Merge pull request #37 from antoniomtz/loss-prevention-logic

docs: loss prevention business logic

docs_src/use-cases/loss-prevention/advanced.md

@@ -41,6 +41,29 @@ To find all supported classes by YOLOv8, you can find them in this file `src/ext
 CLASS_IDS="0,2"
 ```
 
+## MQTT Inference Export and ROI Detection
+
+This applicaiton enables monitoring object entry and exit within a defined Region of Interest (ROI), allowing real-time event tracking and external message handling.
+The `yolov8s_roi.json` pipeline exports the inference data through MQTT using mosquitto broker defined in the `docker-compose.yml` file. 
+
+To change the default MQTT URL, edit the file `src/pipelines/yolov8s_roi.sh` and update the variable `MQTT_HOST="127.0.0.1:1883"`.
+
+We have developed a business logic application in `src/app/loss_prevention.py` that tracks objects entering and exiting a defined ROI, generating corresponding events.
+
+To configure the app to connect to external MQTT broker, modify the `src/docker-compose.yml` and change the following env variables:
+
+| Variable   | Default        | Description                              |
+|------------|----------------|------------------------------------------|
+| `MQTT_URL` | `127.0.0.1`    | MQTT Broker URL                          |
+| `MQTT_PORT`| `1883`         | MQTT Broker Port                         |
+| `MQTT_TOPIC` | `event/detection` | Topic for publishing inference data   |
+| `ROI_NAME` | `BASKET`       | The name of the ROI used to filter objects |
+
+
+The following diagram illustrates the containers running:  
+
+[![MQTT export](./images/mqtt-diagram.jpg)](./images/mqtt-diagram.jpg)
+
 
 For enviroments variables, follow the same tutorial as the automated self checkout [HERE](../automated-self-checkout/advanced.md)