Increasing dimension of bounding box

[tadteo]

I am using octomap server to explore a zone. The problem is that I am unable to modify the bounding box of the map. The points are passed in the map reference frame since I have done a package to do transform the point cloud from the robot frame to the map frame. For this reason, I think that the map is limited by the sensor/max_range parameter as it takes the origin as the position of the sensor even if the sensor is moving with the drone.
My questions are:

• There is a way to modify the overall bounding box of the octomap from the launchfile?
• Is correct to provide the point cloud in the map reference frame or the octomap automatically convert the reference frame of the points? In that case, is the senor/max_range parameter correct since in reality, the sensor is not in that position?

Thank you in advance for the information

[GeneralAdmin]

Have you found a solution? I have the same issue.

[tadteo]

In the end, I have increased the sensor/max_range parameter ad that worked for me. But I think there is a parameter to change the external bounding box of the octomap.

[GeneralAdmin]

This is the same issue than in:

#69

I've also seen that the folks in:

https://github.com/PX4/PX4-Avoidance/blob/master/global_planner/src/nodes/global_planner_node.cpp

Have done this to solve the issue:

// Go through obstacle points and store them
void GlobalPlannerNode::depthCameraCallback(const sensor_msgs::PointCloud2& msg) {
try {
// Transform msg from camera frame to world frame
ros::Time now = ros::Time::now();
listener_.waitForTransform(frame_id_, camera_frame_id_, now, ros::Duration(5.0));
tf::StampedTransform transform;
listener_.lookupTransform(frame_id_, camera_frame_id_, now, transform);
sensor_msgs::PointCloud2 transformed_msg;
pcl_ros::transformPointCloud(frame_id_, transform, msg, transformed_msg);
pcl::PointCloudpcl::PointXYZ cloud; // Easier to loop through pcl::PointCloud
pcl::fromROSMsg(transformed_msg, cloud);

// Store the obstacle points
for (const auto& p : cloud) {
  if (!std::isnan(p.x)) {
    // TODO: Not all points end up here
    Cell occupied_cell(p.x, p.y, p.z);
    global_planner_.occupied_.insert(occupied_cell);
  }
}
pointcloud_pub_.publish(msg);

} catch (tf::TransformException const& ex) {
ROS_DEBUG("%s", ex.what());
ROS_WARN("Transformation not available (%s to %s)", frame_id_.c_str(), camera_frame_id_.c_str());
}
}

I'll try to do something like that in my code...