A method for pulse Doppler radar image signal compression and transmission

Abstract

This invention discloses a method for compressing and transmitting pulse Doppler radar image signals, comprising the following steps: S1, acquiring pulse Doppler radar RD data and performing CFAR calculations on the RD data; S2, compressing the pulse Doppler radar RD data using the CFAR calculation results; S3, transmitting the compressed pulse Doppler radar RD data; S4, forming an image from a reconstructed RD map and map subframes using the pulse Doppler radar RD data. This invention can transmit the main information contained in the RD map with minimal loss of pulse Doppler radar RD map information using a smaller bandwidth.

Description

Technical Field

[0001] This invention relates to the field of signal processing technology, and more specifically to a method for compression and transmission of pulse Doppler radar image signals. Background Technology

[0002] Pulse Doppler radar signal processing systems are characterized by high floating-point computation density, high parallelism, and centralized algorithms. The resulting detection results mainly include information such as the presence of the detected target, its range, Doppler velocity, and the intensity of the target's reflected signal. Formally, the presence of the target and related information can be extracted by interpreting the RD (range-Doppler) domain image of the radar signal. The RD map is the foundation of pulse Doppler radar information interpretation, and the target position is usually directly obtained by the CFAR (Constant False Alarm Rate) algorithm. The result of radar information processing is equivalent to transmitting one or more points on the RD map, i.e., the target's location and Doppler velocity, intensity, etc. Because targets vary in form, the target patterns displayed on the RD map are also diverse. Manual observation of the RD map and subsequent RD map pattern recognition are necessary. Furthermore, the CFAR detection location and the image reading location may be in different locations, requiring transmission and storage for later processing.

[0003] Current methods for compressing and transmitting RD images typically employ image processing compression algorithms. These algorithms generally perform correlation operations between the original image and the basis functions (also known as kernel functions) used for compression, storing the correlation coefficients between the image and the basis functions of each order. During subsequent reconstruction, these coefficients are multiplied by the known basis functions of each order and then summed to restore the original image. Basis functions can be chosen from sine, cosine, or wavelet functions, prioritizing the preservation of low-frequency information while appropriately ignoring high-frequency information, thus preserving the main features of the image. However, RD images are not images in the traditional sense; pixels are generally floating-point numbers rather than fixed-point numbers. Basis functions may not be suitable for the characteristics of RD images, and the parts preserved by traditional compression algorithms may not be what is needed. The compression effect may not be what is required for subsequent observation and processing of RD images. Furthermore, image compression algorithms, even lossless ones, are not essential for radar information processing. They require additional logic or computational resources to complete the calculations. The related operations in image information compression are generally complex, and fast, low-resource algorithms are not always available. Therefore, it is necessary to find algorithms more suited to the characteristics of RD images and integrate them with radar algorithms to achieve better results. Summary of the Invention

[0004] To address the aforementioned shortcomings in the existing technology, the present invention provides a method for compression and transmission of pulse Doppler radar image signals, which solves the problem of compressing and transmitting RD images.

[0005] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows: a method for pulse Doppler radar image signal compression and transmission, comprising the following steps:

[0006] S1. Acquire pulse Doppler radar RD data and perform CFAR calculations on the RD data;

[0007] S2. Compress the pulse Doppler radar RD data using the CFAR calculation results;

[0008] S3. Transmit compressed pulse Doppler radar RD data;

[0009] S4. Using pulse Doppler radar RD data, a reconstructed RD map and an image formed from map subframes are generated.

[0010] Further: Step S1 specifically includes:

[0011] S11. For pulse Doppler radar, after transmitting electromagnetic pulses and being reflected by the target, several pulse trains are formed. Pulse compression calculation is performed on each pulse to complete the range and direction calculation.

[0012] S12. Perform Doppler accumulation processing by accumulating Doppler pulses through several pulse trains, and arrange the distance and Doppler accumulated data in a two-dimensional manner to form an RD image;

[0013] S13. Perform CFAR calculation to detect whether the target point exists. If it exists, obtain the range position and Doppler velocity of the target point. Otherwise, determine whether to transmit map subframes.

[0014] Furthermore: The CFAR operation adopts the CA-CFAR algorithm or other algorithms, such as G0-CFAR, SO-CFARWCA algorithm. The purpose of the CFAR operation is to give one or more signal points as target points, or to assume that there is no target. If there is a target, the reflected energy of the signal point reflected by the target on the RD image is significantly higher than that of the surrounding points, and also higher than that of the noise. The ratio of the signal point to the noise is higher than a preset fixed value.

[0015] Further: Step S2 specifically includes:

[0016] S21. Using CFAR calculations to obtain the number of targets in the RD map, the position of each target in the RD, and given the range of the RD map associated with each target, several CFAR keyframes are formed.

[0017] S22. Place several CFAR key subframes into the transmission frame. If the transmission frame is not full, proceed to step S23. If the transmission frame is full, transmit directly.

[0018] S23. Determine whether map subframes need to be transmitted. If yes, proceed to step S24; otherwise, proceed to step S25.

[0019] S24. Extract the RD portion of the image where the Doppler velocity is 0 to form a map subframe. Place the map subframe into the transmission frame. If the transmission frame is full, transmit directly. If the transmission frame is not full, proceed to step S25.

[0020] S25. After the idle frames are placed into the remaining positions of the transmission frame in turn, they are transmitted.

[0021] Furthermore: the idle frame is the portion of the RD image excluding the CFAR key subframe and the map subframe.

[0022] Furthermore: the transmission frame consists of a frame header and data. The frame header defines the size of the transmission packet, the number of CFAR key subframes it contains, the position of each CFAR key subframe in the RD map, the waveform information, the CFAR key subframe data, whether it includes map subframes, the map subframe data, whether it includes idle subframe positions, the idle subframe position data, and the idle subframe data.

[0023] Further: Step S4 specifically involves: using the obtained CFAR keyframe image, restoring the RD image at the corresponding position of the CFAR keyframe, filling the RD image with the obtained free subframes where there are no CFAR keyframes, and setting the image data at the RD image to zero if there are no free subframes. Combining the map subframes scanned in the azimuth direction into a map formed by the reflection of static objects, if no map subframe is obtained in the azimuth dimension, it is used to replace the previously obtained map subframe, and if there is no previous map subframe, the image data at that location is set to zero.

[0024] The beneficial effects of this invention are as follows: This invention utilizes the CFAR correlation algorithm detected by radar targets to indicate key information areas in the RD map, reducing the transmission of non-critical information. Simultaneously, it maximizes the use of transmission bandwidth resources, transmitting information captured by other radar RD maps when the amount of critical information to be transmitted is small, thus making full use of bandwidth resources and RD map information. This invention can transmit the main information contained in the RD map using a relatively small bandwidth without significantly losing pulse Doppler radar RD map information. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the overall system.

[0026] Figure 2 This is a schematic diagram illustrating the formation and transmission components of an RD diagram;

[0027] Figure 3 This is a flowchart illustrating the formation of a transmission frame.

[0028] Figure 4 This is a diagram showing the format of the transmission frame;

[0029] Figure 5 The RD diagram recovered in the example;

[0030] Figure 6 The "map" in this embodiment is a static target composed of map subframes. Detailed Implementation

[0031] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0032] For current radar systems using DSP+FPGA or GPU platforms, the devices that display and store images are generally not located at the same location as the radar processing unit, requiring a certain transmission process. Simultaneously, the display unit typically needs to display multiple types of information—information from multiple radars and sensors. A single radar may need to share transmission bandwidth with other sensors, such as… Figure 1 As shown, based on the characteristics of the above platform, this invention adopts a basis function guided by the CFAR function, which allows the CFAR calculation process to point out the features of the RD image while retaining some RD image information according to the characteristics of radar detection, thereby realizing the information compression and transmission of the RD image.

[0033] The basis function processing mechanism used in this system preserves the Doppler information of the target of interest to pulse-Doppler radar as much as possible, enabling feature extraction in RD (Radar Radar Detection) to a certain extent to fuse the radar processing flow, but it is not simply about obtaining target descriptors. This enriches the observation of the target, allowing for further analysis of the target's features displayed on the RD map in subsequent processing, including manually observed images or machine learning-based RD images. It also assists in environmental detection. For example, in pulse-Doppler radars with mechanical or electronic phase scanning, if the radar vehicle is stationary, the zero-velocity portion in the RD map is generated by the accumulation of radar waves reflected from stationary objects in the environment. By recording and transmitting the stationary object reflection data shown in the RD at each wave position as important data, and adding the RD data of moving targets, a static radiation environment map can be formed, simultaneously obtaining the position and possible RD patterns of moving targets.

[0034] This method includes the acquisition of complete RD data from pulse Doppler radar and the screening and transmission of key data. The process is as follows: Figure 2 As shown. The main processing steps include:

[0035] (1) Acquisition of pulse Doppler radar RD data and CFAR calculation

[0036] First, the pulse Doppler radar utilizes the emitted electromagnetic pulses, which are reflected by the target to form several pulse trains. Pulse compression is performed on each pulse to calculate the range and direction. Then, Doppler accumulation processing is performed, accumulating Doppler pulses through several pulses. The range and Doppler accumulated data are arranged in a two-dimensional format to form an RD image.

[0037] Next, CFAR calculations are performed to detect the presence of the target and obtain its range position and Doppler velocity. CFAR processing can be carried out in various ways, including using CA-CFAR or other algorithms such as G0 (Greatest of)-CFAR, SO (Smallest of)-CFAR, and WCA (Weighted Cell-Averaging). The purpose of CFAR calculations is to identify one or more points whose reflected energy in the RD image is generally significantly higher than that of surrounding points and significantly higher than that of noise. The ratio of signal points to noise is higher than a fixed value, which is generally related to a fixed false alarm rate and missed alarm rate. For a given target point, it will form a certain pattern with surrounding points. Generally, if it is a point target reflection, it will form a SINC function-like intensity distribution in both the range and Doppler dimensions. If the distribution significantly deviates from the SINC function form, the target may exhibit features such as tumbling, multiple small target clusters, or other moving parts on the target.

[0038] (2) Pulse Doppler radar RD data compression

[0039] The CFAR operation provides the location of the target point in the RD image, and its surrounding points are typically regions where the target's characteristics are concentrated. Outputting the RD image information around the point from the CFAR operation effectively compresses the RD image. The target point location output by CFAR can be viewed as the origin of a radial basis function. The main function range of the basis function can be defined using the range of window functions for range and Doppler velocity in pulse Doppler processing. The basis function can be in the form of a window function or other forms. If a rectangular window function is used, the result of the correlation between the image function and the basis function within the window is the image function itself, requiring no special calculations, simplifying ordinary compression. Outside the window, since it mainly consists of noise information, it can be directly set to zero without calculation or transmission. Since the higher-order parts of the basis function are not very meaningful for the RD image, they can be omitted and not calculated. For scenarios where there may be multiple targets, image values ​​can be extracted from multiple windows. Generally, in practical applications of pulse Doppler radar, the number of targets detected by CFAR in a single RD image frame is not large, or in other words, it's unlikely that most image points in the RD image are targets. Therefore, the range covered by the basis function window is much smaller than the entire RD image range, thus achieving RD image compression. Based on this, RD image compression involves selecting a portion of the RD image. The selection method uses the results of CFAR calculations to indicate the center position of the selected portion. The selection range can also be determined by CFAR calculations, or by the target to be detected by the radar and the radar's own parameters, such as pixels surrounding the center point (e.g., 16 or 32 surrounding pixels, or only the center point itself). After selecting a portion of the RD image, this portion can be directly transmitted, or the sub-image can be further compressed using traditional image compression. Compared to further image compression, since the selected portion already has a small data volume, direct transmission does not require additional computational resources, making it a feasible method.

[0040] Compared to simply transmitting the center point provided by CFAR, transmitting the RD image around the center point allows for the observation and analysis of target characteristics. For example, when a flying target exhibits tumbling behavior or has small moving parts, the RD image can be used to analyze the target's behavior and derive more relevant information. This is the significance of transmitting RD images. The CFAR point and its surrounding points are referred to below as the "CFAR keyframes" of the RD image.

[0041] For ground-based radars observing near-ground targets, due to the small observation elevation angle or the possibility of looking down, the reflected signals from stationary or quasi-stationary objects in the ground environment are usually filtered out as clutter by Doppler accumulation during CFAR processing. However, this clutter may still be beneficial to the overall information acquired by the radar. It might be considered that by opening a window at the Doppler DC (signal reflected by a stationary object) in the RD map, the Doppler DC signals in the RD domain from multiple scan positions can form the outline of a stationary ground object, facilitating the analysis of the relationship between the radar, the target, and the environment. This part is not the area indicated by pulse CFAR. The set of points of the Doppler DC signal in the RD map is referred to below as the "map subframe" of the RD map. In contrast to the "CFAR key frame," the "map subframe" is the "CFAR key frame" that uses the Doppler DC point in the RD map as its "center point." However, this "center point" is a straight line in the RD map, representing the signal reflected by a stationary object. The area selected during compression is a strip-shaped region near the Doppler DC image in the RD map, usually only a straight line.

[0042] When no CFAR target is detected, a portion of the RD image can be transmitted sequentially. Since the radar has a certain scanning speed, the entire RD image can be obtained after a certain period of time. Although the different parts of the RD image are not images of the same frame, during manual observation, the CFAR may be pointed to the RD image outside the window until the part is updated. It is possible to discover some reflectors that the CFAR algorithm ignores. At the same time, the parts outside the non-CFAR key subframes and map subframes are referred to as idle frames. The update of idle frames can obtain dynamic RD images, which are convenient for manual observation.

[0043] The process of forming a transmission frame is as follows: Figure 3 As shown, specifically:

[0044] S21. Using CFAR calculations to obtain the number of targets in the RD map, the position of each target in the RD, and given the range of the RD map associated with each target, several CFAR keyframes are formed.

[0045] S22. Place several CFAR key subframes into the transmission frame. If the transmission frame is not full, proceed to step S23. If the transmission frame is full, transmit directly.

[0046] S23. Determine whether map subframes need to be transmitted. If yes, proceed to step S24; otherwise, proceed to step S25.

[0047] S24. Extract the RD portion of the image where the Doppler velocity is 0 to form a map subframe. Place the map subframe into the transmission frame. If the transmission frame is full, transmit directly. If the transmission frame is not full, proceed to step S25.

[0048] S25. After the idle frames are placed into the remaining positions of the transmission frame in turn, they are transmitted.

[0049] (3) Transmitting pulse Doppler radar RD data compression

[0050] The transmission and updating of the three types of RD maps mentioned above can be adjusted according to the actual system's transmission capacity, setting certain weights to ensure the richness of the acquired information. For example, the update area indicated by CFAR, i.e., the CFAR key subframe, generally has the highest priority. If the transmission bandwidth is insufficient, map subframes can be updated, and finally, idle frames without a target are updated and transmitted. Transmission can be carried out through ordinary transmission tools or lines such as Ethernet, serial buses, and wireless channel transmission tools. A fixed transmission frame format can be formed during transmission, consisting of a frame header and data. The frame header defines the size of the transmission packet, the number of CFAR key subframes included, the position of each CFAR key subframe in the RD map, waveform information, CFAR key subframe data, whether map subframes are included, map subframe data, whether idle subframe positions are included, idle subframe position data, and idle subframe data. The format of a transmission frame is as follows: Figure 4 As shown. Based on the characteristics of the radar system, a frame size and frame transmission frequency are specified. The frame transmission frequency is generally the same as the frequency of radar wave position changes. The frame size is determined by the transmission bandwidth and frequency that the system can provide. Depending on the number of targets in the CFAR and whether map subframes need to be transmitted, the remaining part is filled into idle frames in sequence to achieve transmission.

[0051] (4) The way of perceiving images after transmission

[0052] Using the currently obtained CFAR keyframe images, the RD image is reconstructed at the corresponding locations. Areas in the PD image where no data is available are filled with previously obtained idle subframes. If no idle subframe data is available, the image data at that location can be set to zero. Map subframes scanned in the azimuth direction are combined to form a map formed by reflections of static objects. If a map subframe for that direction dimension is not currently available, it is replaced with a previously obtained map subframe; if no previous map subframe is available, the image at that location is set to zero. The image form resulting from the reconstructed RD map and map subframes is as follows: Figure 5 and Figure 6 As shown.

[0053] Example 1

[0054] A complete pulse-Doppler radar RD map contains 2048*512 pixels, with 2048 in the range direction and 512 in the Doppler velocity direction. Each pixel includes 32 bits of data, generating 40 frames per second. If all transmissions require a bandwidth of 160MB / s, and a CFAR keyframe contains 16*16 pixels, assuming CFAR detects fewer than 10 targets in each complete RD map, transmitting all CFAR keyframes would only require a bandwidth of 400KB / s. If CFAR does not detect any targets, each RD map can output one map subframe and two idle frames with a bandwidth of 400KB / s. The idle frames update a complete RD map after 51.2 seconds.

[0055] Example 2

[0056] In hybrid update mode, the radar completes a 360-degree mechanical circular scan in 4 seconds. The target elevation angle is measured via antenna sum and difference beams. If the radar remains stationary, it will not update map subframes in real time after several scans. The time to update a complete RD map in idle frames can be further shortened. Alternatively, idle frames can be used to extract RD maps focusing on a key area of ​​interest, making the RD map information more meaningful.

[0057] Example 3

[0058] When the distribution of CFAR points is concentrated or has a unique distribution in the velocity or range dimensions, such as when the target's expansion on the RD map is significantly different from the distribution of ordinary point targets, the range of the keyframes can be varied to form variations such as 32*32, 16*32, and 32*16. This allows the obtained keyframes to better reflect the characteristics of the target.

Claims

1. A method of pulse Doppler radar image signal compression and transmission, characterized by, The method comprises the following steps: S1, acquiring pulse Doppler radar RD data and performing CFAR operation on the RD data; The step S1 is specifically: S11, performing pulse compression operation on each pulse of a plurality of pulse trains formed by a target after reflection of a transmitted electromagnetic pulse train by the target, to complete operation in the distance direction; S12, performing Doppler accumulation processing, accumulating Doppler pulses through the plurality of pulse trains, and arranging the distance and Doppler accumulation data in a two-dimensional manner to form an RD image; S13, performing CFAR operation to detect whether a target point exists, and if the target point exists, acquiring the distance position and Doppler velocity of the target point, or judging whether a map subframe needs to be transmitted; S2, compressing the pulse Doppler radar RD data according to the CFAR operation result; The step S2 is specifically: S21, obtaining the number of targets in the RD image, the position of each target in the RD image, and the range of the RD image associated with each target by using the CFAR operation, to form a plurality of CFAR key subframes; S22, placing the plurality of CFAR key subframes in a transmission frame, and if the transmission frame is not full, proceeding to step S23, or if the transmission frame is full, directly transmitting; S23, judging whether a map subframe needs to be transmitted, and if yes, proceeding to step S24, or if no, proceeding to step S25; S24, extracting an RD partial image with a Doppler velocity of 0 to form a map subframe, placing the map subframe in the transmission frame, and if the transmission frame is full, directly transmitting, or if the transmission frame is not full, proceeding to step S25; S25, placing idle frames in the remaining positions of the transmission frame in turn and then transmitting; S3, transmitting the compressed pulse Doppler radar RD data; S4, forming a reorganized RD image and an image formed by a map subframe by using the pulse Doppler radar RD data; The step S4 is specifically: restoring the RD image at the corresponding position of the CFAR key subframe by using the image of the obtained CFAR key subframe, filling the RD image at the position without the CFAR key subframe with an obtained idle subframe, setting the image data at the position to zero if there is no idle subframe, combining the map subframes in the azimuth direction to form a map reflected by a static object, and replacing the previously obtained map subframe with the map subframe if no map subframe is obtained in the azimuth direction, or setting the image data at the position to zero if there is no previous map subframe.

2. The method of pulse-Doppler radar image signal compression and transmission according to claim 1, characterized in that, The CFAR operation adopts a CA-CFAR algorithm or a G0-CFAR, SO-CFAR or WCA algorithm. The purpose of the CFAR operation is to give one or more signal points as target points, or to consider that there is no target. If there is a target, the signal point reflected by the target has a reflection energy obviously higher than that of the surrounding points and higher than noise in the RD image, and the ratio of the signal point to the noise is higher than a pre-set fixed value.

3. The method of pulsed Doppler radar image signal compression and transmission according to claim 1, characterized in that, The idle frame is a part of the RD image excluding the CFAR key subframe and the map subframe.

4. The method of pulse-Doppler radar image signal compression and transmission according to claim 1, characterized in that, The transmission frame is composed of a frame header and data, the frame header defines the size of the transmission packet, the number of CFAR key sub-frames contained, the location of the RD map of each CFAR key sub-frame, the wave position information, the CFAR key sub-frame data, whether to include a map sub-frame, the map sub-frame data, whether to include the idle sub-frame position, the idle sub-frame position data, and the idle sub-frame data.

Images