A multi-radar fusion method based on marine radar

Abstract

A multi-radar fusion method based on marine radar provided by the present invention preprocesses the detection data of each radar for a target ship and calculates the corresponding quality factor; according to the quality factor, the relationship probability between each radar and the target ship is established through joint probability density; the fused track of the target ship is calculated according to the relationship probability. By calculating the radar center point, all radar data are converted into a rectangular coordinate system with the center point as the origin and the positive y-axis pointing north. The joint probability and the adaptive weighting algorithm (i.e., the quality factor) are combined to calculate the probability density of each radar. The detection data of each radar are multiplied by the relationship probability with the target ship and then added together to obtain the fused track of the target ship, which is the track point conforming to the total track, reducing the systematic error and human error, reducing the calculation amount, and improving the accuracy of radar fusion when the gates overlap.

Description

Technical Field

[0001] The present invention relates to the field of radar technology, and particularly to a multi-radar fusion method based on marine radar. Background Art

[0002] In the early stage of the development of radar technology, it was necessary for a special person to read marine targets and target-related information from the radar display screen. In modern warfare, higher requirements are placed on the accuracy of radar detection and the timeliness of target reading. The method of manually operating the radar cannot meet the needs of modern warfare. Combining radar technology with technologies such as electronic computer technology, image information processing, and artificial intelligence, the information processing method of the radar has changed from manual operation to computer automatic target capture. In modern marine radars, an automatic radar plotting aid (ARPA) based on electronic computer technology is equipped.

[0003] Compared with traditional manual operation, the ARPA system composed of a log, a compass, and a radar can capture multiple radar targets manually or automatically and perform automatic tracking. The system can display the speed and course of the target on the radar chart in the form of vectors for the processed data, display multiple important collision avoidance data such as CPA and TCPA through data, and can also perform various functions such as alarm, collision avoidance risk calculation, and avoidance.

[0004] The data fusion scheme of the radar data fusion system is divided into three types based on the processing stage of the radar signal: detection-level fusion, measurement-level fusion, and estimation-level fusion.

[0005] Detection-level fusion is the fusion at the level of the original information collected, that is, analysis and synthesis are carried out before the echo received by the radar is preprocessed. First, the validity of the signal is determined, and then data fusion is performed. This fusion stage can retain the original data information as much as possible, and the data reliability is very high, but at the same time, the data processing volume is large and the processing time is long.

[0006] Measurement-level fusion refers to directly fusing the measurements of each sensor after registering the measurements during the data processing process, that is, comprehensively analyzing the measurements of each sensor without processing. The advantage of measurement-level fusion is that it can provide more accurate information than the estimation-level fusion method. The disadvantage is that the amount of calculation is large, the registration of each radar is required to be consistent, and the real-time performance is not strong.

[0007] Estimation-level fusion belongs to the highest level of fusion. Each sensor in the system independently processes the sampled data to obtain local estimation results, and then these local estimation information is transmitted to the fusion center. The fusion center fuses the various local results obtained to finally obtain a global estimation.

[0008] During navigation, a single radar may be affected by the external environment, resulting in data failure. Therefore, multiple radars are usually installed on ships, and the data of multiple radars are fused. By reasonably fusing the detection data from multiple radars, not only can the monitoring range be directly expanded to all the sea areas covered by each radar, but also the target monitoring quality, the reliability of the system itself and the anti-impact ability can be comprehensively improved. This technology has become an important part of offshore navigation and port environment monitoring.

[0009] Therefore, a multi-radar fusion method based on marine radar is needed, which can reduce system errors and human errors, reduce the amount of calculation, and improve the accuracy of radar fusion when the gates overlap. Summary of the Invention

[0010] (1) Technical problems to be solved

[0011] To solve the above problems of the prior art, the present invention provides a multi-radar fusion method based on marine radar.

[0012] (2) Technical solutions

[0013] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0014] A multi-radar fusion method based on marine radar includes the steps of:

[0015] S1. Obtain the detection data of each radar for the target ship, perform preprocessing, and calculate the corresponding quality factor;

[0016] S2. Establish the relationship probability between each radar and the target ship according to the quality factor through the joint probability density;

[0017] S3. Calculate the fused track of the target ship according to the relationship probability.

[0018] (3) Beneficial effects

[0019] The beneficial effects of the present invention are as follows: By obtaining the detection data of each radar for the target ship and performing preprocessing, and calculating the corresponding quality factor; According to the quality factor, through the joint probability density, establishing the relationship probability between each radar and the target ship; Calculating the fusion track of the target ship according to the relationship probability, by calculating the radar center point, converting all radar data into a rectangular coordinate system with the center point as the origin and the positive y-axis pointing north, combining the joint probability and the adaptive weighting algorithm (i.e., the quality factor), calculating the probability density of each radar, multiplying the detection data of each radar by the relationship probability with the target ship and then adding them up to obtain the fusion track of the target ship, which is the track point that conforms to the total track, reducing the system error and human error, reducing the calculation amount, and improving the accuracy of radar fusion when the gates overlap. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a flowchart of a multi-radar fusion method based on marine radar according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] In order to better explain the present invention for easy understanding, the present invention will be described in detail below with reference to the drawings and through specific embodiments.

[0022] Please refer to Figure 1 , a multi-radar fusion method based on marine radar, comprising the steps of:

[0023] S1. Obtain the detection data of each radar for the target ship, perform preprocessing, and calculate the corresponding quality factor;

[0024] The detection data specifically includes detection time, longitude, latitude, speed and heading of the target ship.

[0025] The preprocessing specifically includes:

[0026] Converting the detection data of each radar in polar coordinates into a rectangular coordinate system with the center radar as the origin;

[0027] Registering the time information of each radar based on the time of the center radar;

[0028] Setting a gate to filter clutter, thereby reducing the calculation amount.

[0029] Specifically, some alternative observations can be filtered by physical or statistical methods (the observation data refers to the data of each radar on the target ship, such as detection time, longitude of the target, latitude of the target, speed of the target, and course of the target). These observations may form unlikely or undesired association pairs with alternative entities, and at the same time, these observations may also come from noise and interference. The trajectory detection ranges of each radar overlap, and clutter is filtered by setting a gate to reduce the amount of calculation. The gates are divided into large gates, medium gates, and small gates.

[0030] Small gate: mainly applied to track targets in a non-maneuvering state of uniform linear motion;

[0031] Medium gate: mainly applied to track targets in a small maneuvering state;

[0032] Large gate: can be used in two cases, namely when the target makes a large maneuvering turn or when the target is lost and needs to be recaptured.

[0033] The calculation formula of the quality factor is as follows:

[0034] Q i = f i + F i + D i + h i

[0035] Where Q i is the quality factor of the i-th radar, f i is the frequency weight of the i-th radar, F i is the goodness-of-fit weight of the i-th radar, D i is the trust weight of the i-th radar, and h(i) is the grey relational degree weight of the i-th radar;

[0036]

[0037] In the formula, count i is the number of reports of the i-th radar, and count is the total number of reports of all radars;

[0038]

[0039] In the formula, is the goodness-of-fit degree of the abscissa x i of the i-th radar on the X-axis, is the goodness-of-fit degree of the ordinate y i of the i-th radar on the Y-axis;

[0040]

[0041]

[0042]

[0043] Wherein, SSE is the sum of squared errors, SSR is the regression sum of squares, y represents the actual navigation trajectory point of the target ship, represents the detection value of the current radar, that is, the ordinate y of the detection point of the i-th radar i , represents the average value of the radar detection values of all detected target ships;

[0044]

[0045] And d 1 + v 1 + θ 1 = 1, d 1 < v 1 < θ 1

[0046] Wherein, D i is the change between the calculated track and the actual track of the i-th radar in terms of distance, course angle, and ship speed;

[0047] Specifically, the calculated track refers to the track detected by each radar, and the actual track refers to the actual track of the target ship.

[0048]

[0049] Wherein, n represents the total number of radars, the value range of i is from 1 to n, x 0 represents the actual navigation trajectory of the target ship, x j is the track detected by the i-th radar at time j, and the value range of j is from 1 to m.

[0050] S2. Establish the relationship probability between each radar and the target ship according to the quality factor through the joint probability density;

[0051] Step S2 is specifically as follows:

[0052] Establish the relationship probability D(k) between the i-th radar and the target ship at time k through the joint probability density;

[0053]

[0054] Wherein, G i (k) is a quantity proportional to the Gaussian likelihood function, indicating the degree of proximity between the i-th radar and the target ship;

[0055] v i (k) is a trajectory association indicator, when the i-th radar is in the association area of the target ship, v i(k) = 1, otherwise v i (k) = 0.

[0056] Specifically, the associated area of the target ship specifically refers to the radar detection track within the gate range around the actual track of the target ship.

[0057] S3. Calculate the fused track of the target ship according to the relationship probability.

[0058] Step S3 is specifically as follows:

[0059] Multiply the detection data of each radar by the relationship probability with the target ship and then sum them up to obtain the fused track of the target ship.

[0060] The above are only embodiments of the present invention, and do not limit the patent scope of the present invention. Any equivalent transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in the relevant technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. A multi-radar fusion method based on marine radar, characterized in that, it includes the steps of: S1. Obtain the detection data of each radar for the target ship, preprocess it, and calculate the corresponding quality factor; S2. Based on the quality factor, establish the relationship probability between each radar and the target ship through the joint probability density; S3. Calculate the fused track of the target ship according to the relationship probability; The specific content of step S2 is: Establish through the joint probability density At the moment, the relationship probability between the ; Among them, is a quantity proportional to the Gaussian likelihood function, representing the proximity between the th radar and the target ship, is the quality factor of the th radar; is a trajectory correlation indicator. When the th radar is in the correlation area of the target ship , otherwise .

2. The multi-radar fusion method based on marine radar according to claim 1, characterized in that, the detection data specifically includes the detection time, longitude, latitude, speed and heading of the target ship.

3. The multi-radar fusion method based on marine radar according to claim 1, characterized in that, the preprocessing specifically includes: Convert the detection data of each radar in polar coordinates to a rectangular coordinate system with the center radar as the origin; Register the time information of each radar based on the time of the center radar; Set a gate to filter clutter.

4. The multi-radar fusion method based on marine radar according to claim 1, characterized in that, the calculation formula of the quality factor is as follows: Among them, is the quality factor of the th radar, is the frequency weight of the th radar, is the fitting degree weight of the th radar, is the change in distance, course angle, and ship speed between the calculated track and the actual track of the th radar, h i is the grey relational degree weight of the th radar; Wherein, is the reporting frequency of the th radar, is the total reporting frequency of all radars; Wherein, is the abscissa of the detection point of the th radar, representing the fitting degree on the X-axis, is the ordinate of the detection point of the th radar, representing the fitting degree on the Y-axis; Wherein, is the difference of squares, is the regression sum of squares, represents the ordinate of the actual navigation trajectory point of the target ship, represents the th ordinate of the detection point of the radar, represents the average value of the ordinates of the radar detection values of all radars that detect the target ship; and , Wherein, n represents the total number of radars, ranges from 1 to n, represents the actual navigation track of the target ship, is the track detected by the th radar at ranges from 1 to m.

5. The multi-radar fusion method based on marine radar according to claim 1, characterized in that, The specific content of step S3 is: Multiply the detection data of each radar by the relationship probability with the target ship and then add them to obtain the fused track of the target ship.

Images