Agnet Object-Detection and Alert System with TensorFlow-Serving and Agnet-API
Igbinidu-Uwuigbe, Augustine (2022)
Igbinidu-Uwuigbe, Augustine
2022
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2022100120706
https://urn.fi/URN:NBN:fi:amk-2022100120706
Tiivistelmä
Recently, Machine Learning has played a major role in the field of science and technology. Object detection in Computer vision systems has gained a lot of use in many industries and is still being developed for many use cases today. It is now an essential technology for many monitoring systems, especially for detecting threats or tracking items.
Agnet as a secure end-to-end communication solution for Airbus can be used as a monitoring system for cameras and drones. Having an integrated and deployable smart monitoring system powered by computer vision technology for risk reporting will bring value to Agnet.
This thesis aimed to investigate the usage of Computer Vision to create a smart monitoring and risk-reporting system for Agnet. To achieve these objectives the theoretical structure will include every step taken in building this computer vision and alerting system: acquisition, processing, model training, model deployment, inference, and risk reporting. This will give an extensive perspective of the key parts and their application. This is followed by the description of Agnet-API, which is the application programming interface for Agnet. Finally, a proof-of-concept Computer Vision and riskreporting system to demonstrate its practicality in a production environment.
Based on the studies illustrated in this paper, it can be concluded that Computer Vision through Agnet-API is a viable and cheap smart-monitoring solution for Organizations. The desired objectives were fulfilled and the applicability of this solution to several communication systems is provided.
Agnet as a secure end-to-end communication solution for Airbus can be used as a monitoring system for cameras and drones. Having an integrated and deployable smart monitoring system powered by computer vision technology for risk reporting will bring value to Agnet.
This thesis aimed to investigate the usage of Computer Vision to create a smart monitoring and risk-reporting system for Agnet. To achieve these objectives the theoretical structure will include every step taken in building this computer vision and alerting system: acquisition, processing, model training, model deployment, inference, and risk reporting. This will give an extensive perspective of the key parts and their application. This is followed by the description of Agnet-API, which is the application programming interface for Agnet. Finally, a proof-of-concept Computer Vision and riskreporting system to demonstrate its practicality in a production environment.
Based on the studies illustrated in this paper, it can be concluded that Computer Vision through Agnet-API is a viable and cheap smart-monitoring solution for Organizations. The desired objectives were fulfilled and the applicability of this solution to several communication systems is provided.