Laser sensing-based method for spatial positioning of agricultural robot
A laser perception-based method for spatial positioning of an agricultural robot: erecting a laser radar with a ranging function in a positioning space, setting a three-dimensional coordinate system, and conducting scanning using the laser radar to obtain point cloud data of an object in the positioning space, where the point cloud data include an azimuth and a distance with respect to the laser radar; installing a laser receiver on the agricultural robot, receiving a laser radar signal using the laser receiver during movement, when a laser beam emitted by the laser radar irradiates the laser receiver, outputting laser signal data and elevation data from the laser receiver; conducting time-event matching on the laser signal data obtained by the laser receiver and the point cloud data scanned by the laser radar within each scanning period of the laser radar to obtain three-dimensional coordinates of a central position of the laser receiver.
1 . A laser sensing-based method for spatial positioning of an agricultural robot, comprising:
Step 1: erecting a laser radar with a ranging function in a positioning space, setting a three-dimensional coordinate system, and conducting scanning using the laser radar to obtain point cloud data of an object in the positioning space, wherein the point cloud data comprise an azimuth and a distance with respect to the laser radar;
Step 2: installing a laser receiver on the agricultural robot, receiving a laser radar signal using the laser receiver during movement, and when a laser beam emitted by the laser radar irradiates the laser receiver, outputting laser signal data and elevation data h from the laser receiver;
wherein the laser receiver comprises a filter housing configured to eliminate ambient light and two photosensitive arrays configured to sense a laser signal, and output the laser signal as sensing laser status data after signal processing; the photosensitive arrays are divided into several segments in a vertical direction, different segments of the photosensitive arrays sense the laser signal and express the laser signal as an elevation data h′ of the laser receiver, each frame of the laser signal data comprises a piece of the sensing laser status data and the elevation data h′ of the laser receiver, the laser receiver is kept within a range of a laser scanning plane by an electric push rod, the laser signal data and the point cloud data output by the laser radar both have a time label, and a set of the point cloud data output by the laser radar and a set of the laser signal data output by the laser receiver are obtained by conducting scanning using the laser radar for one circle; and
Step 3: conducting time-event matching on the laser signal data obtained by the laser receiver and the point cloud data scanned by the laser radar within each scanning period of the laser radar to obtain three-dimensional coordinates of a central position of the laser receiver.
2 . The laser sensing-based method for spatial positioning of an agricultural robot according to claim 1 , wherein in Step 1, each piece of point data of the point cloud data comprises a distance from an obstacle to the laser radar, an angle with respect to an initial scanning line of the laser radar, and a time label.
3 . The laser sensing-based method for spatial positioning of an agricultural robot according to claim 1 , wherein in Step 1, the laser radar is erected to ensure that the laser scanning plane is horizontal, an installation height is known, and the laser receiver is installed within the range of a received laser scanning plane, and an initial height h 1 of the laser receiver is obtained according to the elevation data h fed back by the laser receiver after laser light irradiates the laser receiver.
4 . The laser sensing-based method for spatial positioning of an agricultural robot according to claim 1 , wherein said conducting time-event matching in Step 3 specifically comprises realizing positioning of the agricultural robot by representing a point in the point cloud data corresponding to a same time at which the laser radar signal is obtained and received by the laser receiver as a position of the laser receiver fixed on the agricultural robot in the three-dimensional coordinate system with the laser radar as a coordinate origin.
5 . The laser sensing-based method for spatial positioning of an agricultural robot according to claim 1 , wherein in Step 1, the three-dimensional coordinate system is constructed as follows:
setting a central position O(X, Y, H) of the laser radar as an origin of the three-dimensional coordinate system, setting L 1 as a first laser ray emitted by the laser radar, setting L n as a last laser ray emitted by the laser radar, setting L R1 as a first ray emitted by the laser radar and irradiating the laser receiver, setting L Rn as a last ray emitted by the laser radar and irradiating the laser receiver, and denoting points of intersection of the rays L R1 and L Rn with the laser radar as R 1 (x 1 , y 1 , h) and R n (x n , y n , h), respectively;
solving three-dimensional point cloud coordinates from an included angle ρ between a corresponding laser ray and the first laser ray emitted by the laser radar, and a measured relative distance d by formulas x=d·cos ρ and y=d·sin ρ; and
when the laser radar conducts scanning for one circle, obtaining coordinates of points R 1 (x 1 , y 1 , h), R n (x n , y n , h) by time-event matching between the time label of receipt of the laser radar signal in the laser signal data and the point cloud data, so as to obtain three-dimensional coordinates
(
x
1
+
x
n
2
,
y
1
+
y
n
2
,
h
)
of the central position of the laser receiver within this scanning period.
6 . The laser sensing-based method for spatial positioning of an agricultural robot according to claim 5 , wherein the elevation data h are obtained specifically by comparing an initial height h 1 of the laser receiver with the elevation data h′ of the laser receiver obtained in a measurement process, wherein when a height of the laser receiver is smaller than the initial height h 1 of the laser receiver, h′ is negative, when the height of the laser receiver is greater than the initial height h 1 of the laser receiver, h′ is positive, and calculation formulas are as follows h=H+(h 1 +h′).
7 . The laser sensing-based method for spatial positioning of an agricultural robot according to claim 1 , wherein said keeping the laser receiver within a range of the laser scanning plane by an electric push rod specifically comprises: if the laser receiver is out of the range of the laser scanning plane, and if the last elevation data h′ of the laser receiver are negative, pushing up the laser receiver by the electric push rod until the laser receiver outputs the laser signal data, and feeding back displacement D via the electric push rod, wherein D is positive at this time; if the last elevation data h′ of the laser receiver are positive, pushing down the laser receiver through the electric push rod until the laser receiver outputs the laser signal data, and feeding back displacement D via the electric push rod, wherein D is negative at this time; and changing an initial height h 1 of the laser receiver to h 1 +D.
8 . The laser sensing-based method for spatial positioning of an agricultural robot according to claim 1 , wherein the method specifically comprises conducting operation processing within a scope of use, establishing the three-dimensional coordinate system with the laser radar as an origin, when the mobile agricultural robot is within a working range of the laser radar, and if two rows of photosensitive films of the laser receiver are both sensitive to light, and an included angle between connection lines and longitudinal lines of sensing units of the two rows of photosensitive films is not a vertical angle, removing an error caused by the included angle from attitude; if there is only one row of photosensitive films sensitive to light in the two rows of photosensitive films, correcting, by point cloud computing, a positioning point at this moment according to two points nearby the positioning point, so as to obtain a more precise moving trajectory; if neither of the two rows of photosensitive films are sensitive to light, predicting a positioning point at this moment in combination with a prediction method in point cloud computing to obtain predicted positioning data; and
after obtaining the predicted positioning data, integrating the predicted positioning data into a set global coordinate system, determining to-be-removed outliers by matching with observations, and then fusing predicted three-dimensional coordinates of the central position of the laser receiver and observed three-dimensional coordinates of the central position of the laser receiver to obtain more precise position information.