Target detection device and radar device
By dividing the radar device's antenna rotation into sectors and performing target detection using sector data, the device suppresses false detections and reduces processing load, enhancing target identification accuracy.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- JAPAN RADIO CO LTD
- Filing Date
- 2022-06-17
- Publication Date
- 2026-06-01
AI Technical Summary
Radar systems that detect targets based on each sweep data can mistakenly detect a single target as multiple different targets, and processing large amounts of sweep data during a single scan leads to high processing load.
The radar device divides the rotation of the antenna into multiple sectors, generating sector data for each sector, and performs target detection using sector data, including labeling and centroid position calculation processes to suppress false detections and reduce processing load.
This approach reduces false detections of a single target appearing as multiple targets and decreases processing load by detecting targets sector by sector, allowing accurate target identification and reduced computational requirements.
Smart Images
Figure 0007867739000001 
Figure 0007867739000002 
Figure 0007867739000003
Abstract
Description
Technical Field
[0001] The present invention relates to a target detection device and a radar device that perform target detection based on an echo reflected by a radar wave.
Background Art
[0002] A radar device mounted on a ship or the like transmits a radar wave in all directions around the ship, receives an echo of the radar wave reflected by a target by an antenna, and detects the position (azimuth and distance) of the target based on the received echo. More specifically, while the antenna rotates within a horizontal plane, a pulse-shaped radio wave is transmitted as a radar wave, and a sweep process of receiving an echo reflected by the radar wave is repeatedly performed. Then, an analog echo signal is generated based on the echo received in each sweep process, and the generated echo signal is sequentially converted into digital sweep data, thereby obtaining sweep data for the entire circumferential direction (for one scan).
[0003] Some conventional radar devices perform target detection based on each sweep data. Also, some conventional radar devices perform target detection based on a large number of sweep data obtained during one scan. For example, in the radar device described in Patent Document 1, as shown in FIG. 10, sweep data extending in the horizontal direction are arranged vertically in the order of acquisition, and scan data indicating the distance R and azimuth θ from the antenna in a Cartesian coordinate system are generated. Each sweep data is divided into a plurality of cells based on the distance from the antenna, and cell data indicating the signal intensity of the echo signal is stored in each cell.Among such scan data, this radar device detects cells (indicated by broken lines in the figure) in which the signal intensity of the echo signal is equal to or greater than a predetermined value, and detects the set of the detected cells as one target.
Prior Art Documents
Patent Documents
[0004]
Patent Document 1
[0005] However, radar systems that detect targets based on each sweep data can sometimes mistakenly detect a single target as multiple different targets. When such false detections occur, a single target appears on the radar screen as multiple different targets and is tracked. Furthermore, detecting targets based on a large amount of sweep data obtained during a single scan presents a problem of extremely high processing load.
[0006] Therefore, the present invention aims to provide a target detection device and a radar device that can suppress false target detection and reduce processing load. [Means for solving the problem]
[0007] To solve the above problems, the invention described in claim 1 is a target detection device used in a radar device that transmits pulsed radar waves while rotating an antenna and repeatedly performs a sweep process to receive echoes reflected by the radar waves, and detects a target from sweep data generated based on the echoes, comprising: sector data generation means that divides the rotation of one scan of the antenna into a plurality of sectors and generates sector data in which a plurality of sweep data within the sector are arranged for each sector; and detection processing means that sequentially detects the target from the sector data for each sector. The detection processing means determines, based on the detection result of the previous sector, whether the target exists on the boundary between the previous sector, in which the target is detected first, and the next sector, which is adjacent to the previous sector and in which the target is detected next, and if it is determined that the target exists on the boundary, it detects the target on the boundary based on the sweep data near the boundary of the previous sector and the sector data of the next sector. It is characterized by the following:
[0009] Claim 2 The invention described in the claim 1In the target detection device described above, the sector data represents a Cartesian coordinate system in which the sweep data is divided into a plurality of cells based on the distance from the antenna, the sweep data is arranged in the order it is acquired, and the plurality of cells are arranged in a matrix. The detection processing means reads cell data indicating the signal strength of the echo from the cells, detects target cells in which the cell data is greater than or equal to a predetermined value, performs a labeling process which assigns the same identification label to adjacent target cells among the detected target cells, and performs a centroid position calculation process which calculates the centroid position of the target using the set of target cells to which the same identification label is assigned as the target. In the labeling process, If the target cell is detected from the final sweep data of the previous sector arranged at the boundary between the previous sector and the next sector, the final sweep data and the target cell that has been assigned the same identification label as the target cell in the final sweep data are temporarily stored as continuation processing data without being used in the centroid position calculation process in the previous sector. In the next sector, the labeling process is executed based on the final sweep data and the sector data of the next sector. In the next sector, the centroid position calculation process is executed based on the target cell that has been detected and assigned the same identification label as the target cell in the final sweep data, and the continuation processing data.
[0010] Claim 3 The invention described herein is a radar device that transmits pulsed radar waves while rotating an antenna and repeatedly performs a sweep process to receive echoes reflected from the radar waves, according to claims 1 to 2 It is characterized by comprising the target detection device described above. [Effects of the Invention]
[0011] Claim 1 and Claim 3According to the invention described above, the rotation of one antenna scan is divided into multiple sectors, and targets are detected for each sector. Compared to conventional radar systems that detect targets for each sweep data, this makes it possible to suppress false detections, such as detecting a single target as multiple different targets. Furthermore, since targets are detected for each sector data, the processing load can be reduced compared to detecting targets from scan data for one antenna rotation.
[0012] Claim 1 and claims 3 According to the invention described above, it is possible to determine whether a target exists on the boundary between adjacent previous and next sectors, and if it is determined that a target exists on the boundary, the target on the boundary is detected based on the sweep data near the boundary of the previous sector and the sector data of the next sector, thereby suppressing the detection of a single target existing on the boundary between sectors as multiple different targets.
[0013] Claim 2 and claims 2 Claims citing 3 According to the invention described above, when a target cell is detected from the final sweep data of the previous sector arranged at the boundary between the previous and next sectors, it is possible to determine that a target exists on the boundary between the previous and next sectors. Furthermore, if it is determined that a target exists on the boundary between the previous and next sectors, the previous sector does not perform centroid position calculation processing for targets on the boundary, thus preventing a single target on the boundary from being detected as a different target in the previous and next sectors. In addition, since inter-sector labeling processing is performed based on the sector data of the next sector and the final sweep data of the previous sector, centroid position calculation processing is performed using target cells with the same identification label in the next sector and the continuation processing data of the previous sector, it is possible to appropriately detect targets existing on the boundary and calculate their centroid position. [Brief explanation of the drawing]
[0014] [Figure 1] It is a block diagram showing a schematic configuration of a radar device according to an embodiment of the present invention. [Figure 2] It is an explanatory diagram of one sweep, sector, and one scan of radar waves. [Figure 3] It is a diagram showing a state where labeling processing and relabeling processing are performed on sector data. [Figure 4] It is a diagram showing a procedure for performing centroid position calculation processing on sector data. [Figure 5] It is a diagram showing a state where a target exists on the boundary between sectors. [Figure 6] It is a diagram showing sector data of the previous sector before the target exists on the boundary. [Figure 7] It is a diagram showing a state where the labels of the final sweep data of the previous sector are being rearranged. [Figure 8] It is a diagram showing a state where inter-sector labeling processing is being performed on the sector data of the next sector. [Figure 9] It is a flowchart showing a procedure for target detection. [Figure 10] It is an explanatory diagram showing a state where target detection is performed from sweep data for one scan of the conventional art.
Embodiments for Carrying Out the Invention
[0015] Hereinafter, the present invention will be described based on the illustrated embodiments.
[0016] FIG. 1 is a block diagram showing a schematic configuration of a radar device 1 according to an embodiment of the present invention. The radar device 1 is mounted on, for example, a ship, and includes an antenna 2, a transceiver 3, an A / D conversion unit 4, a signal processing unit 5, a sweep data storage unit 6, a display control unit 7, a display unit 8, and a target detection unit (target detection device) 9.
[0017] Antenna 2 rotates in the horizontal plane, transmitting pulsed radar waves and repeatedly performing a sweep process to receive echoes of the radar waves reflected by the target. The transmitting / receiving unit 3 generates radar waves to be transmitted from antenna 2 and inputs them to antenna 2. The transmitting / receiving unit 3 also detects and amplifies the echoes received by antenna 2 to generate an analog echo signal.
[0018] The A / D conversion unit 4 converts the echo signal output from the transmitting / receiving unit 3 using A / D conversion to generate digital sweep data. The signal processing unit 5 performs signal processing on the sweep data output from the A / D conversion unit 4, including removing unwanted echoes (clutter) caused by sea surface reflection, land reflection, and rain / snow reflection. The signal processing unit 5 also converts the sweep data, which is in polar coordinates indicated by azimuth θ and distance R, into Cartesian coordinates.
[0019] The sweep data storage unit 6 stores the sweep data for one sweep output from the signal processing unit 5 and outputs this sweep data for one sweep to the display control unit 7 and the target detection unit 9 until the next sweep process is executed and new sweep data is input. The target detection unit 9 detects a target based on the sweep data and outputs the detection result to the display control unit 7.
[0020] The display control unit 7 generates a raster-format radar image based on sweep data for one scan. The display control unit 7 also generates a target identification graphic image based on the target detection result of the target detection unit 9. The target identification graphic image is a graphic image used to display the identification name of the target detected by the target detection unit 9, and is displayed superimposed on the radar image. The display unit 8 is a display device equipped with the function of displaying the radar image and graphic image generated by the display control unit 7, and is configured, for example, as a liquid crystal display.
[0021] The target detection unit 9 comprises a sector data generation unit (sector data generation means) 91 and a detection processing unit (detection processing means) 92. The sector data generation unit 91 generates sector data based on sweep data for one sector output from the sweep data storage unit 6.
[0022] Here, as shown in Figure 2, if one transmission and reception from antenna 2 to the reception of an echo is defined as one sweep, and one rotation of antenna 2 is defined as one scan, then one scan contains a large number of sweeps, i.e., sweep data. The sector data generation unit 91 divides the rotation of antenna 2 for one scan into multiple sectors (for example, sector S, sector S+1, sector S-1, etc.), and generates sector data for each sector in which multiple sweep data within the sector are arranged.
[0023] In Figure 3(A), the symbol 91A indicates an example of sector data generated by the sector data generation unit 91. The sector data 91A consists of multiple sweep data Swp units (one sector) arranged along the vertical axis, with each sweep data Swp unit extending in the horizontal direction. Each sweep data Swp unit is divided into multiple cells C based on its distance from the antenna 2, and each cell C stores cell data indicating the signal strength of the echo signal. As a result, the sector data 91A consists of multiple cells C arranged in a matrix, with the distance R and azimuth θ indicated in a Cartesian coordinate system. The horizontal axis of the sector data 91A is set to a data counter value indicating the position of each cell C in the distance direction, and the vertical axis of the sector data 91A is set to a sweep counter value indicating the position (azimuth) of the sweep data Swp unit.
[0024] The detection processing unit 92 sequentially performs the process of detecting targets from sector data sector by sector. As a process for detecting targets, the detection processing unit 92 performs a labeling process and a center of gravity position calculation process. The labeling process is the process of assigning an identification label to each cell of the sector data according to a predetermined procedure. The center of gravity position calculation process identifies a set of cells that have been assigned the same identification label by the labeling process as a target, and calculates the area and center of gravity position for each target. The area of the target is used to determine whether or not the target is another ship or the like. The center of gravity position of the target is used for tracking the detected target, etc.
[0025] In the labeling process of the detection processing unit 92, cell data is read sequentially from the cell closest to the antenna 2 (data count counter value "1") for each sweep data set. Cells with cell data equal to or greater than a predetermined value are detected as target cells and assigned an identification label. The identification label is, for example, a sequential number starting from "1". If multiple target cells are detected within the same sweep data, the same identification label is assigned to the target cells if they are adjacent to each other, and different identification numbers are assigned to the target cells if they are far apart.
[0026] Furthermore, if adjacent target cells are detected between adjacent sweep data, the later-detected target cell will be assigned the same identification label as the earlier-detected target cell. Note that if a target label is adjacent to an earlier-detected target cell both within the sweep data (horizontally) and between sweep data (vertically), the identification label of the target cell adjacent between sweep data (vertically) will take precedence.
[0027] Furthermore, if, after the labeling process described above, a set of adjacent target cells is assigned different identification labels, a relabeling process is performed. In this relabeling process, the identification labels of the target cells in that set are unified to the identification label that is most frequently assigned within that set. As a result, target cells that represent the same target are assigned the same identification label. In addition, the labeling and relabeling processes include a LUT (not shown) that stores used identification labels to prevent the duplication of identification labels.
[0028] For example, in the sector data 91A shown in Figure 3(A), according to the above rules, cells C with coordinates "8,5", "4,6", "8,6", "9,5", and "8,7" indicated by the data count counter value and sweep counter value are detected as target cells Tc, and each is assigned an identification label of "1" or "2". Furthermore, since these sets of target cells Tc are assigned different identification labels, a relabeling process is performed, and as shown in Figure 3(B), the identification labels of target cells Tc within the same set are unified to "1". As a result, the set of target cells Tc assigned the identification label "1" is detected as target T.
[0029] Next, the center of gravity calculation process will be explained. Figure 4(A) shows the procedure for calculating the center of gravity for target T, which has been relabeled in Figure 3(B), and Figure 4(B) shows the actual calculation content in each calculation step. From left to right, Figures 4(A) and (B) show "labeled data," "calculation of moment, area, and volume," and "calculation of the center of gravity." In Figures 4(A) and (B), each target cell Tc of target T is assigned an identification code a, b, c, d, and e.
[0030] In the "Labeled Data" section, for each target cell Tc of target T, the label value of the identification label (Label), the data counter value (D_cnt), the sweep counter value (L_cnt), and the data value of the cell data (Data) are obtained.
[0031] In the following "Calculation of Moment, Area, and Volume," for each target cell Tc, the moment in the azimuthal direction (M_TH) and the moment in the distance direction (M_R) are calculated based on the following formulas (1) to (3), and the volume (V) is set. The data value (Data) of the cell data is used for the volume (V). Note that in this explanation of the centroid position calculation process, the moment in the azimuthal direction (M_TH) and the moment in the distance direction (M_R) are calculated after the labeling process, but in reality, the moment in the azimuthal direction (M_TH) and the moment in the distance direction (M_R) are calculated when the target cell is detected during the labeling process. M_TH=L_cnt×Data (1) M_R=D_cnt×Data (2) V=Data (3)
[0032] In the "Calculation of Moment, Area, and Volume," the azimuthal moment (M_TH), distance moment (M_R), volume (V), and area (S) are calculated and set for each target cell Tc, and these are added together to calculate the total azimuthal moment (TM_TH), total distance moment (TM_R), total volume (TV), and total area (TS) for the entire target T. Note that the total area (TS) is the total number of cells in the target T.
[0033] In the following "Centroid Calculation," the azimuthal component centroid position (G_TH) and the distance component centroid position (G_R) of the target T are calculated based on the following formulas (4) and (5). In formulas (4) and (5) below, "R1Width" represents the distance between each cell, and "TH1Width" represents the angle of one sweep interval. The azimuthal component centroid position (G_TH) and distance component centroid position (G_R) of the target T calculated in the centroid position calculation process are output from the target detection unit 9 to the display control unit 7. G_TH=(TM_TH / TV)×TH1Width (4) G_R=(TM_R / TV)×R1Width (5)
[0034] As explained above, by performing labeling and centroid position calculation processes for target detection in each sector, it is possible to suppress false detections, such as detecting a single target as multiple different targets, compared to when target detection is performed for each sweep data. Furthermore, since targets are detected sector by sector, the processing load can be reduced compared to when targets are detected from scan data equivalent to one rotation of the antenna.
[0035] Furthermore, even if targets are detected in each sector, if a target exists on the boundary between adjacent sectors, there is a possibility of misidentifying the same target as multiple different targets. Figure 5 shows a situation where, among multiple sectors, two targets exist on the boundary between the previous sector, where target detection is performed first, and the next sector, which is adjacent to the previous sector and where target detection is performed afterward. The two targets enclosed by the dashed lines in the figure are detected in the previous and next sectors respectively, and there is a possibility of misidentification as four different targets.
[0036] In this embodiment, the target detection unit 9 determines whether a target exists on the boundary between adjacent sectors based on the detection result of the previous sector, in order to appropriately detect targets that exist on the boundary between adjacent sectors as described above. If it determines that a target exists on the boundary, it performs inter-sector labeling processing to detect the target on the boundary based on the sweep data near the boundary of the previous sector and the sector data of the next sector. The inter-sector labeling processing will be described below.
[0037] Figure 6 shows the sector data 91B of the previous sector after the labeling and relabeling processes have been completed by the detection processing unit 92, and the LUT in which the used identification labels are stored. This sector data 91B shows that after the labeling process assigned identification labels "1" to "10" to the target cells, the relabeling process sorted the identification labels to "1", "2", "7", and "9". Similarly, in the LUT, identification labels "1" to "10" were stored during the labeling process, but after the relabeling process, identification labels "3" to "6", "8", and "10" were deleted, leaving only the identification labels "1", "2", "7", and "9".
[0038] The detection processing unit 92 executes inter-sector labeling processing if, in the label data 91B of the previous sector as described above, the target cell has been detected from the final sweep data Swp-E, in which the last target cell was detected. In this inter-sector labeling processing, the final sweep data Swp-E and target cells that have the same identification label as the target cells in the final sweep data Swp-E (for example, "7" and "9") (target cells enclosed by a dashed line in Figure 6) are not used in the centroid position calculation processing in the previous sector, but are stored as continuation processing data in a temporary storage memory (not shown). The temporary storage memory also stores the azimuthal moment (M_TH), distance moment (M_R), and cell data as related data for the target cells that have been set as continuation processing data. Furthermore, among the identification labels stored in the LUT, the identification labels "7" and "9" that have been assigned to the target cells in the final sweep data Swp-E are set as continuation labels.
[0039] In other words, when performing inter-sector labeling, the centroid position calculation process is performed only for the set of identification labels "1" and "2" in the previous sector's sector data 91B, which are not related to the target cell of the final sweep data Swp-E, and the centroid position calculation process is not performed for the set of identification labels "7" and "9".
[0040] The detection processing unit 92 uses the final sweep data Swp-E of the previous sector for labeling the next sector, and as shown in Figure 7, it reassigns the identification labels "7" and "9" of the target cells sequentially starting from the identification label starting number "1". In addition, the identification labels of the target cells stored as continuation processing data in the temporary memory are also changed to "1" and "2", just like the target cells in the final sweep data Swp-E. Furthermore, the identification labels "7" and "9" set as continuation labels in the LUT are changed to "1" and "2", just like the target cells in the final sweep data Swp-E.
[0041] As shown in Figure 8, the detection processing unit 92 places the final sweep data Swp-E of the previous sector, to which the identification label has been reassigned, before the start sweep data Swp-S of the sector data 91C of the next sector, performs labeling processing, and then performs relabeling processing. Note that the sector data 91C of the next sector shown in Figure 8 has a reduced number of display columns to avoid complexity in the diagram.
[0042] After the relabeling process for the next sector is completed, the detection processing unit 92 performs a centroid position calculation process based on the target cell detected in the next sector that has the same identification label as the target cell of the final sweep data Swp-E, and the target cell and its related data stored as continuous processing data. This makes it possible to appropriately detect targets located on the boundary between sectors.
[0043] Furthermore, if there are targets surrounding the radar device 1, the target cell will always be detected from the final sweep data of each sector, and therefore the inter-sector labeling process will not terminate. For this reason, in the labeling process, the number of processed sweep data can be counted based on the sweep counter value, and the inter-sector labeling process may be terminated when the count value exceeds a predetermined value. Preferably, the maximum number of sweeps counted to terminate the inter-sector labeling process is the number of sweeps for one scan.
[0044] Next, the operation of the target detection process in the above embodiment will be explained based on the flowchart shown in Figure 9. The sweep data storage unit 6 outputs sweep data generated based on the transmission and reception of a radar wave by the antenna 2 to the target detection unit 9. The sector data generation unit 91 of the target detection unit 9 generates sector data based on the sweep data for one sector output from the sweep data storage unit 6 (step S1).
[0045] The detection processing unit 92 of the target detection unit 9 sequentially performs the process of detecting targets from sector data for each sector. Specifically, the detection processing unit 92 performs a labeling process (step S2) to detect target cells from each cell of the sector data and assign identification labels, and a relabeling process (step S3) to organize the assigned identification labels according to a predetermined procedure.
[0046] If, after the relabeling process is complete, the detection processing unit 92 does not detect a target cell from the final sweep data of the previous sector (NO in step S4), it performs a centroid position calculation process (step S5) on the sector data for which the relabeling process has been completed in order to calculate the area and centroid position of the target. The area and centroid position of the target calculated by the centroid position calculation process are output to the display control unit 7 and used for displaying the radar image.
[0047] Furthermore, after the relabeling process is completed, if a target cell has been detected from the final sweep data of the previous sector (YES in step S4), the detection processing unit 92 performs inter-sector labeling (step S6) on the sector data of the next sector in order to detect a target on the boundary based on the sweep data near the boundary of the previous sector and the sector data of the next sector. After the inter-sector labeling process is completed, the detection processing unit 92 performs centroid position calculation (step S5) and outputs the calculated target area and centroid position to the display control unit 7.
[0048] As described above, according to the radar device 1 and target detection unit 9 of this embodiment, the rotation of one scan of the antenna 2 is divided into multiple sectors, and a target is detected for each sector. Therefore, compared to conventional radar devices that detect a target for each sweep data, it is possible to suppress false detections that would otherwise occur, such as detecting a single target as multiple different targets. Furthermore, since a target is detected for each sector data, the processing load can be reduced compared to detecting a target from scan data for one rotation of the antenna.
[0049] Furthermore, according to this embodiment, the radar device 1 and target detection unit 9 determine whether a target exists on the boundary between adjacent previous and next sectors. If it is determined that a target exists on the boundary, the target on the boundary is detected based on sweep data near the boundary of the previous sector and sector data of the next sector. This makes it possible to suppress the detection of a single target existing on the boundary between sectors as multiple different targets.
[0050] Furthermore, according to this embodiment, with the radar device 1 and target detection unit 9, if a target cell Tc is detected from the final sweep data Swp-E of the previous sector, which is arranged at the boundary between the previous and next sectors, it is possible to determine that a target exists on the boundary between the previous and next sectors. Also, if it is determined that a target exists on the boundary between the previous and next sectors, the previous sector does not perform centroid position calculation processing for targets on the boundary, thus preventing a single target on the boundary from being detected as different targets in the previous and next sectors. Moreover, in the next sector, since inter-sector labeling processing is performed based on the final sweep data Swp-E of the previous sector and the sector data of the next sector before centroid position calculation processing is performed, it is possible to appropriately detect targets existing on the boundary and calculate their centroid positions.
[0051] Although embodiments of this invention have been described above, the specific configuration is not limited to the embodiments described above, and any design changes, etc., that do not depart from the gist of this invention are also included. For example, although the above embodiments were described using a radar system for ships as an example, the invention is also applicable to other radar systems in which radar waves are scanned.
[0052] 1. Radar equipment 2 antennas 3. Transmitter / Receiver Unit 4. A / D conversion section 5. Signal Processing Unit 6. Sweep Data Storage Unit 7 Display Control Unit 8 Display 9. Target detection unit (target detection device) 91 Sector data generation unit (sector data generation means) 92 Detection Processing Unit (Detection Processing Means)
Claims
1. A target detection device used in a radar system that transmits pulsed radar waves while rotating an antenna and repeatedly performs a sweep process to receive echoes reflected from the radar waves, and detects a target from sweep data generated based on the echoes, Sector data generation means divides the rotation of the antenna for one scan into multiple sectors, and generates sector data in which multiple sweep data within the sector are arranged for each sector, A detection processing means for sequentially detecting the target from the sector data for each sector, Equipped with, The detection processing means is Based on the detection result of the previous sector, it is determined whether the target exists on the boundary between the previous sector, which is adjacent to the previous sector, and the next sector, which is adjacent to the previous sector and is the next sector to which the target is detected. If it is determined that the target exists on the boundary, the target on the boundary is detected based on the sweep data near the boundary of the previous sector and the sector data of the next sector. A target detection device characterized by the following features.
2. The sector data represents a Cartesian coordinate system in which the sweep data is divided into multiple cells based on the distance from the antenna, and the sweep data is arranged in the order it is acquired, with the multiple cells arranged in a matrix. The detection processing means is A labeling process is performed which involves reading cell data indicating the signal intensity of the echo from the cell, detecting target cells for which the cell data is equal to or greater than a predetermined value, and assigning the same identification label to adjacent target cells among the detected target cells. The set of target cells to which the same identification label is assigned is defined as the target, and a centroid position calculation process is performed to calculate the centroid position of the target. In the labeling process, if the target cell is detected from the final sweep data of the previous sector arranged at the boundary between the previous sector and the next sector, The final sweep data and the target cell that has been assigned the same identification label as the target cell in the final sweep data are temporarily stored as continuation processing data and are not used in the centroid position calculation process in the previous sector. In the next sector, the labeling process is performed based on the final sweep data and the sector data of the next sector. Based on the target cell detected in the next sector, which has the same identification label as the target cell in the final sweep data, and the continuation processing data, the centroid position calculation process is executed. The target detection device according to claim 1, characterized in that it is a target detection device.
3. A radar device that transmits pulsed radar waves while rotating an antenna and repeatedly performs a sweep process to receive echoes reflected from the radar waves, A target detection device according to claim 1 or 2, A radar device characterized by the following features.