brkljac / ViTech

Computer Vision Gateway for real-time stream processing

ViTech

Computer Vision Gateway for real-time stream processing

Gstreamer, OpenCV, TensorRT, H.265

Tested on Jetson Nano

---

• Stream processing and AI deployment with low power consumption and low memory footprint

• Hardware encoding and decoding for any combination of user defined input and output image/video streams

• Versatile tool for performing both low-level and high-level vision tasks

• Jetson implementation utilizing heterogeneous computing environment, suitable for outdoor application scenarios

• Support for model translators and compilers enabling creation of optimized AI inference engines

• Test examples: semantic segmentation, image classification, real-time monocular depth estimation on drone platform

---

Functional blocks:

---

Examples

Semantic segmentation:

--->>> testVideoOut1_pedestrian

--->>> testVideoOut2_traffic

---

Monocular depth estimation:

--->>> testVideoOut3_unprocessed

--->>> testVideoOut3_depth

---

UAV integration:

---

If you would like to use this software, please consider citing the following publication:

• Brkljač, B., Brkljač, M., Antić, B. (2024). Computer vision gateway for real-time on-site analytics in smart cities and peri-urban agriculture. In Proceedings: 1st International Scientific and Professional Conference - ALFATECH 2024 – Smart cities and modern technologies, Belgrade, Serbia, March 15, 2024, Vol. 1,

    @inproceedings{Brkljac_Computer_vision_gateway_2024,
    author = {Brklja{\v{c}}, Branko and Brklja{\v{c}}, Milan and Anti{\'c}, Boris},
    booktitle = {1st International Scientific and Professional Conference - ALFATECH 2024 – Smart cities and modern technologies, Belgrade, Serbia, March 15, 2024, Proceedings},
    doi = {-},
    month = mar,
    pages = {1--12},
    title = {{Computer vision gateway for real-time on-site analytics in smart cities and peri-urban agriculture}},
    year = {2024}
    }

​