A real-time correction method for the lag of track detection

Abstract

The present invention discloses a real-time correction method for track detection lag. For the azimuth spectrum of track detection output in real time, when the azimuth of a certain target changes rapidly, the local peak region of its detection has a peak shift due to time accumulation, resulting in corresponding track detection lag. In order to correct this lag, the present invention first traverses and searches the local peak region along the azimuth dimension, then performs azimuth dilation search on each local peak region, with the search object being the original azimuth spectrum before time accumulation, and finally takes the peak azimuth of the original azimuth spectrum as the corrected azimuth detection output. The method of the present invention has clear logic and simple operation. Compared with the existing method of reducing the accumulation time length to slow down the detection lag, the above method of the present invention can accurately detect track changes, perform real-time correction on the accumulated azimuth spectrum, and will not sacrifice the detection gain brought by time accumulation.

Description

Technical Field

[0001] The present invention relates to the technical field of weak signal detection, and particularly to a real-time correction method for track detection lag. Background Art

[0002] In track detection applications, in order to improve the detection probability of weak targets, the time accumulation effect is usually utilized to accumulate the energy of multiple course azimuth spectra, so that weak targets exceed the detection threshold and at the same time reduce the false alarm probability. However, the accumulated track detection azimuth spectrum reflects the average azimuth of the target in the past period of time. If the target azimuth changes rapidly, the lag deviation error between the accumulated detection azimuth and the true azimuth at the current moment will be relatively significant. To alleviate the above lag phenomenon, either the accumulation time length is reduced to slow down the detection lag, or the azimuth change rate is estimated for correction; the former still exists after the lag phenomenon is alleviated and sacrifices the detection gain brought by time accumulation; the latter increases the computational complexity of real-time estimating the azimuth change rate and additionally introduces parameter estimation errors.

[0003] The above content is only used to assist in understanding the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The main object of the present invention is to provide a real-time correction method for track detection lag, aiming to solve the technical problems of insufficient real-time accuracy and cumbersome calculation in the prior art.

[0005] To achieve the above object, the present invention provides a real-time correction method based on local peak region expansion search, which includes the following steps:

[0006] S1. Produce an accumulated detection azimuth spectrum. According to the time accumulation effect, produce an accumulated detection azimuth spectrum with multiple courses.

[0007] S2. Solve the local peak region of the accumulated detection azimuth spectrum. Traverse and search the local peak region along the azimuth dimension of the accumulated detection azimuth spectrum to detect the local peak region in the accumulated detection azimuth spectrum.

[0008] S3. Peak azimuth correction. Correct the azimuth of the peak in the detected corresponding local peak region according to the original single-course azimuth spectrum.

[0009] In a preferred technical solution, the specific steps of producing the accumulated detection azimuth spectrum in step S1 are as follows:

[0010] S11. Determine the number of courses participating in the azimuth spectrum accumulation according to the single-course duration and engineering suggestions. For example, if the single-course duration is 0.1 second and the engineering suggestion is that the accumulation duration is preferably within 4 seconds, the number of courses participating in the azimuth spectrum accumulation can be selected as 40.

[0011] S12. Multiple single-process azimuth spectra enter the track detection module in real time and accumulate all the received single-process azimuth spectra; until the preset number of processes is accumulated, the sliding accumulation of multiple processes begins. For example, multiple single-process azimuth spectra enter the track detection module in real time. When the number of processes is less than 40 at the initial stage of processing, all the received single-process azimuth spectra are accumulated. That is, when the first process is received, the accumulated azimuth spectrum is the first process itself; when the second process is received, the accumulated azimuth spectrum is the sum of the first and second processes; when the third process is received, the accumulated azimuth spectrum is the sum of the first three processes... and so on. Until the 40th process, the sliding accumulation of 40 processes begins;

[0012] S13. Start sliding accumulation from the preset starting point until the last process to obtain the accumulated detection azimuth spectrum. For example, start sliding accumulation from the 40th process. The accumulated azimuth spectrum of the 40th process is the sum of the 1st to 40th processes. When the 41st process enters the track detection module, what is needed is the sum of the 2nd to 41st processes. On the basis of the accumulated azimuth spectrum of the 40th process, add the azimuth spectrum of the 41st process and subtract the azimuth spectrum of the 1st process to obtain the accumulated azimuth spectrum of the 41st process. And so on for the subsequent processes until the last process (denoted as the Mth process), and the accumulated azimuth spectrum is the sum of the (M - 39)th to Mth processes.

[0013] In the preferred technical solution, the specific steps for solving the local peak region of the accumulated detection azimuth spectrum in step S2 are as follows:

[0014] S21. Let the azimuth number counting variable in the local peak region to be detected be cnt, and the detection flag be flg; then the initial value of the azimuth number counting variable cnt is 0;

[0015] S22. Start searching from the first azimuth of the accumulated detection azimuth spectrum, and perform threshold judgment on the local peak region detection marks corresponding to each azimuth in the accumulated detection azimuth spectrum one by one;

[0016] If it exceeds the initial detection threshold (specific operation: the value of the accumulated detection azimuth spectrum of the current azimuth is greater than the background statistical values of several azimuths on the left and right. The background statistical value is the mean of the two azimuths on the left, the mean of the two azimuths on the right, and then take the maximum value of these two means), then add 1 to the corresponding counting variable cnt, and at the same time set the detection flag flg to 0, indicating that it is in the process of detecting a certain local peak region; if it does not exceed the initial detection threshold, then set the detection flag flg to 1, indicating that it is not in the process of detecting a certain local peak region;

[0017] S23. Determine whether both cnt > 0 and flg > 0 are satisfied in the current search direction. If the determination is negative, indicating that the corresponding local peak region has not been detected or is still in the process of detecting a certain local peak region, then return to the previous step S22 and continue to search for the next direction. Determine whether both cnt > 0 and flg > 0 are satisfied in the current search direction. If the determination is positive, indicating that the detection of a certain local peak region has ended, then enter the next peak direction correction step.

[0018] In the preferred technical solution, the specific steps of peak direction correction in step S3 are as follows:

[0019] S31. Expand the search range of the current local peak region, expand the search range of the current local peak region by n azimuth units to the left and right respectively to obtain an expanded search range for the local peak region.

[0020] S32. Read the current original single - process azimuth spectrum.

[0021] S33. Search for the peak of the original azimuth spectrum within the corresponding expanded search range of the local peak region.

[0022] S34. Use the azimuth where the peak of the original azimuth spectrum searched in S33 is located as the azimuth of the corresponding accumulated detection azimuth spectrum peak, obtain the corrected accumulated detection azimuth spectrum peak, and output the correction.

[0023] In the preferred technical solution, the value of n for expanding the search range of the current local peak region by n azimuth units to the left and right in step S31 is less than or equal to 20.

[0024] In the preferred technical solution, the value of n is 3.

[0025] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0026] 1. The present invention can output the track detection azimuth spectrum in real time.

[0027] 2. When the azimuth of a certain target changes relatively fast, the present invention overcomes the problem that the peak offset caused by time accumulation during the process of detecting the local peak region leads to the lag of the corresponding track detection.

[0028] 3. The present invention first traverses and searches for the local peak region along the azimuth dimension, then performs azimuth expansion search on each local peak region, the search object is the original azimuth spectrum before time accumulation, and finally uses the peak azimuth of the original azimuth spectrum as the corrected azimuth detection output, improving the accuracy and timeliness of the track detection output.

[0029] 4. The method of the present invention has clear logic and simple operations. Compared with the existing method of reducing the accumulation time length to slow down the detection lag, the above method of the present invention can accurately detect the track change, perform real-time correction on the accumulated azimuth spectrum, and will not sacrifice the detection gain brought by time accumulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is a schematic flow chart of a method for real-time correction of track detection lag of the present invention;

[0031] Figure 2 is a schematic flow chart of the local peak region step in Embodiment 1;

[0032] Figure 3 is the original azimuth spectrum in Embodiment 1;

[0033] Figure 4 is the accumulated detection azimuth spectrum in Embodiment 1;

[0034] Figure 5 is a schematic diagram of track output based on the accumulated detection azimuth spectrum;

[0035] Figure 6 is a schematic diagram for explaining the initial detection threshold detection mark of the accumulated detection azimuth spectrum in the first process;

[0036] Figure 7 is a lag comparison diagram of the track based on the accumulated detection azimuth spectrum relative to the original azimuth spectrum;

[0037] Figure 8 is a misalignment comparison diagram of the real-time azimuth spectrum in the 100th process and the track based on the accumulated detection azimuth spectrum;

[0038] Figure 9 is a misalignment comparison diagram of the real-time azimuth spectrum in the 300th process and the track based on the accumulated detection azimuth spectrum;

[0039] Figure 10 is a schematic diagram of track output based on the real-time correction of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0040] Taking the actually received multi-process original azimuth spectrum as an example for illustration, as Figure 3 shown, the horizontal axis is the azimuth, the numerical range is 1° to 360°, the vertical axis is the process index, there are 300 processes in total, the time flow direction is from bottom to top, and one row of azimuth spectrum is output in real time for each single process. For the consistency of gray-scale display after subsequent accumulation, the first 39 processes are not drawn. That is, when calculating the accumulated detection azimuth spectrum in the first process in the figure, the original azimuth spectra of the current and the past 40 processes are already available.

[0041] Figure 3The noise floor background fluctuation of the original azimuth spectrum is relatively significant. Except for the two brighter tracks near 70° and 200°, other tracks are relatively blurred and lack continuity, making real-time detection difficult. To solve this problem, in engineering, the original azimuth spectrum is usually accumulated by sliding along the process, trading time for energy, so that continuously appearing weak targets can be highlighted.

[0042] Figure 4 Shown in Figure 3 is the accumulated detection azimuth spectrum of multiple processes obtained by sliding and accumulating 40 processes of the original azimuth spectrum. That is, one row of the azimuth spectrum of each process is the incoherent accumulation of the azimuth spectra of the previous 40 single processes of this process. The specific operation steps for making the accumulated detection azimuth spectrum are as follows:

[0043] S11. According to the duration of a single process and engineering suggestions, determine the number of processes participating in the azimuth spectrum accumulation. For example, if the duration of a single process is 0.1 second and the engineering suggestion is that the accumulation duration is preferably within 4 seconds, then the number of processes participating in the azimuth spectrum accumulation can be selected as 40;

[0044] S12. Multiple single-process azimuth spectra enter the track detection module in real time. When there are less than 40 processes at the initial stage of processing, all received single-process azimuth spectra are accumulated. That is, when the first process is received, the accumulated azimuth spectrum is the first process itself; when the second process is received, the accumulated azimuth spectrum is the accumulation of the first and second processes; when the third process is received, the accumulated azimuth spectrum is the accumulation of the first three processes... and so on. Until the 40th process, the sliding accumulation of 40 processes begins;

[0045] S13. Start sliding accumulation from the 40th process. The accumulated azimuth spectrum of the 40th process is the accumulation of the 1st to 40th processes. When the 41st process enters the track detection module, what is needed is the accumulation of the 2nd to 41st processes. On the basis of the accumulated azimuth spectrum of the 40th process, adding the azimuth spectrum of the 41st process and then subtracting the azimuth spectrum of the 1st process can obtain the accumulated azimuth spectrum of the 41st process. And so on for the subsequent processes until the last process (denoted as the Mth process), and the accumulated azimuth spectrum is the accumulation of the (M - 39)th to Mth processes.

[0046] Figure 4 Compared with Figure 3 it can be seen that the background of the accumulated azimuth spectrum is darker and smoother, and the continuity of some weak tracks is better. The accumulation brings a certain detection signal-to-noise ratio gain.

[0047] Based on the accumulated detection azimuth spectrum, signal detection is carried out to obtain as Figure 5The track output shown, two tracks with stronger brightness near 70° and near 200°, the detected output of the highlighted track thin line fits perfectly with the original divergence area, truly reflecting the intensity change track of the accumulated azimuth spectrum. The detected output of the highlighted track thin line is Figure 2 The detection result of the "exceeding the preliminary detection threshold" judgment in the flowchart, and each row slice reflected in the figure is a curve along the azimuth. There is a prominent local peak area at each highlighted position. Taking the slice of the first process as an example, as Figure 6 shown, the horizontal axis is the azimuth, the dotted line is the normalized accumulated detection azimuth spectrum of the first process, and the solid line is the previous detection of the accumulated detection azimuth spectrum. An envelope (that is, the peak area) is formed along the azimuth at each local peak area, and this area is enhanced for display. As the process flows through, the Figure 5 highlighted track thin line shown is formed.

[0048] However, when the highlighted track thin line is superimposed on the original azimuth spectrum, as Figure 7 shown, it can be seen that the track detection based on the accumulated azimuth spectrum lags behind the original azimuth spectrum. Taking the slice of the 100th process as an example, as Figure 8 shown, the dotted line is the normalized real-time single-process azimuth spectrum, and the solid line is the normalized accumulated azimuth spectrum of the previous 40 processes. It can be seen that there is a group of local peak areas near 53° in the real-time azimuth spectrum of the 100th process, and the peak is at 53°. However, the peak of the previous accumulated azimuth spectrum is at 55°, and the overall local peak areas are misaligned; the local peak area at 194° is also misaligned to 196°. The reason for the analysis is that at the 100th process, both of these two local peak areas have a tendency to move to the left, that is, the azimuth value gradually becomes smaller, and the accumulated azimuth spectrum reflects the average azimuth of the previous 40 processes, so it does not track the change of the current 100th process in real time, and the track detection has a lag.

[0049] Similarly, the slice of the 300th process is as Figure 9 shown. At this time, the local peak area has a tendency to move to the right, that is, the azimuth value gradually becomes larger, and the real-time peak at 68° is misaligned to 65°, and the real-time peak at 209° is misaligned to 206°. If the track detection result based on the accumulated azimuth spectrum is directly output, there will be an obvious lag phenomenon. In this embodiment, according to the real-time correction method based on the expansion search of the local peak area, the local peak area of the accumulated azimuth spectrum is traversed and searched along the azimuth dimension, and then the azimuth expansion search is performed on each peak area. The search object is the original azimuth spectrum before time accumulation. Finally, the peak of the original azimuth spectrum is used as the corrected detection output. The specific steps are as follows:

[0050] S21. Let the azimuth number counting variable in the local peak area to be detected be cnt, and the detection flag be flg; then the initial value of the azimuth number counting variable cnt is 0;

[0051] S22. Start searching from the first azimuth of the accumulated detection azimuth spectrum, and perform threshold judgment on each local peak corresponding to each azimuth in the accumulated detection azimuth spectrum one by one;

[0052] Since the search will expand left and right in the peak region interval later, start searching from the fourth azimuth from the leftmost. Let's assume the azimuth index of the current search azimuth is kb, that is, the initial value of the azimuth index kb = 3, and the first search kb = 4.

[0053] If it exceeds the initial detection threshold, increment the corresponding count variable cnt by 1, and at the same time set the detection flag flg to 0, indicating that it is in the process of detecting a certain local peak region; if it does not exceed the initial detection threshold, set the detection flag flg to 1, indicating that it is not in the process of detecting a certain local peak region;

[0054] S23. Determine whether both cnt > 0 and flg > 0 are satisfied under the current search azimuth. If the judgment is no, indicating that the peak region has not been detected or is still in the process of detecting a certain local peak region, then return to the previous step S22 and continue to search the next azimuth; determine whether both cnt > 0 and flg > 0 are satisfied under the current search azimuth. If the judgment is yes, indicating that the detection of a certain local peak region is completed, then enter the next peak azimuth correction step, and at the same time reset cnt to 0. If the current azimuth index kb is less than the total azimuth index Nb - 3 in the current process (Nb refers to the total number of azimuths in the current process), then continue to search the next azimuth; otherwise, end the current process correction and enter the next process correction.

[0055] The specific steps for peak azimuth correction are as follows:

[0056] S31. Expand the search interval of the current local peak region, expand the search interval of the current local peak region by three azimuth units to the left and right respectively to obtain the expanded search interval of the local peak region; assume the index of the current search azimuth is kb, the right endpoint of the azimuth interval of the detected peak region is kb - 1, and the left endpoint of the interval is kb - cnt, then the expanded search interval is [kb - cnt - 3, kb - 4];

[0057] S32. Read the current original single - process azimuth spectrum;

[0058] S33. Search for the original azimuth spectrum peak in the corresponding expanded search interval of the local peak region;

[0059] S34. Take the azimuth where the original azimuth spectrum peak searched in S33 is located as the azimuth of the corresponding accumulated detection azimuth spectrum peak, obtain the corrected accumulated detection azimuth spectrum peak, and output the correction. The corrected detection process diagram is as Figure 10 shown, compared with Figure 7, the detection lag has been significantly improved.

[0060] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. It should be noted that any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A real-time correction method for track detection lag, characterized in that: It includes the following steps: S1. Produce an accumulated detection azimuth spectrum. According to the time accumulation effect, produce an accumulated detection azimuth spectrum with multiple processes. S2. Solve the local peak region of the accumulated detection azimuth spectrum. Traverse and search the local peak region along the azimuth dimension of the accumulated detection azimuth spectrum to detect the local peak region in the accumulated detection azimuth spectrum. S3. Peak azimuth correction. Correct the azimuth of the peak in the detected local peak region according to the original single-process azimuth spectrum. The specific steps of peak azimuth correction in the above step S3 are as follows: S31. Expand the search range of the current local peak region. Expand the search range of the current local peak region by n azimuth units to the left and right respectively to obtain the expanded search range of the local peak region. S32. Read the current original single-process azimuth spectrum. S33. Search for the original azimuth spectrum peak in the corresponding expanded search range of the local peak region. S34. Take the azimuth where the original azimuth spectrum peak searched in S33 is located as the azimuth of the corresponding peak of the accumulated detection azimuth spectrum, obtain the corrected peak of the accumulated detection azimuth spectrum, and output the correction.

2. The real-time correction method for detecting a lag in a flight path according to claim 1, characterized in that : The specific steps of producing the accumulated detection azimuth spectrum in the above step S1 are as follows: S11. Determine the number of processes participating in the accumulation of the accumulated detection azimuth spectrum according to the duration of the original single process and engineering suggestions. S12. Multiple single-process azimuth spectra enter the track detection module in real time, and all received single-process azimuth spectra are accumulated. Until the preset number of processes is accumulated, start the sliding accumulation of multiple processes. S13. Start sliding accumulation from the preset start until the last process to produce the accumulated detection azimuth spectrum.

3. A real-time correction method for detecting and lagging a track according to claim 2, characterized in that : The specific steps of solving the local peak region of the accumulated detection azimuth spectrum in the above step S2 are as follows: S21. Let the azimuth number counting variable in the local peak region to be detected be cnt, and the detection flag be flg; then the initial value of the azimuth number counting variable cnt is 0. S22. Start searching from the first azimuth of the accumulated detection azimuth spectrum, and perform threshold decision on the detection marks of the local peak regions corresponding to each azimuth in the accumulated detection azimuth spectrum one by one. If it exceeds the preliminary detection threshold, add 1 to the corresponding counting variable cnt, and at the same time set the detection flag flg to 0, indicating that it is in the process of detecting a certain local peak region; if it does not exceed the preliminary detection threshold, set the detection flag flg to 1, indicating that it is not in the process of detecting a certain local peak region. S23. Judge whether at the current search azimuth, cnt>0 and flg>0 are satisfied at the same time. If the judgment is no, it means that the corresponding local peak region has not been detected or is still in the process of detecting a certain local peak region, then return to the previous step S22 and continue to search for the next azimuth; judge whether at the current search azimuth, cnt>0 and flg>0 are satisfied at the same time. If the judgment is yes, it means that the detection of a certain local peak region is over, and then enter the next peak azimuth correction step.

4. The real-time correction method for the lag of track detection according to claim 3, characterized in that: In the above step S31, the value of n for expanding the search range of the current local peak region by n azimuth units to the left and right respectively is less than or equal to 20.

5. A real-time correction method for detecting lag of a track, according to claim 4, characterized in that: The value of n is 3.

Images