Realsense RGBD msg format and Rtabmap RGBD msg format diffrent !

[AltinayGrass]

I am trying to use realsense D435i Depth camera with rtabmap. I can able to get map with depth and color images combination. But when I want to connect to the rgbd topic, I see that the rgbd msg formats are different. Is there a setting I can make for this?

Cannot echo topic '/camera/rgbd', as it contains more than one type: [realsense2_camera_msgs/msg/RGBD, rtabmap_msgs/msg/RGBDImage]

[matlabbe]

Looks like realsense2 has their own version of rgbd msg.

• Realsense: https://github.com/IntelRealSense/realsense-ros/blob/ros2-master/realsense2_camera_msgs/msg/RGBD.msg (created in 2023)
• Rtabmap: https://github.com/introlab/rtabmap_ros/blob/ros2/rtabmap_msgs/msg/RGBDImage.msg (created in 2016)

You cannot connect realsense's RGBD msg to rtabmap nodes. I think however that it would not be difficult to convert a realsense2_camera_msgs/RGBD topic to rtabmap_msgs/RGBDImage topic, as the same fields exist in the rtabmap_msgs version.

Here a chatgpt example to convert them (I didn't test it, but looks roughly what I was thinking):

import rclpy
from rclpy.node import Node

from realsense2_camera_msgs.msg import RGBD as RealSenseRGBD
from rtabmap_msgs.msg import RGBDImage as RtabmapRGBDImage
from std_msgs.msg import Header
from sensor_msgs.msg import CameraInfo, Image, CompressedImage

class RGBDConverter(Node):
    def __init__(self):
        super().__init__('rgbd_converter')
        self.subscription = self.create_subscription(
            RealSenseRGBD,
            '/camera/rgbd',  # Change to your actual input topic
            self.rgbd_callback,
            10)
        self.publisher = self.create_publisher(
            RtabmapRGBDImage,
            '/rtabmap/rgbd_image',  # Output topic
            10)

    def rgbd_callback(self, msg: RealSenseRGBD):
        converted = RtabmapRGBDImage()

        # Copy shared fields
        converted.header = msg.header
        converted.rgb_camera_info = msg.rgb_camera_info
        converted.depth_camera_info = msg.depth_camera_info
        converted.rgb = msg.rgb
        converted.depth = msg.depth

        # Optional: Leave other fields empty/default
        converted.rgb_compressed = CompressedImage()
        converted.depth_compressed = CompressedImage()
        converted.key_points = []
        converted.points = []
        converted.descriptors = []
        # global_descriptor can be left default or initialized if needed

        self.publisher.publish(converted)
        self.get_logger().info('Converted and published RGBDImage')

def main(args=None):
    rclpy.init(args=args)
    node = RGBDConverter()
    rclpy.spin(node)
    node.destroy_node()
    rclpy.shutdown()