1. Naviyationsgerät for missiles, which has a first Contains memory in which at least one image of a part of the earth's surface, over the flight path of the missile is digitized as a two-dimensional reference image is stored, of which when approximately reaching this part of the earth's surface an image is recorded by means of sensors, which is temporarily stored in a second memory is stored and correlated with the reference image, and in which from the in this Comparison obtained correlation function the positional deviation of the location where from the image was recorded by the sensors from the location assigned to the reference image it is determined, since you do not know, that for individual pixels of the reference image in the first memory (7) only the information black or white is stored that from the image recorded by the sensor (1) several Comparison images are obtained for whose pixels in the second memory (5) only the information is stored in black or white, whereby the individual Differentiate comparison images in that the decision criterion is whether a pixel is black or white, different gray tone thresholds are used that at least some of the comparison images are correlated with the reference image, and that the correlation function is used to determine the deviation, in which the ratio of main to side lobes is greatest. 2. Navigation device according to claim 1, d a d u r c h g e k e n n -z E i c h n e t that the gray tone threshold values used in the generation of the comparison images (4) lie in the area of the middle gray tone of the reference image. 3. Navigation device according to claim 1, d a d u r c h g e k e n n -z e i n e t that the reference image covers a smaller part of the earth's surface than that Comparison images shows that the black / white information of the pixels of the reference image in a first shift register (Fig.1,7; Fig.2,2i, 22), which is at each storage location has a tap, that the black / white information of the Image points of the comparison images in columns or lines by at least one other Shift register (23, 24, 25, 26) are clocked through that the further shift register has the same number of taps as the first shift register, since (3 to form of the correlation functions of the reference image and comparison images, the taps in each case of the corresponding pixels of the reference image and part of a comparison image an EXCLUSIVE-OR circuit (28) and the output signals of this EXCLUSIVE-OR circuits are summed up (27). 4. Navigation device according to one of claims 1 to 3, d a d u r c h e k e n n n n e i n e t that, in addition to the reference image, also the one during the creation the threshold value used for the reference image is stored.