Custom Connector to make Get Request to Google Big Query API??

Question

We are working on a pilot project with Google to use the Google Vision API to detect train's at railroad crossings by monitoring live stream cameras. The Google Vision API would then out put its data in JSON format. However, in order to make a Get request to the Google Vision API, we would need to make that request using the Google Big Query API.

Has anyone ever made a custom input connector that could make this type of API call? Is it possible? I am not a developer but I do understand basic development procedures. Can anyone add their two cents on this idea?

Thanks

JeffSilberberg · Answer

@Gene_Sipes

Thank you, I have a better idea of what you are after and after reading this post a better idea of the API you are interfacing with.   https://medium.com/@annycarolinegnr/using-google-vision-api-22d1fdb755d8

I would suggest you start by interfacing with Postman to see if you have the details and headers correct and specifically what is being returned by the API.  https://learning.postman.com/docs/introduction/overview/

Then you should be able to build a Simple Polling Feed to get your information.  Having said that I will tell you that for a similar interface to Motorola's Avigilon Alta, I built a small refactor process that runs on one of our servers and pools the Alta interface, and then posts to Velocity.

Additionally, depending on what your Emergency Services departments are doing, an Alta or other solution may offer more capabilities than the Google offering as well as the ability to place your cameras where you need them.

https://www.avigilon.com/alta

Gene_Sipes · Answer

I'm not sure how a hot box detector helps either, seems its more for detecting rail issues to prevent derailment. Additionally, we are a local municipality and have zero access to install or otherwise interfere with Union Pacific's operations. We have crossing closure detection at three intersections in town, but it was costly and required extension engineering and safety plans to install. Using a cheap camera's livestream and AI video detection is a passive approach. It is theoretically much cheaper and faster than installing equipment that is wired to crossings. The purpose of this exercise is to be able to communicate via a web app, text alert, and email when a train crossing arm is down due to passing or stalled train (preventing vehicles from crossing). This can benefit the public by alerting them to longer than normal train crossing closures.

JeffSilberberg · Answer

Not sure what the goal of this project is but I would think that capturing data from the Hotbox detectors on a train would be a lot more reliable than using Cameras and AI.

https://cdlelectric.com/hot-box-detectors/