CN106918807B - A kind of Targets Dots condensing method of radar return data - Google Patents

Abstract

The invention belongs to the technical field of radar signal processing, and discloses a method for condensing target point traces of radar echo data. The pulse non-coherent accumulation is used to obtain the echo data after pulse accumulation; the echo data after pulse accumulation is subjected to constant false alarm detection; binary sliding window detection; take the echo data detected by the binary sliding window as the binary image data, and perform the operation of dilating and then corroding the binary image data, so as to obtain all connected domains in the binary image data; filter In addition to the first connected domain and the second connected domain, the remaining connected domain containing the target point trace is obtained; the target information is obtained according to the remaining connected domain containing the target point trace, so as to realize the condensation of the target point trace; it can improve the estimation accuracy of the target parameters and the target Distinguishing ability.

Description

A Target Point Trace Condensation Method for Radar Echo Data

technical field

The invention belongs to the technical field of radar signal processing, and in particular relates to a method for condensing target point traces of radar echo data, which is used for preprocessing of radar echo data and extraction of target parameters.

Background technique

With the development of radar technology, people require radar to have more functions and roles. The radar in the traditional sense is mainly used to measure the basic information such as distance, azimuth, and pitch of the target. Modern radar requires detection, point trace recording and target tracking to be realized in a more complex environment (interference and clutter).

Modern radar technology is developing in the direction of intelligence, miniaturization, informatization and precision. People require radar to be able to obtain more precise target information and achieve more stable target tracking. Point trace condensation technology is the process of estimating target parameters and recording point traces by processing radar echo data. The radar data processing system processes the recorded point traces to achieve target tracking display. Compared with foreign countries, the domestic research on radar data processing technology started later, and the research on point trace condensation processing technology is mainly aimed at the radar system of the MTD system. At this stage, the main methods of point trace condensation are point trace condensation based on sliding window method and point trace condensation based on image processing.

The main problems of point trace condensation at this stage are how to accurately discriminate target point traces in complex environments, and how to improve the processing methods of point trace condensation under different radar systems. The point trace condensation method based on the sliding window method is a commonly used point trace condensation method. The main problem is that the target point trace will be split when the target end threshold is large. The method of threshold value is used to solve the problem, and the threshold value should not be too low. Too low threshold value will cause too many false traces and affect the quality of traces. The classical method based on image processing is to use the method of image contour search to obtain the centroid to achieve point trace condensation, which cannot improve the splitting problem of point traces in distance and azimuth.

SUMMARY OF THE INVENTION

In view of the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a target point trace condensation method for radar echo data, which not only adopts the binary sliding window detection in the sliding window method, but also adopts the morphological filtering algorithm in mathematics. The connected domain is expanded and corroded. By selecting appropriate structural elements, the splitting phenomenon of the target point trace in the sliding window method can be improved not only in azimuth, but also in distance.

In order to achieve the above object, the present invention adopts the following technical solutions to achieve.

A method for condensing target point traces of radar echo data, the method comprises the following steps:

Step 1: Obtain the raw echo data of multiple scanning cycles of the radar after pulse compression processing, the raw echo data is two-dimensional data including the distance dimension and the azimuth dimension, and the raw echo data of the multiple scanning cycles are The orientation dimension is arranged in order;

Step 2, sequentially performing envelope detection and non-coherent accumulation on the original echo data to obtain echo data after pulse accumulation;

Step 3, performing constant false alarm detection on the echo data after the pulse accumulation, to obtain echo data after constant false alarm detection, and then performing binary quantization processing on the echo data after constant false alarm detection in turn; The binary sliding window detection along the azimuth dimension is used to obtain the echo data after the binary sliding window detection.

Step 4: Convert the echo data detected by the binary sliding window into corresponding binary image data, and the length of the binary image corresponds to the distance unit of the echo data detected by the binary sliding window. The width of the image corresponds to the pulse number of the echo data detected by the binary sliding window, and the binary image data is first expanded and then eroded to obtain all connected domains in the binary image data;

Step 5: In all the connected domains, filter out the first connected domain and the second connected domain, the first connected domain is a connected domain that contains only one isolated point trace, and the second connected domain is a connected domain in the distance dimension. The connected domain composed of the point traces exceeding the threshold value is upwardly expanded, thereby obtaining the remaining connected domain containing the target point trace; the target information is obtained according to the remaining connected domain containing the target point trace, so as to realize the agglomeration of the target point trace.

The characteristics and further improvement of the technical solution of the present invention are:

(1) After the step 1 and before the step 2, the method further includes: performing secondary cancellation on the original echo data to suppress clutter in the original echo data.

(2) Described step 2 is specifically:

Envelope detection is performed on the original echo data, so as to obtain the amplitude information of the original echo data in the corresponding distance-azimuth unit; the distance unit is the sampling point of the original echo data in the distance dimension, and the azimuth The unit is the sampling point of the original echo data in the azimuth dimension.

The echo data after envelope detection is accumulated on adjacent pulses by sliding window to obtain echo data after pulse accumulation.

(3) the step 3 is specifically:

Sliding window detection is performed on the echo data after the pulse accumulation along the distance unit, and the value obtained by multiplying the mean value of all reference units in the window by the threshold factor is set as the threshold value of sliding window detection; the amplitude value of the unit to be detected is set. Compared with the threshold value of the sliding window detection, if the amplitude of the unit to be detected is greater than the threshold value of the sliding window detection, the data of the unit to be detected is retained, otherwise the data of the unit to be detected is set is 0;

Then, the data of the unit to be detected whose amplitude value is greater than the threshold value of the sliding window detection is set to 1, and the process of performing binary quantization processing on the echo data after the constant false alarm detection is completed;

When binary sliding window detection is performed on the echo data after binary quantization processing along the azimuth dimension, the M/N criterion is adopted, where M represents the detection threshold of binary sliding window detection, and N represents the detection threshold of binary sliding window detection. The length of the window.

(4) In step 4, the binary image data is first expanded and then eroded, so as to obtain all connected domains in the binary image data, specifically including:

Select the structuring elements, use the selected structuring elements to dilate and then corrode the binary image data, and then search the connected domain of the binary image data that is dilated and then corroded.

(5) In step 5, the target information is obtained according to the remaining connected domain including the target point trace, so as to realize the agglomeration of the target point trace, which specifically includes:

Obtain the centroid position of the remaining connected domain including the target trace, take the distance unit corresponding to the centroid position in the original echo data as the distance information of the target, and take the centroid position in the original echo data The corresponding azimuth The orientation information of the unit as the target;

Obtain the original echo data corresponding to the remaining connected domain containing the target point trace, according to the distance-azimuth unit where the original echo data corresponding to the remaining connected domain containing the target point trace is located and the corresponding envelope detection result, find Get the distance information, bearing information and amplitude information of the target.

Compared with the prior art, the present invention has the following advantages:

(1) Since the present invention converts echo data into binary images for processing, software implementation and offline processing can be used; it has respective advantages based on image processing and point trace condensation methods based on the sliding window method, and is in the classical point trace condensation method. On the basis of the morphological filtering method, the point trace is further processed, which makes up for the shortcomings of the two classical methods, further improves the splitting phenomenon of the target point trace, and at the same time has a higher point trace quality;

(2) The target point trace agglomeration method provided by the present invention adopts a binary image connected domain search method, which can accurately determine the distribution area of the target point trace, and has a higher resolution for the target point trace. After searching, the centroid method is used to determine the regional centroid Or use the echo amplitude weighting method to estimate the target parameters, and the target parameter estimation is more accurate than the sliding window method;

(3) It is more effective for the discrimination of target traces and the filtering of false traces. First, combining the discrimination of target traces based on the sliding window method, the point trace aggregation criterion and the characteristics of target distribution, the following target discrimination criteria can be formulated: The target point traces with a larger extension in the distance are filtered out and the isolated point traces are filtered out, and then the small connected domain area is filtered out (the specific parameters should be set according to the situation, there is no unified standard). Judgment and dot trace filtering criteria can effectively identify target dot traces, and at the same time can flexibly control the number of dot traces;

(4) Connected domain merging through image expansion and erosion operation can effectively reduce the impact of target splitting, and further improve the effect of point trace condensation based on the sliding window method.

Description of drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can also be obtained according to these drawings. On the premise of no creative work, other drawings can also be obtained from these drawings.

1 is a schematic flowchart of a method for condensing target point traces of radar echo data according to an embodiment of the present invention;

2 is a schematic diagram of a binary sliding window detection process provided by an embodiment of the present invention;

3 is a schematic diagram of the expansion results of different structural elements on the same target provided by an embodiment of the present invention;

4 is a schematic diagram of the corrosion results of different structural elements on the same target according to an embodiment of the present invention;

5 is a schematic diagram of a pixel adjacency relationship provided by an embodiment of the present invention;

6 is a schematic diagram of the connectivity relationship of points P, Q, and S provided by an embodiment of the present invention;

7 is a schematic diagram of a connected domain search process provided by an embodiment of the present invention;

FIG. 8 is a schematic diagram of original echo data provided by an embodiment of the present invention;

9 is a schematic diagram of echo data after non-coherent accumulation provided by an embodiment of the present invention;

10 is a schematic diagram of echo data after a constant false alarm detection result provided by an embodiment of the present invention;

11 is a schematic diagram of a binary sliding window detection result provided by an embodiment of the present invention;

12 is a schematic diagram of a binary image dilation corrosion result provided by an embodiment of the present invention;

FIG. 13 is a schematic diagram of a result of target point trace condensation according to an embodiment of the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

The development of radar technology requires the radar system to provide higher target parameters. After the antenna beam scans the target, the echo data will be split after signal processing and target detection, which can be divided into distance splitting and azimuth splitting. For two-coordinate radar, point trace condensation is the extraction of primary information (range and azimuth) of the target. Usually, a binary sliding window detector is used to extract the target information, and the degree of false alarm reduction of the sliding window detector depends on the detection threshold and the window length. On the one hand, dot trace condensation requires filtering out false dot traces, and on the other hand, it requires high resolution. Whether the impact of target splitting can be reduced and the quality of target parameters can be improved is an important indicator for point trace condensation. The image processing-based point trace condensation method provided by the embodiments of the present invention is, on the one hand, implemented in a binary image, and is not affected by other conditions, so it is relatively easy to implement; The criterion for the discrimination of the point trace is added, which is beneficial to improve the quality of the target point trace. The image processing-based point trace condensation method provided by the embodiment of the present invention is the result of combining the idea and principle of point trace condensation based on the sliding window method with an image processing related algorithm, and improving on this basis.

An embodiment of the present invention provides a method for condensing target point traces of radar echo data. As shown in FIG. 1 , the method includes the following steps:

Step 1: Obtain the raw echo data of multiple scanning cycles of the radar after pulse compression processing, the raw echo data is two-dimensional data including the distance dimension and the azimuth dimension, and the raw echo data of the multiple scanning cycles are The orientation dimensions are arranged in order.

It should be supplemented that, after the step 1 and before the step 2, the method further includes: performing secondary cancellation on the original echo data to suppress clutter in the original echo data.

Step 2: Envelope detection and multi-pulse non-coherent accumulation are sequentially performed on the original echo data to obtain echo data after pulse accumulation.

Pulse accumulation is divided into coherent accumulation and non-coherent accumulation. Coherent accumulation can improve the target signal-to-noise ratio well, but due to the high computational complexity and the large error in the estimation of point trace aggregation after coherent accumulation, especially when the target Doppler frequency appears in multiple channels Or when the target moves tangentially, the condensing effect is not good. For the two-coordinate radar, the main parameters of the target are distance, azimuth and speed, and the target parameters are extracted according to the data after the signal target is detected. Adopting non-coherent accumulation has low computational complexity and strong adaptability in many scenarios, so the technical solution of the present invention adopts multi-pulse non-coherent accumulation.

The step 2 is specifically:

Envelope detection is performed on the original echo data, so as to obtain the amplitude information of the original echo data in the corresponding distance-azimuth unit; the distance unit is the sampling point of the original echo data in the distance dimension, and the azimuth The unit is the sampling point of the original echo data in the azimuth dimension.

The echo data after envelope detection is accumulated on adjacent pulses by sliding window to obtain echo data after pulse accumulation.

Step 3, performing constant false alarm detection on the echo data after the pulse accumulation, to obtain echo data after constant false alarm detection, and then performing binary quantization processing on the echo data after constant false alarm detection in turn; The binary sliding window detection along the azimuth dimension is used to obtain echo data after binary sliding window detection, which further reduces the influence of false point traces. The echo data detected by the binary sliding window is two-dimensional data, wherein one dimension represents the distance unit, and the other dimension represents the number of pulses in the azimuth dimension;

The constant false alarm detection adopts cell averaged constant false alarm detection (CA-CFAR).

In the step 3, further perform binary quantization processing and binary sliding window detection along the azimuth dimension on the echo data after the constant false alarm detection, to obtain the echo data after the binary sliding window detection, specifically: :

First, the echo data after the pulse accumulation obtained by the processing in step 2 is processed by using the unit average constant false alarm detection. The value obtained by the threshold factor is set as the threshold value of the sliding window detection, the amplitude of the unit to be detected is compared with the threshold value of the sliding window detection, the data that has passed the threshold is retained, and the data that has not passed the threshold is set to 0; then The data that has passed the threshold value is set to 1, so as to complete the process of performing constant false alarm detection and binary quantization processing on the echo data after the pulse accumulation;

When binary sliding window detection is performed on the echo data after binary quantization processing along the azimuth dimension, the M/N criterion is adopted, where M represents the detection threshold of binary sliding window detection, and N represents the detection threshold of binary sliding window detection. The length of the window.

The target start threshold of binary sliding window detection has little effect on the resolution of target traces, so the target end threshold is set to be the same as the start threshold, and the relationship between M and N is selected according to the empirical formula M=1.5*sqrt(N) , but in order to reduce the target point trace splitting, the detection threshold of the binary sliding window detection should be appropriately reduced.

It should be noted that the target end threshold value refers to the process of distinguishing the target traces in the sliding window detection process, and the accumulated value of the binary sliding window is a process of change. The starting point and ending point on the azimuth, so as to complete the resolution of the target trace in the azimuth.

Specifically, the binary sliding window detection process is shown in Figure 2, in which Figure 2(a) is the input signal after binary quantization processing, Figure 2(b) is the binary sliding window detection accumulation process, Figure 2(c) ) is a schematic diagram of the binary sliding window detection result; the M/N criterion refers to the window length N=5, the detection threshold M=3, when the binary accumulation result in the window is greater than or equal to the threshold value, it is considered that the detected target is marked as 1, Otherwise, it is recorded as 0.

It should be added that the original echo data refers to the data after the antenna is scanned and processed by pulse compression, and the processing of the echo data includes clutter suppression, signal accumulation and constant false alarm detection.

For the two-coordinate radar, the target parameter refers to the distance and azimuth information of the target. Here, the echo data is non-coherently accumulated, and the clutter suppression is realized by a canceller.

However, according to specific circumstances, for example, in order not to affect the target detection of tangential flight, the echo data can be simply accumulated and the secondary cancellation process can be omitted. The present invention uses sliding window accumulation, constant virtual methods for alarm detection and binary sliding window detection.

Step 4: Take the echo data detected by the binary sliding window as binary image data, and perform an operation of dilating and then corroding the binary image data, so as to obtain all connected domains in the binary image data. .

Specifically, a morphological filtering algorithm is used to perform a dilation and erosion operation on the binary image data, in order to realize the merging of connected domains.

Image expansion and erosion operations, from the perspective of image processing, binary image processing is to move and compare a small structural element on a large binary image point by point, and make corresponding processing according to the comparison results. Structural elements refer to background images with a certain size. Structural elements have no fixed shape. When designing morphological transformation algorithms, structural elements need to be designed according to the input image and the required information. Usually the shape of the structuring element is square, rectangle, circle and linear etc. Simply put, expansion is to merge the background points around the object into the object, and the object expands outward through expansion, so that it is possible for two adjacent objects to be connected together, which is represented as a connected domain in a binary image. 's merger. In contrast, the erosion operation of the image is to remove the boundary points of objects, and also to remove smaller objects.

(1) The expansion processing of the image can be simply described as the traversal of the structuring element on the binary image. Taking the black point of the binary image as an example, when the origin of the structuring element is the same as the pixel of a certain point in the binary image, then the structuring element is the same as the binary image. All points corresponding to the value image become black points. Figure 3 is a schematic diagram of the expansion results of different structural elements for the same target, Figure 3 (a) is the structural element 1, Figure 3 (b) is the structural element 2, Figure 3 (c) is the original binary image, Figure 3 (d) is the result of expanding the original binary image with structural element 1, and Figure 3(e) is the result of expanding the original binary image with structural element 2, where the symbol "+" in the structural element represents the origin of coordinates location.

(2) The corrosion operation of the image can be simply described as the traversal of the structuring element on the binary image, taking the black point of the binary image as an example, when the origin of the structuring element is the same as the pixel of a certain point in the binary image, if the rest of the structuring elements If one or more pixels of the pixel are different from the corresponding pixels of the binary image, the point of the binary image corresponding to the origin becomes a white point. Figure 4 shows the schematic diagram of the corrosion results of different structural elements on the same target, Figure 4(a) is the structural element 1, Figure 4(b) is the structural element 2, Figure 4(c) is the original binary image, Figure 4 (d) is the result of corroding the original binary image by using structural element 1, and FIG. 4(e) is the result of corroding the original binary image by using structural element 2.

The structural elements used in the embodiments of the present invention are line-segment structural elements, the center point is at the center of the line segment, and the length is 1/4 of the number of coherent processing pulses. The binary image dilation realizes the merging of connected domains, and the erosion operation restores the dilated binary image, but the restored image is different from the original binary image.

Specifically, in step 4, when all the connected domains in the binary image data are obtained, it is necessary to search the connected domains after expansion and corrosion.

Adjacency and connectivity are the basic relationships between pixels. Except for edge pixels in a binary image, there are 8 adjacent points around each pixel, and 4 adjacent points are distinguished according to the position of the adjacent points. As shown in Fig. 5, Fig. 5(a) is a schematic diagram of 8 adjacencies around the black dots, and Fig. 5(b) is a schematic diagram of 4 adjacencies around the black dots.

In a binary image, if two points are connected together by a series of point sequences with the same gray level according to the 4-adjacency relationship or the 8-adjacency relationship, then the two points are said to be connected domains, so that all the points are connected to these two points. The set of points constitutes a connected domain. As shown in Figure 6, the points P, Q, and S whose pixel value is 1, where P and Q are 8 connected, S and Q are 4 connected, and S and P are 8 connected.

Taking 8 connectivity as an example, the connected domain is searched for the pixel whose pixel value is 1. Assuming that A is a connected domain in the binary image, and a point in A is known as P, then the search for the connected domain A can be done through the following Iteratively:

X ₀ =P

The algorithm converges when _Xk = _Xk-1 , and A= _Xk . where B represents a structural element, Indicates the image dilation operation, and Y is the original binary image.

Figure 7 is a schematic diagram of the connected domain search, using 8 connectivity relationships, the structural elements are shown in Figure 7(a), Figure 7(b) is the starting point P of the connected domain search, the black elements are the extracted elements, and the gray is not yet. The extracted elements, Figure 7(c) is the result of the first iteration of finding the connected domain, Figure 7(d) is the result of the second iteration of finding the connected domain, and Figure 7(e) is a schematic diagram of the final result of finding the connected domain. According to the connected domain search method, the searched connected domain needs to be marked until all the points whose pixels are 1 are marked, thus completing the connected domain search.

Step 5: In all the connected domains, filter out the first connected domain and the second connected domain, the first connected domain is a connected domain that contains only one isolated point trace, and the second connected domain is a connected domain in the distance dimension. The connected domain composed of the point traces exceeding the threshold value is upwardly expanded, thereby obtaining the remaining connected domain containing the target point trace; the target information is obtained according to the remaining connected domain containing the target point trace, so as to realize the agglomeration of the target point trace.

In step 5, the target information is obtained according to the remaining connected domain including the target point trace, and the condensation of the target point trace is realized, which specifically includes:

Obtain the centroid position of the remaining connected domain including the target trace, take the distance unit corresponding to the centroid position in the original echo data as the distance information of the target, and take the centroid position in the original echo data The corresponding azimuth The orientation information of the unit as the target;

Obtain the original echo data corresponding to the remaining connected domain including the target point trace, according to the distance unit, the azimuth unit and the corresponding envelope detection result where the original echo data corresponding to the remaining connected domain including the target point trace is located, Obtain the distance information, bearing information and amplitude information of the target.

Specifically, in the embodiment of the present invention, when the target echo of the measured data is weak, in order to retain the target spot trace as much as possible, the adopted spot trace filtering criterion is:

(1) Filter out independent point traces. According to the distribution characteristics of target echoes, the target usually expands in azimuth or distance. Therefore, some false points can be filtered out according to the size of the target point trace expansion in distance or azimuth. trace;

(2) Filter out the point traces with more expansion on the distance corresponding to the echo data, and the maximum value of the distance unit is determined according to the estimated parameters of the target and the system;

(3) When the system allows, as many dot traces should be retained as possible, so the dot trace discrimination and filtering criteria should be increased or decreased according to the actual situation.

Experiment content and results

Experiment 1. The echo data is the echo data recorded by a certain air-to-air radar. Figure 8 shows the intercepted part of a certain air-to-air data. The data is spliced from the data on the same azimuth and distance segment intercepted by the radar within 9 scanning periods. made.

The target echo data display is shown in Figure 8, and the echo data accumulated by the sliding window is shown in Figure 9. Figure 10 shows the result after constant false alarm processing, Figure 11 shows the result after binary sliding window detection, Figure 11(a) is the binary sliding window detection M/N criterion, M=2, The detection result when N=5, Fig. 11(b) is the detection result when M=3 and N=5 in the binary sliding window detection M/N criterion. FIG. 12 shows the result of the binary image after dilation and corrosion, and FIG. 13 shows the point trace condensation result of the target point trace condensation method provided by the embodiment of the present invention.

As can be seen from Figure 8, there is a strong clutter distribution around the target, and it is difficult to determine the position of the target trace from a single intercepted data. Therefore, the position of the target can be roughly seen through the spliced data. The estimation of point trace condensation effect is verified.

It can be seen from Figure 9 that after the target is accumulated non-coherently, the improvement of the weak target is not obvious, but it still has certain benefits, mainly because it has a certain smoothing effect on the meteorological clutter distributed along the radial direction, which is very important for the next step. It is advantageous to perform constant false alarm detection.

It can be seen from Figure 10 that the target traces and false traces in the remaining traces are greatly affected by signal processing and constant false alarm detection. After the constant false alarm detection, it is compared with the original echo data. It can be seen that the dot trace splitting phenomenon of the same target is more obvious, and the marked place in the figure is the dot trace of the target splitting.

As can be seen from Figure 11, the binary sliding window detection can effectively improve the impact of target splitting and reduce false alarms, but when the M/N criterion is used, if the threshold value is too large, the point traces of weaker targets will be lost. The target point trace splitting phenomenon will occur during detection, which is usually solved by reducing the threshold value. If the threshold value is too low, false traces will increase, which has a great impact on the quality of the point traces. Therefore, the threshold value needs to be considered as a compromise. It can be seen from Figure 12 that, by performing the dilation and corrosion operation on the binary image detected by the sliding window, on the one hand, the target resolution is guaranteed, and on the other hand, the improvement of target splitting is also quite effective.

It can be seen from Figure 13 that the improved method successfully achieves point trace condensation, and the target point trace condensation achieves good results.

To sum up, it can be seen from the experimental results that the method of the present invention can effectively realize the point trace agglomeration, and at the same time reduce the influence of target splitting on the point trace agglomeration.

Compared with the point trace condensation of the sliding window method, the technical solutions provided by the embodiments of the present invention greatly improve the target parameter accuracy and target resolution ability based on the image processing-based point trace condensation technology, and the related algorithms in the image, such as the binary image used in the present invention The dilation corrosion method can further improve the problem of target splitting in the point trace condensation of the sliding window method. The advantage of the sliding window method is that it can effectively overcome the influence of target splitting and false alarms, but the selection of the threshold value of the sliding window method has a great influence on the point trace condensation, which can be improved by morphological filtering in image processing. Therefore, combining the advantages of the sliding window method and image processing, the point trace is transformed into the image domain for processing, and the connected domain search algorithm and image expansion and corrosion processing are used. On the one hand, binary sliding window detection is used to process the target echo data and point traces. On the other hand, image processing methods can be used to overcome dot trace splitting, so as to obtain higher quality target estimation parameters. The method proposed in this patent combines the sliding window method and the image processing method. The improved method is mainly reflected in the first is to use the sliding window detector and the constant false alarm detection to process the echo data, and the second is to use the morphological filtering method in the image processing. Improves the effect of target trail splitting. The improved point trace condensation method has important reference value and research significance for the research of point trace condensation technology, and the measured data verify the effectiveness of the proposed method.

The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited to this. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention. should be included within the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims (6)

1. a target point trace condensation method of radar echo data, is characterized in that, described method comprises the steps: Step 1: Obtain the raw echo data of multiple scanning cycles of the radar after pulse compression processing, the raw echo data is two-dimensional data including the distance dimension and the azimuth dimension, and the raw echo data of the multiple scanning cycles are The orientation dimension is arranged in order; Step 2, sequentially performing envelope detection and non-coherent accumulation on the original echo data to obtain echo data after pulse accumulation; Step 3, performing constant false alarm detection on the echo data after the pulse accumulation, to obtain echo data after constant false alarm detection, and then performing binary quantization processing on the echo data after constant false alarm detection in turn; The binary sliding window detection along the azimuth dimension obtains the echo data after the binary sliding window detection. The echo data after the binary sliding window detection is two-dimensional data, in which one dimension represents the distance unit, and the other dimension represents the distance unit. number of pulses in the azimuth dimension; Step 4: Convert the echo data detected by the binary sliding window into corresponding binary image data, and the length of the binary image data corresponds to the distance unit of the echo data detected by the binary sliding window, two The width of the binary image data corresponds to the number of pulses of the echo data detected by the binary sliding window, and the binary image data is first expanded and then eroded to obtain all connected domains in the binary image data. ; Step 5: In all the connected domains, filter out the first connected domain and the second connected domain, the first connected domain is a connected domain that contains only one isolated point trace, and the second connected domain is a connected domain in the distance dimension. The connected domain composed of the point traces exceeding the threshold value is upwardly expanded, thereby obtaining the remaining connected domain containing the target point trace; the target information is obtained according to the remaining connected domain containing the target point trace, so as to realize the agglomeration of the target point trace. 2 . The method for condensing target point traces of radar echo data according to claim 1 , wherein, after the step 1 and before the step 2, the method further comprises: analyzing the original The echo data is cancelled twice to suppress the clutter in the original echo data. 3. a kind of target point trace condensation method of radar echo data according to claim 1, is characterized in that, described step 2 is specifically: Envelope detection is performed on the original echo data, thereby obtaining the amplitude information of the original echo data in the corresponding distance-azimuth unit; the distance unit is the sampling point of the original echo data in the distance dimension, and the azimuth unit is the The sampling points of the original echo data in the azimuth dimension; The echo data after envelope detection is accumulated on adjacent pulses by sliding window to obtain echo data after pulse accumulation. 4. a kind of target point trace condensation method of radar echo data according to claim 1, is characterized in that, described step 3 is specifically: Sliding window detection is performed on the echo data after the pulse accumulation along the distance unit, and the value obtained by multiplying the mean value of the respective amplitudes of all reference units in the window by the threshold factor is set as the threshold value of the sliding window detection; The amplitude value of the unit to be detected is compared with the threshold value of the sliding window detection. If the amplitude value of the unit to be detected is greater than the threshold value of the sliding window detection, the data of the unit to be detected will be retained, otherwise the unit to be detected will be retained. The data is set to 0; the threshold factor is a preset value; Then, the data of the unit to be detected that is greater than the threshold value of the sliding window detection is set to 1, and the process of performing binary quantization processing on the echo data after the constant false alarm detection is completed; When using the M/N criterion for binary sliding window detection along the azimuth dimension of the echo data after binary quantization, M represents the detection threshold of binary sliding window detection, and N represents the length of the window used in binary sliding window detection. . 5 . The method for condensing target point traces of radar echo data according to claim 1 , wherein in step 4, the binary image data is first expanded and then eroded to obtain the binary image data. 6 . Values all connected domains in image data, including: Select the structuring element, and use the selected structuring element to perform the operation of dilating and then eroding the binary image data to obtain the binary image data that is dilated and then eroded. Thus, all connected domains in the binary image data are obtained. 6. The method for condensing target point traces of a kind of radar echo data according to claim 1, it is characterized in that, in step 5, obtain target information according to the remaining connected domain that contains target point traces, realize to target point traces agglomeration, including: Obtain the centroid position of the remaining connected domain including the target trace, take the distance unit corresponding to the centroid position in the original echo data as the distance information of the target, and take the centroid position in the original echo data The corresponding azimuth The orientation information of the unit as the target; Obtain the original echo data corresponding to the remaining connected domain containing the target point trace, according to the distance-azimuth unit where the original echo data corresponding to the remaining connected domain containing the target point trace is located and the corresponding envelope detection result, find Get the distance information, bearing information and amplitude information of the target.