Project report on smart animal rescue system with AI recognition
Loading...
Files
Date
2026
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
New Horizon College
Abstract
Managing stray and domestic animal welfare stands as a major municipal challenge in rapidly developing urban environments. Traditional animal rescue networks are highly fragmented and suffer from extreme operational friction, relying on manual phone calls, unverified communication lines, and vague verbal descriptions of the scene. This lack of structure leads to inaccurate location mapping, making it difficult for rescue teams to navigate to a crisis site efficiently, while central dispatch centers lack real-time data visibility to determine medical severity. To bridge this gap between public citizens and specialized regional dispatch squads, this project introduces the Smart Animal Rescue System with AI Recognition, an end-to-end automated emergency response ecosystem designed to streamline civic intake and field resource logistics. The application architecture relies on a dual-environment network ecosystem that runs entirely on localized infrastructure via loopback configurations (localhost:5006). On the citizen-facing presentation layer, the framework uses native hardware-level API event handlers to capture high-precision geographic telemetry metrics from the device's location sensors. By extracting absolute latitude and longitude floating-point decimal coordinates silently in the background, the system entirely removes human navigation error from the tracking loop. Simultaneously, a multi-modal computer vision model handles incoming multi-part form-data streams. When a citizen attaches a local photo asset, the neural network executes deep feature extraction layers across pixel matrices to perform real-time animal species identification. The system then merges this computed classification label with user-supplied qualitative trauma data strings through an internal conditional rules engine to calculate a severity priority rating and deliver on-site first-aid guidelines. The backend database tier is governed by relational storage tables designed for strict transaction security and absolute ACID compliance. When a report satisfies all client-side validation guard rails, it is assigned a short, user-friendly auto-incrementing integer tracking number (e.g., #27) to eliminate communication friction. On the administrative control end, a centralized management terminal consolidates active records into a tabular logging grid. The console anchors four interactive analytics widgets displaying total emergencies, pending cases, rescue results, and active volunteers, functioning as dynamic list filters that isolate sub-arrays on click without page reloads. Furthermore, the dashboard automatically converts coordinates into visual marker pins plotted across an integrated spatial tracking map while calculating paths to match each incident to the closest physical regional rescue hub. The operational lifecycle concludes through a strict dual-photo verification closure loop where field personnel submit a real-time post-rescue photograph to the server, firing atomic SQL update queries that transition the case state from "Pending" to "Completed" across the admin console and public tracking paths concurrently. This project successfully replaces legacy manual processes with a highly resilient public utility infrastructure that reduces communication gaps and accelerates life-saving response operations at a systemic scale.