rostest
commented
1 year ago Thank you for reporting this. Performance depends on the execution time of the callbacks. The Python function pySickScanCartesianPointCloudMsgToXYZ shows an example of how a point cloud message can be converted and visualized with matplotlib. For real-time processing, we recommend using C++ and/or ROS (ROS-1 or ROS-2 on Linux or Windows).
Please test the performance with a simple callback, that just prints a message:
# Simple callback for cartesian point cloud messages, just prints a message
def pySickScanCartesianPointCloudMsgCallback(api_handle, pointcloud_msg):
pointcloud_msg = pointcloud_msg.contents # dereference msg pointer (pointcloud_msg = pointcloud_msg[0])
print("pySickScanCartesianPointCloudMsgCallback: api_handle={}, {}x{} pointcloud".format(api_handle, pointcloud_msg.width, pointcloud_msg.height))
# Simple callback for polar point cloud messages, just prints a message
def pySickScanPolarPointCloudMsgCallback(api_handle, pointcloud_msg):
pointcloud_msg = pointcloud_msg.contents # dereference msg pointer (pointcloud_msg = pointcloud_msg[0])
print("pySickScanPolarPointCloudMsgCallback (ROS-{}): api_handle={}, {}x{} pointcloud".format(api_handle, pointcloud_msg.width, pointcloud_msg.height))You should be receiving the pointcloud messages with the expected frequency. If this works as expected, you can process the data within the callback. We recommend to copy the scan data into a FIFO and to process the data in a separate thread to avoid bottlenecks due to blocking callbacks. For performance critical applications, we recommend to process the scan data using C++ and/or ROS.
The launch file sick_mrs_1xxx.launch should work, if you replace the default ip address of the lidar (i.e. "192.168.6.53" instead of "192.168.0.1"). Alternatively, you can overwrite parameter hostname by commandline. Did you get any error messages when using sick_mrs_1xxx.launch?
SuhaibAlHindi
I'm using a MRS1104C SICK LiDAR and I have noticed that when I open the interface of the LiDAR, it has a realtime scanning of the device but when I run this code, I found that the scans are much slower, each full scan takes about 1 sec.
Second note: the launch file sick_mrs_1xxx.launch does not support my LiDAR.
### launch file:
`
`
### code: ` def pySickScanCartesianPointCloudMsgToXYZ(pointcloud_msg): num_fields = pointcloud_msg.fields.size msg_fields_buffer = pointcloud_msg.fields.buffer field_offset_x = -1 field_offset_y = -1 field_offset_z = -1 for n in range(num_fields): field_name = ctypesCharArrayToString(msg_fields_buffer[n].name) field_offset = msg_fields_buffer[n].offset if field_name == "x": field_offset_x = msg_fields_buffer[n].offset elif field_name == "y": field_offset_y = msg_fields_buffer[n].offset elif field_name == "z": field_offset_z = msg_fields_buffer[n].offset cloud_data_buffer_len = (pointcloud_msg.row_step pointcloud_msg.height) # length of polar cloud data in byte assert(pointcloud_msg.data.size == cloud_data_buffer_len and field_offset_x >= 0 and field_offset_y >= 0 and field_offset_z >= 0) cloud_data_buffer = bytearray(cloud_data_buffer_len) for n in range(cloud_data_buffer_len): cloud_data_buffer[n] = pointcloud_msg.data.buffer[n] points_x = np.zeros(pointcloud_msg.width pointcloud_msg.height, dtype = np.float32) points_y = np.zeros(pointcloud_msg.width pointcloud_msg.height, dtype = np.float32) points_z = np.zeros(pointcloud_msg.width pointcloud_msg.height, dtype = np.float32) point_idx = 0 for row_idx in range(pointcloud_msg.height): for col_idx in range(pointcloud_msg.width): pointcloud_offset = row_idx pointcloud_msg.row_step + col_idx pointcloud_msg.point_step points_x[point_idx] = np.frombuffer(cloud_data_buffer, dtype = np.float32, count = 1, offset = pointcloud_offset + field_offset_x)[0] points_y[point_idx] = np.frombuffer(cloud_data_buffer, dtype = np.float32, count = 1, offset = pointcloud_offset + field_offset_y)[0] points_z[point_idx] = np.frombuffer(cloud_data_buffer, dtype = np.float32, count = 1, offset = pointcloud_offset + field_offset_z)[0] point_idx = point_idx + 1 return points_x, points_y, points_z
def pySickScanCartesianPointCloudMsgCallback(api_handle, pointcloud_msg): pointcloud_msg = pointcloud_msg.contents # dereference msg pointer (pointcloud_msg = pointcloud_msg[0]) plot_points_x, plot_points_y,plot_points_z = pySickScanCartesianPointCloudMsgToXYZ(pointcloud_msg) global x_points,y_points,z_points x_points = plot_points_x y_points = plot_points_y z_points = plot_points_z print(f"{plot_points_x},{np.max(plot_points_y):.2f}, and {np.max(plot_points_z):.2f}")
def pySickScanPolarPointCloudMsgCallback(api_handle, pointcloud_msg): global msg msg = pointcloud_msg pointcloud_msg = pointcloud_msg.contents # dereference msg pointer (pointcloud_msg = pointcloud_msg[0])
cartesian_pointcloud_callback = SickScanPointCloudMsgCallback(pySickScanCartesianPointCloudMsgCallback) SickScanApiRegisterCartesianPointCloudMsg(sick_scan_library, api_handle, cartesian_pointcloud_callback)`