Blanco Romero, Franco T. (2024) Desarrollo del modelado 3D de una habitación usando visión por computadora y/o un LIDAR para la filmadora gamma / Development of 3D modeling of a room using computer vision and/or a LIDAR for gamma camera. Proyecto Integrador Ingeniería en Telecomunicaciones, Universidad Nacional de Cuyo, Instituto Balseiro.
![[img]](/style/images/fileicons/application_pdf.png)
| PDF (Tesis) Español 14Mb |
Resumen en español
La filmadora de radiación gamma de mano desarrollada por INVAP, llamada comercialmente ”Dragonfly”, permite realizar mediciones de tasa de radiación gamma y determinar la dirección de donde proviene. Hay estimaciones que se pueden realizar conociendo la tasa de radiación, por ejemplo, el estimar la tasa de radiación en una posición distinta a la que uno se encuentra, o estimar la actividad de una fuente radioactiva. Sin embargo, solo se pueden estimar si se conoce la distancia a la fuente. En este trabajo se desarrollaron distintas soluciones de medición de distancia utilizando ROS2, que es un framework orientado a robótica. Para realizar las mediciones se utilizaron un LIDAR y una cámara estéreo, para implementarse con el Drangonfly, con el fin de poder extender la familia de este producto y ampliar sus aplicaciones. Para la cámara estéreo se hizo un estudio de los modelos de las cámaras y como a partir de ellas se pueden obtener estimaciones de distancia de los planos que captura. Utilizando una placa de desarrollo Jetson Nano para hacer la adquisición de los sensores utilizando ROS2. También se realizaron gemelos digitales de los distintos sensores, utilizando el simulador Gazebo, para poder realizar mediciones sin tener que estar utilizando los sensores físicos, lo cual optimiza el tiempo de prueba de los algoritmos. Se desarrollaron algoritmos en ROS2 que permitieron la implementación de los mensajes que da como solución el Dragonfly y se mapearon a las correspondientes distancias obtenidas de los distintos sensores. A partir de ellas se pudieron analizar ciertos casos en los que conviene usar el LIDAR o la cámara estéreo. Por ultimo se realizó un análisis sobre la edometría y SLAM, y como esto acerca el Dragonfly a implementarse en conjunto con la realidad virtual realizando mapeos y reconstrucciones de salas o entornos.
Resumen en inglés
The handheld gamma radiation video recorder developed by INVAP, commercially named ”Dragonfly,” allows for measurements of gamma radiation rate and determining its source. Estimations can be made based on the radiation rate, for example, estimating the radiation rate at a position different from one’s own or estimating the activity of a radioactive source. However, these estimations can only be made if the distance to the source is known. In this work, various distance measurement solutions were developed using ROS2, which is a framework geared towards robotics. To perform the measurements, a LIDAR and a stereo camera were used to be implemented with the Dragonfly, aiming to expand the family of this product and broaden its applications. For the stereo camera, a study of camera models was conducted and how distance estimates from the captured planes can be obtained. A Jetson Nano development board was used to acquire the sensors using ROS2. Digital twins of the different sensors were also created using the Gazebo simulator to facilitate measurements without the need for physical sensors, thus optimizing algorithm testing time. Algorithms were developed in ROS2 to implement the messages provided by Dragonfly as solutions and mapped them to corresponding distances obtained from the various sensors. Certain cases where it is advantageous to use either the LIDAR or the stereo camera could be analyzed based on these mappings. Finally, an analysis was conducted on odometry and SLAM, and how this brings Dragonfly closer to being implemented in conjunction with virtual reality for mapping and reconstructing rooms or environments.
| Tipo de objeto: | Tesis (Proyecto Integrador Ingeniería en Telecomunicaciones) |
|---|---|
| Palabras Clave: | Argentine INVAP; INVAP Argentina; [Stereo camera; Cámara estéreo; Gamma video recorder; Filmadora gamma; Distance estimation; Estimación de distancia; LIDAR] |
| Referencias: | [1] Pesco, V. D. Cadena de cálculo para cámara gamma. 2 [2] Detector de centelleo. URL https://es.wikipedia.org/wiki/Centelleador. 2 [3] Autoridad Regulatoria Nuclear. URL https://www.argentina.gob.ar/arn. 3 [4] Rico, F. A Concise Introduction to Robot Programming with ROS2. CRC Press, 2022. URL https://books.google.com.ar/books?id=iCKIEAAAQBAJ. 7, 9 [5] Ros. URL https://www.ros.org/. 8 [6] Wiki ros. URL http://wiki.ros.org/. 8 [7] Ros2 understanding nodes. URL https://docs.ros.org/en/humble/Tutorials/ Beginner-CLI-Tools/Understanding-ROS2-Nodes/Understanding-ROS2-Nodes.html. 9 [8] Ros2 nodes. URL https://docs.ros.org/en/humble/Concepts/Basic/About-Nodes.html. 9 [9] Ros2 topics. URL https://docs.ros.org/en/humble/Concepts/Basic/About-Topics.html.9 [10] Ros2 understanding topics. URL https://docs.ros.org/en/humble/Tutorials/ Beginner-CLI-Tools/Understanding-ROS2-Topics/Understanding-ROS2-Topics.html.9 [11] std msgs. URL https://index.ros.org/p/std_msgs/github-ros2-common_interfaces/#humble. 10 [12] geometry msgs. URL https://index.ros.org/p/geometry_msgs/github-ros2-common_interfaces/#humble. 10 [13] sensor msgs. URL https://index.ros.org/p/sensor_msgs/github-ros2-common_interfaces/#humble. 10 [14] Estructura del mensaje tipo camerainfo. URL http://docs.ros.org/en/melodic/api/sensor_msgs/html/msg/CameraInfo.html. 10 [15] Estructura del mensaje tipo image. URL http://docs.ros.org/en/melodic/api/sensor_msgs/html/msg/Image.html. 10 [16] Estructura del mensaje tipo laserscan. URL http://docs.ros.org/en/melodic/api/sensor_msgs/html/msg/LaserScan.html. 10 [17] Estructura del mensaje tipo pointcloud2. URL http://docs.ros.org/en/melodic/api/sensor_msgs/html/msg/PointCloud2.html. 10 [18] Estructura del mensaje tipo header. URL http://docs.ros.org/en/melodic/api/std_msgs/html/msg/Header.html. 10 [19] Estructura del mensaje tipo pointfield. URL http://docs.ros.org/en/melodic/api/sensor_msgs/html/msg/PointField.html. 12 [20] Herramienta rqt de ros. URL http://wiki.ros.org/rqt. 12 [21] Herramienta rviz de ros. URL http://wiki.ros.org/rqt. 12 [22] Simulador gazebo. URL https://gazebosim.org/home. 13 [23] Sdf simulation description format. URL http://sdformat.org/. 14 [24] Ros2 urdf. URL https://docs.ros.org/en/humble/Tutorials/Intermediate/URDF/URDF-Main.html. 17 [25] Ros2 tf2. URL https://docs.ros.org/en/humble/Tutorials/Intermediate/Tf2/Tf2-Main.html. 17 [26] Raspberry pi. URL https://www.raspberrypi.com/. 19 [27] Jetson nano. URL https://www.nvidia.com/es-la/autonomous-machines/embedded-systems/jetson-nano/product-development/. 19 [28] Jetpack sdk. URL https://developer.nvidia.com/embedded/jetpack. 20 [29] Tfmini-s. URL https://en.benewake.com/TFminiS/. 22 [30] Rplidar-a1. URL https://www.slamtec.ai/product/slamtec-rplidar-a1/. 23 [31] Paquete sllidar ros2. URL https://github.com/Slamtec/sllidar_ros2/. 23 [32] Repositorio de contenedores de docker de dustynv. URL https://hub.docker.com/r/dustynv/ros/. 24 [33] Imx219-83. URL https://www.waveshare.com/wiki/IMX219-83_Stereo_Camera. 25 [34] Gstreamer. URL https://gstreamer.freedesktop.org/. 25 [35] csi-camera-ros-wrapper. URL https://github.com/packbionics/csi-camera-ros-wrapper.25 [36] Paquete ros deep learnig. URL https://github.com/dusty-nv/ros_deep_learning. 25 [37] Fusion360 de autodesk. URL https://autodesk.com/products/fusion-360/overview. 27 [38] Shapiro, L., Stockman, G. Computer Vision. Prentice Hall, 2001. 29, 30, 32 [39] Foto de ruta de noche extraido de una nota del diario”minuto uno”. URL https://www.minutouno.com/economia/ en-las-rutas-se-siente-el-exodo-turistico-el-fin-semana-largo-n333276. 30 [40] Daniel Scharstein, A. V.-R., Szeliski, R. 2003 stereo datasets with ground truth. URL https://vision.middlebury.edu/stereo/data/scenes2003/. 32 [41] Daniel Scharstein, P. U., Szeliski, R. 2001 stereo datasets with ground truth. URL https://vision.middlebury.edu/stereo/data/scenes2001/. 33 [42] Patrick Mihelich, J. L., Kurt Konolige. stereo image proc ros2 package. URL http://wiki.ros.org/stereo_image_proc. 33, 39, 40 [43] Opencv. URL https://opencv.org/. 33 [44] OpenCV clase StereoBM. URL https://docs.opencv.org/3.4/d9/dba/classcv_1_1StereoBM.html#details. 33 [45] Hirschmuller, H. Stereo processing by semiglobal matching and mutual information. IEEE Transactions on Pattern Analysis and Machine Intelligence, 30 (2), 328–341, 2008. 33 [46] Estructura del mensaje de mapa de disparidad. URL http://docs.ros.org/en/api/stereo_msgs/html/msg/DisparityImage.html. 34 [47] Adrián Kaehler, G. B. Learning OpenCV 3. O’Reilly Media, Inc., 2016. 35, 37 [48] Camera Calibration Mathworks. URL http://wiki.ros.org/camera_calibration. 35, 37 [49] Camera calibration and 3d reconstruction. URL https://in.mathworks.com/help/vision/ug/camera-calibration.html. 36 [50] Elso, I. J. Reconstrucci´on de objetos en 3d mediante un sistema est´ereo binocular. URL https://academica-e.unavarra.es/xmlui/bitstream/handle/2454/32849/MemoriaTFG.pdf?sequence=1&isAllowed=y. 36, 37 [51] Ros2 camera calibration. URL http://wiki.ros.org/camera_calibration. 37 [52] Calibrate camera from opencv. URL https://docs.opencv.org/2.4/modules/calib3d/doc/camera_calibration_and_3d_reconstruction.html. 37 [53] Zhang, Z. A flexible new technique for camera calibration. IEEE Transactions on Pattern Analysis and Machine Intelligence, 22 (11), 1330–1334, 2000. 37 [54] Image proc node. URL http://wiki.ros.org/image_proc. 38, 39 [55] Protocol buffers. URL https://protobuf.dev/. 41 [56] Calib.io. URL https://calib.io/. 44 [57] Choosinggoodstereoparameters. URL http://wiki.ros.org/stereo_image_proc/Tutorials/ChoosingGoodStereoParameters. 46 [58] NVIDIA Isaac ROS. URL https://developer.nvidia.com/isaac-ros. 54 [59] Paquete rf2o laser odometry. URL https://github.com/MAPIRlab/rf2o_laser_odometry?tab=readme-ov-file. 55 [60] Jaimez, M., Monroy, J. G., González-Jiménez, J. Planar odometry from a radial laser scanner. arange flow-based approach. En: 2016 IEEE International Conference on Robotics and Automation (ICRA), p´ags. 4479–4485. 2016. 55 [61] Paquete slam toolbox. URL http://wiki.ros.org/slam_toolbox. 55 [62] Paquete rtabmap ros. URL http://introlab.github.io/rtabmap/. 55 [63] RTABMAP. URL http://introlab.github.io/rtabmap/. 55 [64] ORB-SLAM3. URL https://github.com/UZ-SLAMLab/ORB_SLAM3. 55 [65] Introduction to FAST (Features from Accelerated Segment Test). URL https://medium.com/@deepanshut041/introduction-to-fast-features-from-accelerated-segment-test-4ed33dde6d65.56 [66] Zed2. URL https://www.stereolabs.com/products/zed-2. 56 [67] Intel Depth Camera D435 URL https://www.intelrealsense.com/depth-camera-d435/. 56 |
| Materias: | Ingeniería > Sistemas embebidos |
| Divisiones: | INVAP |
| Código ID: | 1254 |
| Depositado Por: | Tamara Cárcamo |
| Depositado En: | 12 Sep 2024 15:26 |
| Última Modificación: | 12 Sep 2024 15:26 |
Personal del repositorio solamente: página de control del documento
