The analytics data produced in a camera is powerful for all sorts of use cases. Different use cases needs different consumer-friendly data structures and behaviour. The bandwidth should be held to a minimum as well as the need to process and filter the data on the consumer side. I share three ACAPs, all using the same camera video anlytics producing different data for easy integraion with most use case such as automation, messeging, forensic, counting and forcasting.
An alternative solution that provides full flexibility and customization is Customized Objects Analytics in Node-RED.

Object Tracker

Sends a JSON structure on MQTT while the object moves through the scene. The data includes the bounding box, age, distance, direction, and object type (human, car, bus, truck, motorcycle, face, license plate). You can optimize the behaviour by applying different types of filters to remove unnessesery detections. Object Tracker targets real-time systems for automation, messeging and control.

Download Object Tracker

Instructions for configuration and integration is in the README.md file inside the ZIP-file.

Object Path

Sends one data JSON structure on MQTT after an object left the scene. The data includes an array of sampled positions (with duration) the object had in the scene, age, distance, direction, dwell time, and object type (human, car, bus, truck, motorcycle, face, license plate). Object Path targets post-processing system where data needs to be stored in a database.

Download Object Path

Instructions for configuration and integration is in the README.md file inside the ZIP-file.

Object Counter

Counts object moving up, down, left and right. The data includes direction, counter value and individual counter values for each object type. The data is fired as en Event and integration can be accomplished with the cameras event system, ONVIF Event stream or MQTT (using the camera client or SIMQTT). User can configure events to be fired on each counter change or time based. Users can select automatic reset on every midnight or time group. The counter values are retained if the camera reboots. Object Counter targets post-processing system where counter data is stored in a database.

Download Object Counter

Instructions for configuration and integration is in the README.md file inside the ZIP-file.