RussianPatents.com

Method of detecting objects

IPC classes for russian patent Method of detecting objects (RU 2331084):

G01S17/06 - Systems determining position data of a target

Another patents in same IPC classes:

Instrument for daytime and night observation and aiming / 2310219
The instrument has a front-view mirror with a system of stabilization and aiming of the sighting line optically linked with the sight channel, which includes an optical module including the daytime and night channels. The instrument has also a radiator unit, scanning unit, photodetector unit, the output of the photodetector unit is connected to the input of the analog-to-digital converter, whose output is connected to the control electronics unit.

Method for determining direction to source of optical radiation on basis of component, dissipated in atmosphere, and device for realization of said method / 2285275
In accordance to method, second optical-electronic coordinator with matrix photo-receivers is additionally installed, field flatness of which is perpendicular to field plane of first optical-electronic coordinator, coordinate alignment of photo-elements of first optical-electronic coordinator is performed in coordinate plane x0z and of photo-elements of second optical-electronic coordinator in coordinate plane y0z, determining of angular coordinates of optical radiation source on basis of formulae where ε, β - tilt angle and azimuth of optical radiation source; d - distance between upper and lower lines of photo-elements of optical-electronic coordinator; Δx=x_1B-x_1H, Δy=y_2B-y_2H; x_1B and x_1H - coordinates of upper and lower photo-elements of lines of first coordinator, signal at output of which has maximal value; y_2B and y_2H - coordinates of upper and lower photo-elements of lines of seconds coordinator, signal at output of which has maximal value. Device for realization of method consists of first and second optical-electronic coordinators with matrix photo-receivers, first and second subtraction blocks, first and second square-law generators, first and second dividers, adder, square root extractor block, first and second arctg calculation blocks, outputs of which are outputs of device, while first and second outputs of optical-electronic coordinator are connected respectively to first and second inputs of first subtraction block, first output of which is connected to input of first square-law generator, second output - to first input of first divider, first and second outputs of second optical-electronic coordinator are connected respectively to first and second inputs of second subtraction block, first output of which is connected to input of second square-law generator, second output - to second input of first divider, output of first divider is connected to input of first arctg calculation block, outputs of first and second square-law generators are connected to appropriate inputs of adder, while output of adder is connected to input of square root extractor block, output of which is connected to first input of second divider, onto second input of which value d is fed, output of divider is connected to input of output arctg calculation block.

Method for detecting optical and optical-electronic surveillance means and device for realization of said method / 2278399
Method includes simultaneously receiving signals of natural background radiation in spectral range of reflected laser radiation, and receiving continuous optical signals in spectral range of natural background radiation from observed volume of space, after transformation of received optical signals to video signal from first video signal, appropriate for reflected laser radiation, video signal is subtracted element-wise, appropriate for optical signal of natural background radiation in spectral range of reflected laser radiation, threshold selection of difference video signal is performed, and from the number of video signals, exceeding the threshold, video signals are selected, appropriate for code-pulse modulation of laser radiation and if such signals are present, alarm signal is generated, simultaneously from difference video signal, exceeding the threshold, and video signal, appropriate for continuous optical signal in spectral range of natural background radiation with consideration of parallax, and after transformation of composite video signal to optical image, position of detected optical means is fixed relatively to characteristic objects on the map of local area.

Method for one-positional measurement of laser emission source coordinates and device for realization of said method / 2269795
In known method for one-positional measurement of angular coordinates of laser emission source, based on use of scanning optical systems and determination of angular coordinate of laser emission source on basis of angular position of scanning system at the moment when maximum value of useful signal is achieved, scanning of vicinity in selected observation sector is performed due to alternation of heterodyne signal phase front position, and angular coordinate of laser emission source is determined on basis of its angular position at the moment when maximum value of useful signal is achieved. Device for realization of method consists of photo-detector mounted on rotary platform with forming optics and of semi-transparent, held, mounted at an angle of 45° to optical axis, plate, block for processing signal at output of photo-detector, heterodyne, and also of serially connected block for forming parameters of phase front of heterodyne signal and block for controlling phase front of heterodyne signal, while second output of block for forming parameters of phase front of heterodyne signal is connected to second output of signal processing block, second input of block for controlling phase front of heterodyne is connected to output of heterodyne, and output - to second input of photo-detector.

Method of detecting objects / 2250478
Method comprises subtracting reference and current images, breaking the image series to be processed into fragments, and converting the characteristic features of the images into signals. The signals from one of the images are recorded as reference ones and are compared, e.g., by subtracting, with corresponding current signals, and, after the threshold processing, the difference signals obtained are converted into the binary signals for control of spatial filtration . As a result, the fragments of the current image, for which the control signals exceed the threshold, are transmitted, whereas the fragments, for which the signals are equal or less than the threshold value, are suppressed.

Method for detection of sources of laser radiation and counteraction of their receiving systems / 2248587
The device has a frequency-pulse laser with an objective lens secured on platforms rotating about the vertical axis, detection unit consisting of two detection sensors, and a control unit having k-flip-flops of the first and second detection sensors, k-AND gates of the first and second detection sensors, two NO gates two OR gates, two coding units, two permanent storage units, master oscillator, two arithmetic-logical devices, initial data input unit, azimuth computing unit, elevation angle computing unit and a laser radiation source height computing unit.

Method of detecting objects / 2331084
Invention pertains to optical methods of detecting foreign objects on a complex dynamic variable background in a surveillance zone. The technical outcome of the invention is the increased accuracy of detecting objects with simultaneous increase in speed and visibility. The method involves receiving and forming two images on two points spread in space and single registration of each fragment of the reference and compared digital images by two identical viewing systems. Each viewing system is made based on a multi-element high-speed light detector, for example, a CMOS - matrix, and contains an objective. Both viewing systems are fitted at a shorter, compared to the displacement from the presumed location of the object, distance between them, parallel each other and directed at the surveillance area. Analysis of the images is done by determining the displacement value Δ of characteristic fragments of the compared image with the corresponding fragments of the reference image with their maximum possible coincidence in the direction of parallactic shift with subsequent identification of the selected and background objects from the obtained displacements Δ. Based on the obtained displacements Δ, three dimensional images of objects can be formed and the distance to each one of them can be determined.

Method of object detection / 2338222
High efficiency method of object detection is developed. It includes preliminary detection of object search zones, image scanning with the help of scanning window within the limits of the search zones, the size of the scanning window corresponds to the size of an object's image, quantity determination of amplitude intervals N by the square of the scanning window, taking into account the most significant amplitude intervals and choosing the coordinates with maximum value N as the coordinates of the object. The claimed invention increase probability of correct object detection within lower contrasts range of the object and higher contrast range of background.

Device for definition of angle coordinates of pulsed laser radiation source / 2352959
Invention is related to instruments for determination of angle coordinates of pulsed laser radiation sources and may be used in protection of different objects against directed aiming. Device comprises control unit, photodetector on the basis of photodiode, two superwide-angle lenses, in focal plane of which two light-sensitive matrices are installed accordingly; time for accumulation of signal of every light-sensitive matrices, on the one hand, is so insignificant, that it provides for absence of day background, and on the other hand, it exceeds time for reset of accumulated signal by second matrix, which provides for continuity of space viewing; control unit connected to the first and second light-sensitive matrices and with photodetector device, provides for alternate accumulation of signal by light-sensitive matrices and determines angle coordinates of pulsed laser radiation source and time of signals appearance at the outlet of photodetector; light-sensitive matrix in mode of signal accumulation at the moment of signal appearance at photodiode outlet, on completion of accumulation mode transferring into mode of signal reading by control unit; at that analysis of time-pulse modulation of signal at the outlet of photodetector may be used for recognition of friend or foe.

Method of processing information in coherent laser locator with photodetector array / 2354994
Present invention relates to measuring techniques and instrument making and can be used in laser Doppler location of stealth objects flying at low altitudes above water basins. The method of processing information in a coherent laser locator with a photodetector array is based on reception of laser radiation from glare of the sea surface, arising when probing radiation is scattered by the stealth object. The current location of an object and its velocity vector can be reconstructed through measurement of angle of arrival of radiation from the glare of the sea surface using a photosensitive reception matrix and through measurement of Doppler frequency shifts in a multi-channel unit for optimum filtration based on heterodyne reception methods using multi-channel dispersive delay lines using statistical averaging methods. Cutting on the number of information processing units is achieved due to creation of two- or three-dimensional groups of elements of a photodetector array, connected to information processing channels. In the processing channel, the signal is converted to a linear-frequency-modulated equivalent with subsequent amplification, spectro-time "compression" in the dispersive delay line, detection and minimum threshold cutting with a given threshold value, which allows for converting a signal with Doppler frequency shift to a short pulse, the time position of which, relative the strobe-pulse for the beginning of the measuring cycle, uniquely characterises the value of the given Doppler frequency shift. This time position of the pulse is coded in a digital code and stored in the corresponding buffer memory of a memory device, in the code of which there is also a code of the number of the measurement cycle and the code of the number of the channel, on which the signal from an element of the photodetector array was processed. From the set of such code records in the given measurement cycle, information is obtained on Doppler frequency shifts in signals of corresponding elements of the array and the position angle on the glare of the sea surface, detected by the locator in the given measuring cycle relative the optical axis of the receiving-transmitting objective of the locator, as well as scatter angles of the probing radiation of the stealth object, generating the said glares. If conditions are met for detecting an object and its bearing auto-tracking, where the inclined range line and the optical axis of the receiving-transmitting objective of the locator lie in the same plane, location of the object and measurement of its radial velocity is done through calculation, using a minimum of two different reflected radiation in a given measurement cycle, based on the method of overlapping circles. The radiation pattern of the locator is fan-shaped - wide on the position angle and extremely narrow on the azimuth.

FIELD: physics; optics.

SUBSTANCE: invention pertains to optical methods of detecting foreign objects on a complex dynamic variable background in a surveillance zone. The technical outcome of the invention is the increased accuracy of detecting objects with simultaneous increase in speed and visibility. The method involves receiving and forming two images on two points spread in space and single registration of each fragment of the reference and compared digital images by two identical viewing systems. Each viewing system is made based on a multi-element high-speed light detector, for example, a CMOS - matrix, and contains an objective. Both viewing systems are fitted at a shorter, compared to the displacement from the presumed location of the object, distance between them, parallel each other and directed at the surveillance area. Analysis of the images is done by determining the displacement value Δ of characteristic fragments of the compared image with the corresponding fragments of the reference image with their maximum possible coincidence in the direction of parallactic shift with subsequent identification of the selected and background objects from the obtained displacements Δ. Based on the obtained displacements Δ, three dimensional images of objects can be formed and the distance to each one of them can be determined.

EFFECT: increased accuracy of detecting objects with simultaneous increase in speed and visibility.

6 cl, 3 dwg

The invention relates to localization (selection), in particular passive optical methods of selection (detection) of foreign objects on the complex contrast dynamically changing background in the protected (controlled) area.

Known way to select objects according to the patent of Russian Federation №2081439 from 27.09.84, IPC G01S 17/06, which is the formation of images of objects and background in the plane of the corresponding a priori the distance to the object, and optionally, in the plane that is offset from the first plane, with the subsequent mathematical processing. However, this method, as it follows from the description of the patent is applicable only in the case of the use of the telescope or telephoto lens, resulting in extremely low visibility and, consequently, low speed when the control of the entire protected area, and not only narrow lines. All this leads to a low detecting ability in General. In addition, a fast moving object or a sharp change in the background if no special measures for overall improvement of the performance of the tract of registration (for example, simultaneous recording of all pixels of the image will blur and inefficiency of the method applied mathematical processing, the contours of the object can be closed or even partially ice the us.

There is a method of object detection in complex backgrounds, including operation of the preprocessing of the signals represented by the temporal sequence of images of a scene, where the appearance of the object, storing the reference signals and the subtraction of the respective current signals corresponding reference signals, and the processing algorithm involves dividing each image into fragments, the measurement of the characteristics of each fragment, such as optical density, memorizing the corresponding signals, and after the subtraction of the signals of the comparison of the obtained difference with a given threshold value, generation of control signals, performing granular filtering time sequence of the current image to pass only those differential signals, the absolute values of which exceed the threshold value (U.S. Pat. Of the Russian Federation No. 2250478 from 11.04.2003, IPC G01S 17/06).

The disadvantage of this method is the complexity of the algorithm processing the differential image and the restriction of its use, if necessary, the detection of rapidly moving object, since for a given processing algorithm will not be enough data for reliable results.

As a prototype of the selected method of selection of an object on a remote background on the patent of Russian Federation №2081435 from 02.04.84, IPC G01 17/00, consists in the admission and formation of two images at two spatially separated points, simultaneous recording of formed images, determining the parallactic displacement of the background by forming a mutual correlation function of the two registered images and determining its maximum, the offset of the first of the registered images by the value of the parallactic displacement of the background in the opposite direction of this offset, obtaining a differential image by subtracting the offset and the second registered image, dividing the differential image into regions having opposite signs, and the analysis of fragments of the fields.

The known method has the following disadvantages.

1. Limited functionality, because the known method does not allow to reliably detect low-contrast object on a high-contrast heterogeneous background, such as the sun, clouds, even with a slight difference of spatially-separated optical systems for the formation and registration of images, for example, caused by operating conditions or technological error in manufacturing. As a result, the desired object can be taken by random fluctuations in the differential image.

2. Limited applicability, as were elaborated to the operation of the known method is not designed for background objects, are comparable with the target object distance, and not at an infinite distance.

3. The relatively low selectivity in the case of the simultaneous registration of images of a group of objects when the application of the algorithm of this method may lead to selection "false" objects.

The task of the present invention is to improve the accuracy of detection (selection) of one or more objects in high-contrast conditions relatively closely spaced objects in the background while increasing performance and visibility.

The problem is solved as follows.

In the method of selection of the object on the remote background, which consists in the admission and formation of two images at two spatially separated points, simultaneous recording of images formed according to the invention reference and compare digital images recorded simultaneously for each piece of image two identical systems based on multi-element high-speed photodetectors, such as CMOS sensors with lenses that are pre-fixed small compared with the distance from the proposed appearance of the object, the distance between them parallel to each other in the direction in controlled space, and image analysis the deposits made by the estimate bias Δ characteristic fragments of the comparison image with the same reference fragments, as much as possible their coincidence in the direction of the parallactic displacement and subsequent identification of breeding and background objects from the obtained displacements Δ.

The technical essence of the invention consists in the application of the effect of binocular vision, similar to human vision, when each of the video performs the function of a private eye, receiving the image of the object is shifted by a certain amount relative to the distant background. The shift is greater, the closer the object. As in the human vision to detect (detection) of an object is the combination of his images posted a video to match, while the background image is blurred.

In addition, to increase awareness of the kind of the detected object, offset Δ build in the form of three-dimensional distribution relative to the reference image, analyze the connectivity of the obtained three-dimensional images of objects, their geometrical dimensions and location relative to each other in a controlled environment.

In addition, it may be determined the distance to the detected object R from the expression R=L·A/Δ, where L is the distance between the centers of the lenses of the video, And the distance from which idiomatic to the main optical axis of the lens system.

In addition, to reduce the processing time allowed to determine offsets Δ not all of the characteristic fragments of the comparison image with the same reference fragments, and optionally, a specified step.

In addition, to reduce the processing time allowed to determine offsets Δ offsets determined at the previous processing time of a pair of digital images.

In addition, to improve the detecting ability of the proposed method and simplify the analysis of three-dimensional displacement distribution Δ Pets preliminary determination of the calibration of the three-dimensional displacement distribution Δ for the case of background objects, obviously considerably remote from the controlled area, and subtract this calibration distribution of the current.

Figure 1 shows a schematic arrangement of the two video systems and the object relative to the background. Video contain multi-element CMOS sensors 1 and 2 and the lenses 3 and 4, which are completely identical to each other. Matrix and the lens is fixed on the base 5 at a small distance L relative to each other. The outputs of the matrices 1 and 2 are connected to the unit 6 of the information processing.

Figure 2 shows how each of the pixels 7 of the reference digital image 8 is allocated rectangular FR is gment (matrix) 9 reference image, which is searched for similar or matching a similar fragment (matrix) 10, but in the compared image 11 in the direction of the parallactic displacement of 12, starting with the pixel 13 in the comparison image corresponding to the current pixel 7 of the reference image.

Figure 3 images of the objects 14, 15 and 16 in the bearing 17 and compare 18 digital images showing the construction of the three-dimensional distribution 19 offsets Δ relative to the reference image 17.

The method is implemented as follows.

The beginning of the observations associated with the absence of the object, characterized by the registration only of the background of each of the two systems at the same time. Remote background, as a rule, is a combination of a dynamically resizable sky through clouds moving in different directions, and more stable land cover (although the stability of its contingent by wind, causing the fluctuation of vegetation). The object can appear in a known location and move in a wide range of speeds. The object may move in different directions, including chaotic or observed in nature. An object can have different sizes and can be protected by camouflage. Check both the reference and compare digital images produced with a given frequency, the definition is caused by the characteristics of CMOS sensors and processing unit. At the moment of appearance of the object in the control area (area) both CMOS sensors 1 and 2 will register the image of the object on the remote background (figure 1) with different offset relative to the background.

To determine offsets Δ and, as a consequence, the distances R to be recorded by the images of objects and their fragments, including background, for each pixel 7 of the reference digital image 8 is allocated, as shown in figure 2, the rectangular piece (matrix) 9 reference image, which is searched for similar or matching a similar fragment (matrix) 10, but in the compared image 11. To reduce the time of account matching is performed in the direction of the parallactic displacement of 12, starting with the pixel 13 in the comparison image corresponding to the current pixel 7 of the reference image. To reduce the time of account matching is not performed, if the rectangular piece (matrix) 9 reference image does not contain significant heterogeneity (detail), for example, is homogeneous. The matching rectangular slices (matrix) 9 and 10 can be determined, for example, as a maximum in the mutual correlation function of the data fragments in the direction of the parallactic displacement 12. Defined this way offsets Δ for the pixels of the reference image are constructed in the form of three-dimensional distribution as shown in figure 3, where the images of the objects 13, 14 and 15 in the bearing 16 and compare 17 digital image corresponds to a three-dimensional distribution 19. In the future, an analysis of connectivity allocated on a three-dimensional distribution of the 19 objects, defining their geometric dimensions and location in a controlled environment. Thus, localizing (detection) of all objects, including the background. Simultaneously determines the distance to each of the detected objects according to the formula given above.

This method provides a relatively high speed, since it is possible to eliminate processing those fragments that do not contain significant for the comparison of parts of images, they are simply omitted. Such significant fragments is necessary so that you will only define the contours of the object, and this is enough to distinguish one object from another and to draw conclusions about coming from danger. Performance can be increased by limiting the number of processed fragments, selecting them with the given step.

Pets also carry out the calculation of intercorrelation functions in the field of offset Δdefined during the processing of the previous pair of digital images, it also reduces the time of processing.

In the preliminary determination kalibrovochnoj the three-dimensional distribution of displacements Δ for the case of background objects, obviously considerably remote from the controlled area, and the subtraction of this calibration distribution of current the opportunity significantly to take into account the identity of the optical paths of registration aberration and distortion.

The absence of the desired fragments of the reference image in the comparison image correspond to the output of an object or part of a controlled area, which begins after the nearest dead zone and ends at the border of the background objects that can create the desired shading of objects. The shading of the object it is also possible other objects from the group, but it is rare (due to possible shading only in the selected direction) and short (due to the motion of the objects). Continuous tracking of a group of objects above the shading will be perceived as a local short-term loss goals.

The inventive method is realizable as used technical tools already developed by the industry of developed countries. The creation of software is a completely solved problem, because the techniques of image processing to date worked.

The method will increase the reliability and efficiency of modern means about which Eugenia those objects, for which various measures have been secretive appearance in areas of control.

1. The method of selection of the object on the remote background, which consists in the admission and formation of two images at two spatially separated points, as well as simultaneous recording of the generated digital image, characterized in that the supporting and compare digital images recorded simultaneously for each fragment (pixel) images of two identical systems based on multi-element high-speed photodetectors, such as CMOS sensors with lenses that are pre-fixed small compared with the distance from the proposed appearance of the object, the distance between them parallel to each other in the direction in controlled space, and the image analysis is performed through determination of the offset values Δ characteristic fragments compared images with the same reference fragments, as much as possible their coincidence in the direction of the parallactic displacement and subsequent identification of breeding and background objects from the obtained displacements Δ.

2. The method of selection of an object on a remote background according to claim 1, characterized in that the offset Δ build in the form of three-dimensional distribution relative to the reference image, analyze swathes the ü obtained three-dimensional images of objects, their geometrical dimensions and location relative to each other in a controlled environment.

3. The method of selection of an object on a remote background according to any one of claim 1 or 2, characterized in that the distance to breeding or background objects R is determined from the expression R=L·A/Δ, where L is the distance between the centers of the lenses of the video, And distance from the video to the main optical axis of the lens system.

4. The method of selection of an object on a remote background according to claim 1, characterized in that the determination of the values of the displacements of the characteristic fragments of the comparison image with the same reference fragments carried out selectively, at a specified step.

5. The method of selection of an object on a remote background according to claim 1, characterized in that the determination of the values of the displacements of the characteristic fragments of the comparison image with the same reference fragments is carried out in the field offsets determined at the previous processing time of a pair of digital images.

6. The method of selection of an object on a remote background according to any one of claim 2 or 4, wherein the pre-determined calibration three-dimensional distribution of displacement Δ for the case of background objects, obviously considerably remote from the controlled area, and subtract this calibration distribution of the current.