基于Intel RealSense D435的户外苹果园3D重建问题求助

研究背景与问题 #

• 硕士论文聚焦苹果园3D重建，采用Intel RealSense D435相机执行重建工作
• 已知该相机支持室外作业，但实际采集树木点云时，代码输出结果与realsense_viewer显示内容差异极大，调整depth_units参数后仍无改善
• 项目需利用ICP（Iterative Closest Point，迭代最近点）算法对相机采集的多帧点云进行配准，最终生成完整场景点云

当前使用的Python代码（室内效果优异，室外效果极差） #

# First import the library
import pyrealsense2 as rs
from pynput import keyboard

counter = 0

def on_press(key):
    global counter
    if key == keyboard.Key.space:
        print("Saving pcd_{}.ply".format(counter))
        points.export_to_ply("/home/agilex/thesis_ws/src/reconstruction/img/point_cloud/pcd_{}.ply".format(counter), color_frame)
        print("Done")

def on_release(key):
    global counter
    counter += 1
    if key == keyboard.Key.esc:
        # Stop listener
        return False

if __name__ == '__main__':
    pipe = rs.pipeline()
    config = rs.config()
    config.enable_stream(rs.stream.depth)
    config.enable_stream(rs.stream.color)
    align_to = rs.stream.color
    align = rs.align(align_to)
    profile = pipe.start(config)
    device = profile.get_device()
    depthsensor = device.first_depth_sensor()
    if depthsensor.supports(rs.option.depth_units):
        # This value in realsense viewer in micrometers and here is in meters
        depthsensor.set_option(rs.option.depth_units,0.000755)
    frames = pipe.wait_for_frames()
    aligned_frames = align.process(frames)
    aligned_depth_frame = aligned_frames.get_depth_frame()
    color_frame = aligned_frames.get_color_frame()
    pc = rs.pointcloud()
    points = rs.points()
    pc.map_to(color_frame)
    points = pc.calculate(aligned_depth_frame)
    # Collect events until released
    with keyboard.Listener(on_press=on_press,on_release=on_release) as listener:
        listener.join()

    listener = keyboard.Listener(on_press=on_press,on_release=on_release)
    listener.start()

效果对比 #

• RGB图像：室外苹果园场景的彩色画面
• 深度图像：对应场景的深度画面
• realsense_viewer导出的.ply文件效果：点云完整，能清晰呈现树木及周边环境结构
• 本人代码输出的.ply文件效果：点云缺失严重，室外场景重建效果极差

需求与求助 #

目标是获取深度与彩色流，生成并保存多帧场景点云，再通过ICP算法得到完整场景点云。恳请提供基于ROS或SDK的解决方案建议。

内容的提问来源于stack exchange，提问作者Jomana Abdelmoaty