Automated Crop Anomaly Detection with AI and Vector Search

This n8n workflow streamlines the ingestion and processing of agricultural crop image datasets for intelligent analysis. It begins by fetching images from a Google Cloud Storage bucket, specifically targeting the 'agricultural-crops' prefix. Each image is then processed to extract relevant information like its public URL and crop name. Crucially, the workflow leverages VoyageAI's multimodal embeddings to create rich vector representations of each image. These embeddings, along with metadata, are batched and then efficiently uploaded to a Qdrant vector database. The workflow includes steps for checking and creating the Qdrant collection, ensuring a robust foundation for subsequent AI-driven anomaly detection and KNN classification tasks.