CN-116609775-B - Real-time detection method and device for radar detection target and spatial position relation

CN116609775BCN 116609775 BCN116609775 BCN 116609775BCN-116609775-B

Abstract

The invention discloses a real-time detection method and a detection device for the relation between a radar detection target and a space position, wherein the detection method comprises the steps of dividing the detection plane of the radar into a space collection set in a grid form, wherein each grid comprises the position coordinates of a current grid and the position coordinates of adjacent grids, updating signals obtained in real time by the radar into the corresponding grids according to the coordinate values, wherein the information of the grids comprises the number of signal track points in the grids, the record time stamp of the signal track points in the grids and the grid activity value, superposing the space collection set and the division time frame collection to form three-dimensional data, extracting a characteristic signal set, and converting the characteristic signal set into a continuous coordinate curve.

Inventors

ZHANG GUOKUN
WANG JUNFENG
LV DAYONG
SUN WENBO

Assignees

中科蓝卓(北京)信息科技有限公司

Dates

Publication Date: 20260512
Application Date: 20230420

Claims (6)

1. A real-time detection method for the relation between a radar detection target and a space position is characterized by comprising the following steps: dividing a detection plane of the radar into a space set in a grid form, wherein each grid comprises the position coordinates of the current grid and the position coordinates of adjacent grids; updating signals obtained by the radar in real time into corresponding grids according to coordinate values, wherein the information of the grids comprises the number of signal track points in the grids, a timestamp recorded by the signal track points in the grids and a grid activity value; Superposing the space set and the divided time frame set to form three-dimensional data; Extracting a characteristic signal set, and converting the characteristic signal set into a continuous coordinate curve; obtaining the continuous coordinate curve includes: traversing current grid adjacent grid data, and associating active grids to form continuous signal track lines; Projecting the coordinate point set of the correlation signal in each track line to the time axis and the transverse coordinate plane, screening out the space and time continuous coordinate points under the condition of continuous time projection lines, and fitting signals to form the continuous coordinate curve; the method for determining the active grid comprises the following steps: calculating the activity value of the grid according to a formula, and determining the grid as an active grid when the activity value of the grid is greater than an activity threshold; The formula is as follows: v=p*n1–(t0-t1)*n2+s*n3 Wherein v is the active value of the current grid, p is the number of signals in the current grid, n1 is the weight of the number of signals, t0 is the current time stamp, t1 is the time stamp of the latest signal of the current grid, n2 is the time weight, s is the number of adjacent active grids, and n3 is the weight of the number of adjacent grids.
2. The method for real-time detection of a spatial position relationship between a radar detection target according to claim 1, wherein the method for forming the spatial set comprises: Dividing a detection plane of the radar into grids, traversing the grids to obtain the position coordinates of the current grid, calculating the position coordinates of adjacent grids of the current grid, and storing the position coordinates of adjacent grids of the current grid into the current grid.
3. The method for real-time detection of the relationship between a radar detection target and a spatial location according to claim 1, wherein the detection plane of the radar is divided into a predetermined number of grids according to a predetermined resolution.
4. The method of claim 1, wherein forming a continuous signal trace further comprises screening out signals that match movement characteristics as signals that form the trace based on movement characteristics of the detected target.
5. The method of claim 4, wherein the moving features include a moving speed and a moving direction.
6. A real-time detection device for radar detection of a relationship between a target and a spatial location, comprising: the grid module is used for dividing the detection plane of the radar into a space set in a grid form, and each grid comprises the position coordinates of the current grid and the position coordinates of the adjacent grid; The updating module is used for updating signals obtained by the radar in real time into corresponding grids according to coordinate values, wherein the information of the grids comprises the number of signal track points in the grids, a timestamp recorded by the signal track points in the grids and a grid activity value; The curve module is used for superposing the space set and the segmentation time frame set to form three-dimensional data, extracting a characteristic signal set and converting the characteristic signal set into a continuous coordinate curve; The curve module comprises a track line unit and a continuous coordinate curve unit; The trace line unit is used for traversing the adjacent grid data of the current grid, and correlating the active grids to form a continuous signal trace line; the continuous coordinate curve unit is used for forming a three-dimensional coordinate system by the continuous track horizontal coordinate system and a time axis, projecting a coordinate point set of related signals in each track line to the time axis and a transverse coordinate plane, screening out space and time continuous coordinate points under the condition of continuous time projection lines, and fitting signals to form the continuous coordinate curve; the method for determining the active grid comprises the following steps: calculating the activity value of the grid according to a formula, and determining the grid as an active grid when the activity value of the grid is greater than an activity threshold; The formula is as follows: v=p*n1–(t0-t1)*n2+s*n3 Wherein v is the active value of the current grid, p is the number of signals in the current grid, n1 is the weight of the number of signals, t0 is the current time stamp, t1 is the time stamp of the latest signal of the current grid, n2 is the time weight, s is the number of adjacent active grids, and n3 is the weight of the number of adjacent grids.

Description

Real-time detection method and device for radar detection target and spatial position relation Technical Field The invention relates to the field of monitoring, in particular to a real-time detection method and device for the relation between a radar detection target and a spatial position. Background With the gradual maturity of civil millimeter wave radar technology, the method is widely applied to the fields of automatic driving, perimeter security and the like. The millimeter wave Lei Dayin has the characteristics of low cost, all weather and high privacy in the field of perimeter security, and has unique advantages in personnel tracking and identification. When the existing millimeter wave radar scans a target, a trace is formed on the tracked target, and the trace refers to a target coordinate signal which is output by the radar through the radar internal data processing in the scanning process. The point trace is the most basic data unit for realizing target tracking and target track fitting. Detection and tracking of pedestrian targets are always research hotspots and difficulties in the technical field of radars. The pedestrian target RCS is small, the moving speed is low, the target echo is small and weak, the Doppler effect is not obvious, the effective detection of the target is difficult, the surrounding environment of the pedestrian target during moving is complex, and the surrounding environment is easily covered by ground, building clutter and tree disturbance signals. In summary, the pedestrian target has the problems of weak echo, strong clutter influence, and Doppler frequency approaching clutter, so that the interference point trace output by the radar on the tracked signal is excessive, the point trace is discontinuous, and the effective target point trace is difficult to distinguish, so that the actual moving track of the target is calculated. In the present personnel tracking and recognition based on millimeter wave radar, the following problems exist that heuristic data association algorithm, such as Nearest Neighbor (NN) algorithm, is adopted, the principle is simple, but association error rate is high under the scene of high target density and false alarm rate, state estimation cannot be accurately carried out under non-Gaussian and non-linear conditions by adopting Kalman or extended Kalman filtering algorithm, and the target cannot be accurately and robustly recognized by adopting simple machine learning algorithm, such as k Nearest Neighbor (kNN) algorithm or logistic regression (Logistic Regression, LR) algorithm in pedestrian recognition. Therefore, a real-time detection method and a detection device for the relationship between a radar detection target and a spatial position are desired. Disclosure of Invention The invention aims to provide a real-time detection method and a detection device for the spatial position relation between a radar detection target and an area, which can solve the problems that the track coordinates of a security millimeter wave radar are discontinuous, radar clutter signals interfere with discrete signals, the calculation time complexity of the position relation between the coordinates and a complex polygonal defense area is high, and the multi-target motion conversion judgment is complex. Based on the above object, the present invention provides a real-time detection method for a relationship between a radar detection target and a spatial position, comprising: Dividing a detection plane of the radar into a space collection in a grid form, wherein each grid comprises the position coordinates of the current grid and the position coordinates of adjacent grids; updating signals obtained by the radar in real time into corresponding grids according to coordinate values, wherein the information of the grids comprises the number of signal track points in the grids, a timestamp recorded by the signal track points in the grids and a grid activity value; Superposing the space set and the divided time frame set to form three-dimensional data; and extracting a characteristic signal set, and converting the characteristic signal set into a continuous coordinate curve. In an alternative, obtaining the continuous coordinate curve includes: traversing current grid adjacent grid data, and associating active grids to form continuous signal track lines; and projecting the coordinate point set of the correlation signal in each track line to the time axis and the transverse coordinate plane, screening out the space and time continuous coordinate points under the condition of continuous time projection lines, and fitting signals to form the continuous coordinate curve. In an alternative, the method for forming the spatial collection includes: Dividing a detection plane of the radar into grids, traversing the grids to obtain the position coordinates of the current grid, calculating the position coordinates of adjacent grids of the current grid, and storing the position coo