Client: SlateGreen House

Overview: 

SlateGreen House is leveraging an AI-assisted, static information site developed by the Association for the Advancement of Business AI (AABA).   It supports Tom Hallewell in providing new plant owners with detailed background information about their purchase. The system combines image recognition, taxonomy generation, and deterministic URL creation using Slipbox, enabling direct, durable QR code integration with each plant.

Workflow:

  1. Image Capture:
    Tom Hallewell photographs plants intended for sale or gifting.

  2. Image Upload:
    Images are uploaded to a designated directory on the server by Tom.

  3. Automated Identification:

    • Each image is submitted to the PlantNet API: https://my-api.plantnet.org/v2/identify/all
    • The API returns JSON metadata describing the plant.
    • This metadata is stored on the server.

  4. Metadata Processing & Taxonomy:

    • makehash.py reformats the raw API JSON into the structure required by Slipbox.

    • It also builds a navigable taxonomy organized by:

      • scientific name/ – Full binomial species names (e.g., Salvia officinalis) Example
      • family/ – Botanical families (e.g., Lamiaceae)
      • genus/ – Genus-level grouping (e.g., Salvia)
      • gbif/ – Global Biodiversity Information Facility (GBIF) IDs
        GBIF assigns globally unique identifiers to taxa (species, genus, etc.) in its biodiversity data infrastructure. These IDs are used to reference species occurrence records, taxonomy, and datasets from hundreds of institutions worldwide. Format: numeric ID in a URI, e.g. https://www.gbif.org/species/2435099.
      • powo/ – Plants of the World Online (POWO) identifiers
        POWO, maintained by Kew Gardens, uses unique LSIDs (Life Science Identifiers) to refer to plant taxa. These identifiers link to accepted names, synonyms, and taxonomy data. Format: URI with LSID, e.g. http://powo.science.kew.org/taxon/urn:lsid:ipni.org:names:77144975-1.
         

  5. Mosaic View Generation:

    • image_table.py generates a visual mosaic of plant thumbnails, configurable by:
      Creates a tiled layout like the following. Or 200pixel  / 300pixel
      • Thumbnail width (-p)
      • Number of columns (-c)
      • Tiled layout (-t)
      • Verbose output (-v)
      • Target directory (-d)
    • Hierachy based layouts

  1. QR Code Integration:

    • Each plant receives a deterministic URL based on the md5 checksum of its image
    • A QR code pointing to that URL is printed directly on the plant stake
    • New plant owners can scan the stake to access accurate, structured information about their specific plant

Key Components: