A. Goudah
Generated dawn view from an inland vessel bow in river mist, with bow sensor and target detection overlays.

85-Class Vision and Fusion on a 105 m Vessel

Dr.-Ing. Ahmad Goudah is a Computer Vision and Perception Engineer building YOLOv8, LiDAR, DGPS, and AIS fusion for autonomous vessels and vehicles.

PDF, 2 pagesDownload CVEmail Ahmad

Generated river photograph (not Niedersachsen 22); the sensor and target overlays are illustrative.

Niedersachsen 22 Sensor Layout and Fusion Subset

Niedersachsen 22 (105 m) mounts bow, mast, side, and stern sensors. The thesis fuses six Basler acA4096-11gc cameras, Velodyne VLP-16 and Ouster OS2-64 LiDARs, Trimble BX992 GNSS, and AIS; bow stereo and mast radar remain unfused.

Side elevation and deck plan of the 105 m inland vessel Niedersachsen 22 showing numbered sensor nodes (#1 to #11) and foremast radar (C) across stern, sides, mast, and bow, distinguishing fused cameras, LiDARs, and GNSS from unfused stereo and radar.2.20 m13.65 mca. 73.10 m105.50 m9.30 m1234567, 89a/b1011C
  • Camera (Rear View) - Basler acA4096-11gc mono camera
  • Lidar (Short Range) - Velodyne VLP-16 3D LiDAR
  • Camera Bb (Forward view) - Basler acA4096-11gc port
  • Camera Stb (Forward view) - Basler acA4096-11gc starboard
  • Camera Bb (Rear view) - Basler acA4096-11gc port
  • Camera Stb (Rear view) - Basler acA4096-11gc starboard
  • Trimble GNSS Antennas - Trimble BX992 dual RTK DGPS
  • Stereo camera (a: Bb / b: Stb) at bow
  • Lidar (Long Range) - Ouster OS2-64 3D LiDAR
  • Camera Bug (Front View) - Basler acA4096-11gc bow
  • Radar (on Mast) - foremast navigation radar
Sensor positions and hull coordinates on Niedersachsen 22 after dissertation Figures 6.12 and 6.13, adapted from Jarofka et al. (2023). Accent marks fused sensors; dashed marks installed unfused sensors.

YOLOv5 and YOLOv8x Benchmark on 33 River Scenes

Pretrained 80-class COCO detectors (YOLOv5s through YOLOv8x) were benchmarked on 33 waterway and traffic scenes. Smaller weights merged shore cranes with barges or mislabeled cranes as airplanes, motivating a custom 85-class YOLOv8x detector (3.3 ms on A100).

River port frame with three overlapping green boat boxes (0.36, 0.38, 0.31) covering a berthed barge and shore crane.
Figure 4.8 (c)YOLOv5s draws three overlapping boat boxes across the crane and barge while missing the host bow.

Camera-LiDAR Object Matching and Web Fusion

LiDAR point clouds (.ply) are projected into top-down images and clustered with eight-centroid K-Means to compute Euclidean distance in meters for waterway objects labeled in synchronized shipboard camera frames.

Shipboard camera frames and LiDAR plot with boxes labeling Bridge Pillar, L Shore, R shore, Buoy, and Host Vessel.
Figure 5.4Dissertation Figure 5.4 (PDF p. 146): synchronized camera frames and LiDAR map annotating five waterway objects around the host vessel.
Top-down LiDAR plot with eight red centroid distances in meters and boxes for Bridge, shores, Buoyer, and Host Vessel.
Figure 5.7 (and Fig. 5.6)Dissertation Figures 5.6 and 5.7 (PDF pp. 148-149): eight K-Means LiDAR cluster distances matched to five waterway objects.

Six-Camera and GPS Fusion Monitor

The web monitoring program synchronizes six shipboard Basler camera streams around a central stitched panorama, logging folder source, millisecond frame interval, GPS timestamp, and latitude and longitude coordinates.

Folder
20210908_1
Interval
2000 ms
Frame
2
Date
08\09\2021
Time
10:17:22.798
Latitude
50.9331483333333 N
Longitude
6.96744666666667 E
Aft-looking port side camera view showing the river, cable-stayed bridge pylon, and port railing.Back Left Side
Forward bow camera view showing the cable-stayed bridge ahead across the river and bow housing.Front Camera
Aft-looking starboard side camera view showing the bridge, right riverbank buildings, and starboard deck.Back Right Side
Panoramic stitched view combining rear camera frames showing the vessel wake, river channel, and bridge.Stitched Image
Forward-looking starboard side camera view along the hull toward the riverbank and distant bridge.Front Right Side
Rear stern camera view showing the white propeller wake, stern bollards, and bridge astern.Back Camera
Forward-looking port side camera view showing the port gunwale, river surface, and shoreline buildings.Front Left Side
Figure 6.16Dissertation Figure 6.16 (PDF p. 184): web fusion interface showing six camera views, stitched rear panorama, and GPS telemetry.

Career Log: Alexandria to Duisburg (2001-2025)

  1. 2001 - 2006

    B.Sc. in Computer Engineering / Computer Science

    Arab Academy for Science, Technology and Maritime Transport (AASTMT)

    Alexandria, Egypt

    Graduated Very Good (GPA 3.1/4.0) with a thesis on hardware and software implementation of AVR microcontroller-based VoIP systems.

  2. 2005 - 2009

    Software & Web Developer

    Toyota Abdul Latif Jameel Co., Noortech Co., and Al-Amal

    Jeddah, Saudi Arabia & Alexandria, Egypt

    Built customer archiving and cheque systems at Toyota ALJ, access control and attendance systems at Noortech, and AVR LED displays at startup Al-Amal.

  3. 2010 - 2013

    M.Sc. in Computer Engineering / Computer Science

    Arab Academy for Science, Technology and Maritime Transport (AASTMT)

    Alexandria, Egypt

    Graduated with Distinction / Excellent (GPA 3.4/4.0); thesis on algorithms and applications for smart-grid three-phase automatic load balancing using genetic algorithms.

  4. 2009 - Present

    Data Scientist, Team Leader, System / Web Developer & Lecturer

    Arab Academy for Science, Technology and Maritime Transport (AASTMT) & Alexandria University

    Alexandria, Egypt

    Led university admission, payroll, elections, KPI dashboard, and scientific publishing (apc.aast.edu) platforms while teaching AI, pattern recognition, big data, and embedded systems.

  5. 2019 - 2025

    Doctoral Researcher (Dr.-Ing., cum laude)

    University of Duisburg-Essen, Chair of Mechatronics

    Duisburg, Germany

    Built 85-class YOLOv8x detection, 3D LiDAR K-Means ranging, and Camera-LiDAR-DGPS-AIS fusion on the 105 m vessel Niedersachsen 22.

Dissertation and Peer-Reviewed Publications

Doctoral dissertation and five papers spanning YOLOv8 inland vessel object detection, power consumption prediction, and genetic-algorithm smart-grid load balancing.

  1. 2025

    Real-time Detection Models for Surrounded Environmental Objects in Highly Automated Vehicles and Inland Vessels

    Ahmad Abdullatif Metwally Goudah Hussien

    Doctoral Dissertation (Dr.-Ing.), Faculty of Engineering, University of Duisburg-Essen

  2. 2023

    Object Detection in Inland Vessels Using Combined Trained and Pretrained Models of YOLO8

    Goudah, A. A., Jarofka, M., El-Habrouk, M., Schramm, D., and Dessouky, Y. G.

    ACE, 3(2)

  3. 2022

    Customer Active Power Consumption Prediction for the Next Day Based on Historical Profile

    Goudah, A. A., El-Habrouk, M., Schramm, D., and Dessouky, Y. G.

    ACE, 2(1)

  4. 2020

    Smart Electric Grids Three-Phase Automatic Load Balancing Applications Using Genetic Algorithms

    Goudah, A. A., Schramm, D., El-Habrouk, M., AbouEL Farag, A., and Dessouky, Y. G.

    Renewable Energy and Sustainable Development (RESD), 6(1)

  5. 2019

    Smart Load Balancing in Smart Grids Using Genetic Algorithms

    Goudah, A., Gaber, Y., Abul-Farag, A., and Mostafa, H.

    NRSC

  6. 2016

    Three-Phase Automatic Load Balancing for Single Phase Loads in Smart Grids using Genetic Algorithms

    Arboleya, P., and Goudah, A.

    Poster paper, Department of Electrical Engineering, University of Oviedo, Spain

Generated photograph from the bow of an inland cargo vessel on a calm river in morning light.

Contact Dr.-Ing. Ahmad Goudah in Dortmund

Based in Dortmund, Germany, bringing 6 years of doctoral perception research and 15+ years of software engineering to shipped products. Reachable by email, WhatsApp (+49 178 3606461), or LinkedIn.

PDF, 2 pagesDownload CVDownload DOCXDOCX, 2 pagesEmail AhmadWhatsAppLinkedIn