Method for managing an agricultural vehicle during harvesting process in a plantation of fruit trees, such as an orange orchard

1 . A method for managing an agricultural vehicle during a harvesting process in a plantation of fruit trees, comprising:

a process of recognizing trunks of the fruit trees of the plantation to be worked by means of a first three-dimensional (3D) sensor system interfaced with a processor of the agricultural vehicle, wherein the agricultural vehicle is shaped to have a portal configured to move over a crop row as the agricultural vehicle moves over the crop row and the agricultural vehicle comprises a couple of rotors arranged to work simultaneously on opposite sides of each plant of the crop row, and wherein the first 3D sensor system comprises a couple of 3D sensors arranged in a lower portion of the agricultural vehicle on opposite sides of the agricultural vehicle and oriented to converge on a common point approximately on a vehicle center line axis in front of the agricultural vehicle, the process comprising the following steps of:

simultaneously acquiring two point clouds obtained from the couple of 3D sensors;

merging of the two point clouds to obtain a cumulative point cloud;

processing of the cumulative point cloud to obtain a first two-dimensional map as a horizontal slice approximately parallel with a ground; and

performing a first fitting of pseudo-ellipsoids in the first two-dimensional map in order to identify the trunks of plants being worked in the crop row.

2 . The method according to claim 1 , wherein the method further comprises:

tracing a route based on a position of the first fitted pseudo-ellipsoids by performing a linear regression operation; and

controlling steering of the agricultural vehicle according to the route traced.

3 . The method according to claim 1 , wherein the couple of rotors are substantially vertical arranged in such a way to harvest the crop row on the opposite sides of each plant of the crop row while the agricultural vehicle moves over the crop row itself, wherein the rotors are in a direction transverse to the vehicle center line axis, and wherein the method further comprises laterally adjusting a position of the couple of rotors according to a position of a detected trunk closest to the front of the agricultural vehicle so that the detected trunk is in an intermediate position between the couple of rotors.

4 . The method according to claim 3 , further comprising an additional process for identifying a trajectory that overlaps a crop row centerline of the plantation to be worked via a second 3D sensor system arranged in a top position of the agricultural vehicle and interfaced with the processor, the additional process comprising the following steps of:

acquiring a 3D point cloud from the second 3D sensor system corresponding to reflections picked up by the 3D sensor system in the front of the agricultural vehicle;

processing of the 3D point cloud to obtain a second two-dimensional map as a transversal slice of the 3D point cloud according to a transversal cutting plane of the 3D point cloud that is transverse to the vehicle centerline at a first predetermined distance from the front of the agricultural vehicle;

performing a second fitting to obtain a pseudo-ellipsoid interpolating curve on the second two-dimensional map;

identification of a peak (PK 1 ) associated to the pseudo-ellipsoid interpolating curve and calculating a lateral misalignment (OTE) between such peak (PK 1 ) and a vertical line of the second two-dimensional map corresponding to the vehicle centerline (VC); and

controlling a trajectory of the agricultural vehicle in order to minimize the first predetermined distance.

5 . The method according to claim 4 , further comprising performing a first windowing of the second two dimensional map immediately before performing the second fitting in order to cut lateral portions of the second two dimensional map.

6 . The method according to claim 1 , wherein execution of the steps for the process is recursive according to a vehicle forward motion of the agricultural vehicle.

7 . The method according to claim 1 , further comprising:

preliminarily acquiring a GPS crop row position; and

controlling a steering angle of the agricultural vehicle in order to maintain a motion of the agricultural vehicle aligned with the crop row.

8 . A non-transitory computer-readable media, comprising:

a memory encoding processor-executable routines;

a processing system configured to access the memory and to execute the processor-executable routines, wherein the processor-executable routines, when executed by the processing system, cause the processing system to:

perform a process of recognizing trunks of fruit trees of a plantation to be worked by means of a first three-dimensional (3D) sensor system interfaced with a processor of an agricultural vehicle, wherein the agricultural vehicle is shaped to have a portal configured to move over a crop row as the agricultural vehicle moves over the crop row and the agricultural vehicle comprises a couple of rotors arranged to work simultaneously on opposite sides of each plant of the crop row, and wherein the first 3D sensor system comprises a couple of 3D sensors arranged in a lower portion of the agricultural vehicle on opposite sides of the agricultural vehicle and oriented to converge on a common point approximately on a vehicle center line axis in front of the agricultural vehicle, the process comprising the following steps of:

simultaneously acquiring two point clouds obtained from the couple of 3D sensors;

merging of the two point clouds to obtain a cumulative point cloud;

processing of the cumulative point cloud to obtain a first two-dimensional map as a horizontal slice approximately parallel with a ground; and

performing a first fitting of pseudo-ellipsoids in the first two-dimensional map in order to identify the trunks of plants being worked in the crop row.

9 . An agricultural vehicle, comprising:

a propulsion system;

a braking and steering system, wherein the agricultural vehicle is shaped to have a portal configured to move over a crop row as the agricultural vehicle moves over the crop row;

a couple of rotors arranged to work simultaneously on opposite sides of each plant of the crop row;

a processor; and

a first three-dimensional (3D) sensor system interfaced with the processor, wherein the first 3D sensor system comprises a couple of 3D sensors arranged in a lower portion of the agricultural vehicle on opposite sides of the agricultural vehicle and oriented to converge on a common point approximately on a vehicle center line axis in front of the agricultural vehicle,

wherein the processor is configured to:

simultaneously acquire two point clouds obtained from the couple of 3D sensors;

merge the two point clouds to obtain a merged point cloud;

process the merged point cloud to obtain a first two-dimensional map as a horizontal slice approximately parallel with a ground; and

perform a first fitting of pseudo-ellipsoids in the first two-dimensional map in order to identify trunks of plants being worked in the crop row.

10 . The agricultural vehicle according to claim 9 , wherein the processor is further configured to:

trace a route based on a position of the first fitted pseudo-ellipsoids by performing a linear regression operation; and

control steering of the agricultural vehicle according to the route traced.