CN102508251B - Method for rapidly implementing sector conversion in multi-beam image sonar - Google Patents

Abstract

The invention provides a method for rapidly implementing sector conversion in a multi-beam image sonar. An acoustic signal is radiated by a transmitting transducer by using the multi-beam image sonar; a target echo is received by a receiving transducer; beam forming processing is performed on the echo data to acquire a target scene image output in a polar coordinate mode; and a sector conversion module is used for conversing the acoustic image of the polar coordinate into the acoustic image of a rectangular coordinate to conform to the visual sense habit. When the sector conversion is realized on the image sonar based on the beam forming, triangular function operation or anti-trigonometric function operation is not needed in the hardware such as field programmable gata array (FPGA). The method comprises the steps of: (1) identifying the resolution ratio of a display used for displaying the acoustic image of a rectangular coordinate; (2) calculating the resolution ratio of horizontal and vertical distances of the image of the rectangular coordinate according to the actual range of the horizontal and vertical distances of the image sonar; (3) calculating the values of the horizontal and vertical coordinates of the polar coordinate corresponding to some point according to the horizontal and vertical coordinates at the point of the image of the rectangular coordinate; (4) obtaining an index value of the horizontal and vertical coordinates of the polar coordinate of the point according to a beam forming principle; (5) storing the image of the polar coordinate image into a memory, reading the corresponding pixel amplitude value according to the index value of the horizontal and vertical coordinates of the polar coordinate, and assigning the value to the position corresponding to the rectangular coordinate; and (6) completing mapping between each point and the polar coordinate under the rectangular coordinate to obtain the acoustic image of the rectangular coordinate.

Description

The Fast implementation of multi-beam image sonar middle fan fractal transform

Technical field

The invention belongs to image sonar field, relate generally to the implementation method of image sonar middle fan fractal transform.

Background technology

Multi-beam image sonar often adopts active mode work, utilize transmitting transducer to give off acoustical signal, by receiving transducer receiving target echo, echo data obtains target scene image through wave beam formation processing, image is exported with polar form, fan-shaped conversion module is for polar coordinates acoustic image is converted into rectangular coordinate acoustic image, so that graphoscope demonstration is suitable for user's visual custom.

The polar plot that wave beam forms module output similarly is that to take target be horizontal ordinate to the oblique distance r of receiving transducer, and the target of take is ordinate with respect to the orientation angles θ of receiving transducer.Fan-shaped conversion can, by forming mapping between the image pixel value g (x, y) under the image pixel value f (r, θ) under polar coordinates (r, θ) and rectangular coordinate (x, y), finally obtain rectangular coordinate acoustic image.One is presented at rectangular coordinate acoustic image on display, for soundman, carries out target information interpretation.As figure (1), the transformation relation between coordinate is

$\left\{\begin{array}{c}x=r\sin \mathrm{\θ}\\ y=r\cos \mathrm{\θ}\end{array}\right.\⇔\left\{\begin{array}{c}\mathrm{\θ}=\arctan \frac{x}{y}\\ r=\sqrt{x^2+y^2}\end{array}\right.---\left(1\right)$

Fan-shaped conversion can have two kinds of mapping modes.A kind of is that polar coordinates (r, θ) convert to rectangular coordinate (x, y), that is to say polar coordinates (r ₀, θ ₀) pixel amplitude f (r ₀, θ ₀) equal rectangular coordinate (x ₀=r ₀cos θ ₀, y ₀=r ₀sin θ ₀) pixel amplitude g (x ₀, y ₀), now need to calculate the trigonometric function value of sin and cos; Another kind is that rectangular coordinate (x, y) is transformed to polar coordinates (r, θ) the position value that gets on, and that is to say rectangular coordinate (x ₀, y ₀) under pixel amplitude g (x ₀, y ₀) equal polar coordinates

under pixel amplitude f (r ₀, θ ₀), need to calculate arctan trigonometric function value.Because rectangular coordinate acoustic image is presented on display, its horizontal ordinate x utilizes display horizontal resolution, and ordinate y utilizes display vertical resolution.Because the sampling number in polar coordinates acoustic image r direction (being oblique distance direction) is high more a lot of than display vertical resolution, if adopt polar coordinates to the method for rectangular coordinates transformation, every bit under polar coordinates all will be done the mapping of a rectangular coordinate conversion, there will be polar multiple spot to be all mapped to the phenomenon of the same point of rectangular coordinate, and calculated amount can be very large, so often adopt rectangular coordinates transformation to the method for the polar coordinates relevant position value of reaching.

When the receiver of image sonar utilizes FPGA to carry out real-time fan-shaped conversion, it is the most consuming time that what account for hardware resource most is exactly trigonometric function operation, the calculating of trigonometric function value has two kinds of processing modes, a kind of is that trigonometric function value computed in advance is stored in hardware memory, while needing, directly call, but need like this to consume very large hardware resource.Chen Jingyu [Chen Jingyu, the conversion of high-resolution radar displaing coordinate and relevant design, marine electronic antagonism, 2001] utilizes periodicity and the symmetry of trigonometric function, deposits the carry value of first quartile in EPROM, reduces the capacity requirement of storer with the method.But still need to consume hardware resource, store, some system configuration cannot satisfy the demands; Another kind of mode is to utilize iterative approach algorithm (cordic algorithm) to calculate trigonometric function value, for example the long root of poplar is at master thesis [Yang Changgen, acoustic imaging algorithm research and realization based on FPGA, Harbin: Harbin Engineering University's master thesis (CNKI), 2006] utilize cordic algorithm to realize the fan-shaped conversion of image sonar, but in order to obtain high-precision trigonometric function value, often required iterations is very large.Such as the people such as Tan Yiyu [Tan Yiyu, Bian Wenbing, a kind of coordinate conversion circuit based on cordic algorithm, data acquisition and processing, 2001] utilize equally cordic algorithm to realize the fan-shaped conversion of B ultrasonic image, but operational precision need high data quantization digit to guarantee.And the realization of cordic algorithm need to carry out the program design based on streamline, developer is had to certain requirement.

Summary of the invention

The object of the present invention is to provide a kind of Fast implementation of multi-beam image sonar middle fan fractal transform of the real-time that can largely improve fan-shaped conversion.The method makes the image sonar forming based on wave beam when realizing fan-shaped conversion, and the hardware such as FPGA, without loaded down with trivial details triangle or inverse trigonometric function computing, can largely improve the real-time of fan-shaped conversion.

The object of the present invention is achieved like this: (1) determines the monitor resolution that is used for showing rectangular coordinate acoustic image; (2), according to image sonar real standard and vertical distance range, calculate rectangular coordinate image horizontal and vertical range resolution; (3), according to the rectangular coordinate transverse and longitudinal coordinate figure of certain point, calculate this and put corresponding polar coordinates transverse and longitudinal coordinate figure; (4) according to wave beam, form principle and obtain polar coordinates transverse and longitudinal coordinated indexing value; (5) polar coordinate image is stored in storer, read respective pixel amplitude, and assignment is to rectangular coordinate correspondence position according to polar transverse and longitudinal coordinated indexing value; (6) complete every bit and polar mapping under rectangular coordinate system, obtain rectangular coordinate acoustic image.

Each pixel intensity of image is reflected by pixel amplitude size.From formula (1), rectangular coordinate (x ₀, y ₀) under pixel amplitude equal polar coordinates

under pixel amplitude.Polar pixel amplitude is stored in SDRAM or other storeies, and be from SDRAM correct value, does not need to know θ in fact ₀and r ₀specifically what, and need to know this θ ₀horizontal ordinate in polar coordinates (angle) index value (0,1,2. ...), r ₀also like this, like this could be according to the address table value of tabling look-up.And this value assignment is arrived to rectangular coordinate correspondence position (x ₀, y ₀), at this moment also must know x ₀, y ₀index value.

One, the method that obtains coordinated indexing value under polar coordinates and rectangular coordinate system is respectively

$\left\{\begin{array}{c}\mathrm{\θ}\_\mathrm{index}={\mathrm{\θ}}_0/\mathrm{\Δ\θ}\\ r\_\mathrm{index}=r_0/\mathrm{\Δr}\end{array}\right.$ With $\left\{\begin{array}{c}x\_\mathrm{index}=x_0/\mathrm{\Δx}\\ y\_\mathrm{index}=y_0/\mathrm{\Δy}\end{array}\right.---\left(2\right)$

Wherein θ _ index and r_index are respectively polar coordinates lower angles and apart from index value; Δ θ and Δ r are respectively angle and distance resolution, by image sonar performance, are determined; X_index and y_index be respectively under rectangular coordinate horizontal and vertical apart from index value; Δ x and Δ y are respectively horizontal and vertical range resolutions, by sonar range with count definite for the display picture element showing.Wherein solve θ _ index the most complicated, because according to formula (1), need to complete the computing of arctan arctan function, just can know azimuth angle theta ₀.The present invention solves and simplifies mainly for θ _ index.

Multi-beam image sonar utilizes basic matrix receiving target echo, by processing such as wave beam formation, obtains polar coordinates acoustic image.The wavelength of supposing echoed signal is λ, is spaced apart the uniformly-spaced linear base of d for N unit, and when plane wave incident direction is θ, normalization battle array output amplitude is

$R\left(\mathrm{\θ}\right)=\frac{\sin (\mathrm{N\π}d\sin \mathrm{\θ}/\mathrm{\λ})}{N\sin (\mathrm{\π}d\sin \mathrm{\θ}/\mathrm{\λ})}---\left(3\right)$

From formula (3), sin θ=± k λ/d (k=0,1,2 ...) position on get maximum value, the position of principal maximum is sin θ=0, maximum value is got on sin θ axle.According to wave beam, form principle, between adjacent array element, insert phase shift ξ, make the position changeable of principal maximum be

$\sin \mathrm{\θ}=\frac{\mathrm{\ξ\λ}}{2\mathrm{\πd}}---\left(4\right)$

So, as incident direction θ ₀when unknown, if insert different phase shifts or time delay between adjacent array element, by wave beam, form exportable different amplitude.Battle array output can reach principal maximum, by horizontal ordinate sin θ corresponding to principal maximum ₀can obtain incident orientation angle.For image sonar, for each sampling instant, carry out wave beam formation processing, can form a width with target with respect to the azimuth angle theta of receiving transducer (in fact take sin θ) as horizontal ordinate, the target of take is ordinate to the oblique distance r of receiving transducer, the two-dimentional polar coordinate image that the battle array output amplitude of take is pixel amplitude.

The two-dimensional coordinate of polar coordinate image is (sin θ, r), with nn point DFT wave beam, forms example, supposes that time domain waveform sampling rate is fs, and the velocity of sound is c, and collection duration is T, and owing to being active sonar, the distance between neighbouring sample point is c/2f _s, according to formula (4), the transverse and longitudinal coordinate of polar each coordinate points of image is

$\left\{\begin{array}{c}\sin \mathrm{\θ}=\frac{2\mathrm{\π}(0:\mathrm{nn}-1)}{\mathrm{nn}}\·\frac{\mathrm{\λ}}{2\mathrm{\πd}}=\frac{\mathrm{\λ}}{d}(0:\mathrm{nn}-1)\\ r=0:\frac{c}{2f_s}:\frac{\mathrm{cT}}{2}\end{array}\right.---\left(5\right)$

So the transverse and longitudinal coordinated indexing value of each coordinate points of polar coordinates is

$\left\{\begin{array}{c}\mathrm{\θ}\_\mathrm{index}=\sin \mathrm{\θ}\·\frac{d}{\mathrm{\λ}}=\frac{x}{\sqrt{x^2+y^2}}\frac{d}{\mathrm{\λ}}\\ r\_\mathrm{index}=\frac{2r}{c}{f}_s=\frac{2f_s}{c}\sqrt{x^2+y^2}\end{array}\right.---\left(6\right)$ Exactly because the image horizontal ordinate forming based on wave beam be sin θ (

it and horizontal ordinate index value θ index (0,1 ..., there is simple linear relationship between nn-1), at once can obtain horizontal ordinate (being angle) index value.When this value is non-integer, can round by the mode of interpolation or neighboring mean value, but this is not the emphasis that the present invention studies.As the above analysis, no longer need first to utilize arctan function value calculate θ ₀, then calculate angle index value and realize fan-shaped conversion.

By above-mentioned analysis, may be summarized as follows: by wave beam, forming what obtain is the polar coordinate image that the A * D in R * Θ region is ordered, if will be at R _x* R _yon region, form the rectangular coordinate image that M * N is ordered, so horizontal range resolution ax x=R _x/ M, vertical range resolution ax y=R _y/ N.Formula (6) can be rewritten into

$\left\{\begin{array}{cc}\mathrm{\&theta;}\_\mathrm{index}=\frac{x\_\mathrm{index}\&times;\mathrm{\&Delta;x}}{\sqrt{{(x\_\mathrm{index}\&times;\mathrm{\&Delta;x})}^2+{(y\_\mathrm{index}\&times;\mathrm{\&Delta;y})}^2}}\frac{d}{\mathrm{\&lambda;}}& 0\&le;x\_\mathrm{index}<M\\ r\_\mathrm{index}=\frac{2f_s}{c}\sqrt{{(x\_\mathrm{index}\&times;\mathrm{\&Delta;x})}^2+{(y\_\mathrm{index}\&times;\mathrm{\&Delta;y})}^2}& 0\&le;y\_\mathrm{index}<N\end{array}\right.---\left(7\right)$

With regard to known right angle image coordinate point (x_index, y_index) pixel amplitude g (x_index, y_index), equal polar coordinate image coordinate points (θ _ index, r_index) pixel amplitude f (θ _ index, r_index), can form like this rectangular coordinate image that view picture M * N is ordered.

Principal feature of the present invention is embodied in:

1, according to by the institute under rectangular coordinate a little (x, y) be transformed into the get on mode of value of polar coordinates (r, θ) position and carry out fan-shaped conversion.

2, without trigonometric function operation is carried out to program design, can complete the mapping of coordinated indexing value between rectangular coordinate and polar coordinates.

3, without trigonometric function value computed in advance is stored in hardware memory.

4, utilize beam-forming technology, the horizontal ordinate of the polar coordinate image that sonar obtains is sin θ, avoids solving θ and specifically equals how many, direct solution horizontal ordinate index value during fan-shaped conversion.

, there is simple linear relationship between the horizontal ordinate sin θ of polar coordinate image and horizontal ordinate index value θ _ index in the principle 5, forming due to wave beam, at once can obtain horizontal ordinate index value.

Accompanying drawing explanation

Fig. 1 is the conversion schematic diagram between polar coordinate image and rectangular coordinate image.

Embodiment

The function of fan-shaped conversion has been the mapping between polar coordinate image and the address table of rectangular coordinate image, namely the mapping between the transverse and longitudinal coordinated indexing value of image.Again because the horizontal ordinate of the polar coordinate image that the image sonar forming based on wave beam obtains is sin θ, between it and polar angle index value through deriving, there is simple linear relationship, so can not need to calculate triangle or inverse trigonometric function is realized fan-shaped conversion, according to by (the x a little of institute under rectangular coordinate, y) be transformed into the get on mode of value of polar coordinates (r, θ) position and carry out fan-shaped conversion, concrete methods of realizing is as follows:

(1) determine the monitor resolution M * N that is used for showing rectangular coordinate acoustic image, obtain the span of the transverse and longitudinal coordinated indexing value of rectangular coordinate image: 0≤x_index < M, 0≤y_index < N;

(2) distance parameter arranging according to system (supposes that horizontal range scope is for [Xmax Xmax], vertical distance range is [0 Ymax]), calculate rectangular coordinate image level (being horizontal ordinate) range resolution Δ x=2Xmax/M and vertical (being ordinate) range resolution Δ y=Ymax/N;

(3) under rectangular coordinate system, calculate certain some transverse and longitudinal coordinate figure x=x_index * Δ x, y=y_index * Δ y, and then calculate the transverse and longitudinal coordinate figure (sin θ, r) under the polar coordinates of this point;

(4) basis $\left\{\begin{array}{cc}\mathrm{\&theta;}\_\mathrm{index}=\frac{x\_\mathrm{index}\&times;\mathrm{\&Delta;x}}{\sqrt{{(x\_\mathrm{index}\&times;\mathrm{\&Delta;x})}^2+{(y\_\mathrm{index}\&times;\mathrm{\&Delta;y})}^2}}\frac{d}{\mathrm{\&lambda;}}& 0\&le;x\_\mathrm{index}<M\\ r\_\mathrm{index}=\frac{2f_s}{c}\sqrt{{(x\_\mathrm{index}\&times;\mathrm{\&Delta;x})}^2+{(y\_\mathrm{index}\&times;\mathrm{\&Delta;y})}^2}& 0\&le;y\_\mathrm{index}<N\end{array}\right.$ Formula is calculated polar transverse and longitudinal coordinated indexing value: θ _ index, r_index, if non-integer rounds according to the mode of interpolation or neighboring mean value;

(5) polar coordinate image is stored in storer, according to (θ _ index, r_index), reads corresponding pixel amplitude, and assignment is to pixel amplitude corresponding to (x_index, y_index) under rectangular coordinate image.

(6) repeat (3)～(5) step and complete every bit and polar mapping under rectangular coordinate system, finally obtain a complete rectangular coordinate acoustic image.

Claims (1)

1. the Fast implementation of multi-beam image sonar middle fan fractal transform, makes the image sonar forming based on wave beam when realizing fan-shaped conversion, and FPGA, without loaded down with trivial details triangle or inverse trigonometric function computing, can largely improve the real-time of fan-shaped conversion; It is characterized in that: (1) determines the monitor resolution M * N that is used for showing rectangular coordinate acoustic image; (2), according to image sonar real standard and vertical distance range, calculate rectangular coordinate image horizontal range resolution ax x and vertical range resolution ax y; (3) according to rectangular coordinate transverse and longitudinal coordinate figure x=x_index * Δ x of certain point, y=y_index * Δ y, wherein x_index and y_index are rectangular coordinate index values, calculate this and put corresponding polar coordinates transverse and longitudinal coordinate figure $\left\{\begin{array}{c}\sin \mathrm{\&theta;}=\frac{x}{\sqrt{x^2+y^2}}=\frac{x\_\mathrm{index}\&times;\mathrm{\&Delta;x}}{\sqrt{{(x\_\mathrm{index}\&times;\mathrm{\&Delta;x})}^2+{(y\_\mathrm{index}\&times;\mathrm{\&Delta;y})}^2}}\\ r=\sqrt{x^2+y^2}=\sqrt{{(x\_\mathrm{index}\&times;\mathrm{\&Delta;x})}^2+{(y\_\mathrm{index}\&times;\mathrm{\&Delta;y})}^2}\end{array}\right.;$ (4) according to wave beam, form principle and obtain polar coordinates transverse and longitudinal coordinate $\left\{\begin{array}{c}\sin \mathrm{\&theta;}=\frac{2\mathrm{\&pi;}(0:\mathrm{nn}-1)}{\mathrm{nn}}\&CenterDot;\frac{\mathrm{\&lambda;}}{2\mathrm{\&pi;d}}=\frac{\mathrm{\&lambda;}}{d}\frac{(0:\mathrm{nn}-1)}{\mathrm{nn}}\\ r=0:\frac{c}{2f_s}:\frac{\mathrm{cT}}{2}\end{array}\right.,$ Wherein λ is signal wavelength, and d is linear base array element interval uniformly-spaced, 0,1 ..., nn-1 is polar coordinates horizontal ordinate index value, and fs is time domain waveform sampling rate, and c is the velocity of sound, and T is collection duration, the calculating formula of integrating step (3), polar coordinates transverse and longitudinal coordinated indexing value is $\left\{\begin{array}{cc}\mathrm{\&theta;}\_\mathrm{index}=\frac{x\_\mathrm{index}\&times;\mathrm{\&Delta;x}}{\sqrt{{(x\_\mathrm{index}\&times;\mathrm{\&Delta;x})}^2+{(y\_\mathrm{index}\&times;\mathrm{\&Delta;y})}^2}}\frac{d\&CenterDot;\mathrm{nn}}{\mathrm{\&lambda;}}& 0\&le;x\_\mathrm{index}<M\\ r\_\mathrm{index}=\frac{2f_s}{c}\sqrt{{(x\_\mathrm{index}\&times;\mathrm{\&Delta;x})}^2+{(y\_\mathrm{index}\&times;\mathrm{\&Delta;y})}^2}& 0\&le;y\_\mathrm{index}<\end{array}\right.;$ (5) polar coordinate image is stored in storer, read respective pixel amplitude, and assignment is to rectangular coordinate correspondence position according to polar transverse and longitudinal coordinated indexing value θ _ index and r_index; (6) complete every bit and polar mapping under rectangular coordinate system, obtain rectangular coordinate acoustic image.

Images