European Vehicle Data Accumulation & Editing -- 2

PROJECT BRIEFING – VEHICLE DATA COLLECTION AND IMAGE PROCESSING PROJECT GOAL We are building a structured vehicle database for an internal AI showroom simulation and market comparison system. The goal is to create a clean, consistent dataset ready for direct website use. SCOPE OF WORK 20 buses 20 tractors 5 wheel loaders 5 excavators 5 large tractor trailers All vehicles must be located in Europe, excluding Germany. Model years only 2024, 2025, 2026. No duplicates. BRANDS Buses: Mercedes-Benz, MAN, Setra, Volvo, Scania Machinery: John Deere, Fendt, Case IH, New Holland, Massey Ferguson DATA REQUIREMENTS (CRITICAL) No incomplete listings are allowed. Technical specifications are mandatory for every vehicle. Full technical descriptions must be available. If technical data is missing, the seller must be contacted directly via email or phone. Price is optional. Only complete and verified datasets are accepted. VIN / FIN REQUIREMENTS (STRICT – NON-NEGOTIABLE) The VIN/FIN must exactly match the specific vehicle. Publicly available or database-sourced VINs are NOT accepted. IF A VIN/FIN CAN BE FOUND VIA GOOGLE OR ANY PUBLIC SOURCE, THE ENTIRE DATASET IS INVALID AND WILL BE REJECTED. The VIN/FIN must be clearly visible on the type plate image. If it is not visible, the listing must be discarded. No seller contact is required for VIN/FIN collection. Every VIN/FIN must be verified via Google search. There are NO exceptions to this rule. IMAGE REQUIREMENTS (MANDATORY) Minimum 8 images per vehicle. Required images include: * front view * side view * rear view * interior (if available) * TYPE PLATE / VIN IMAGE (ABSOLUTELY MANDATORY) IF THE TYPE PLATE IMAGE IS MISSING, THE DATASET IS AUTOMATICALLY INVALID. DO NOT PROCESS OR EDIT ANY IMAGES IF THE TYPE PLATE IS NOT INCLUDED. IMAGE PROCESSING RULES All images must be processed as follows: * remove background * remove watermarks and text * add a clean showroom background * add a soft, realistic shadow (No image mirroring required.) TASK WORKFLOW (MANDATORY) 1. Check if a type plate / VIN image is included. → If NOT: discard the listing immediately. 2. Extract the VIN/FIN directly from the image. 3. Search the VIN/FIN on Google. 4. Validation: → If the VIN/FIN is found anywhere online: INVALID dataset → If the VIN/FIN is NOT found and only exists in the image: VALID dataset Only after this validation, continue with data processing. DELIVERY FORMAT CSV, Excel or JSON format Structured image folders per vehicle Consistent naming convention Dataset must be ready for direct website integration FINAL NOTE (CRITICAL) Datasets that do not meet the VIN/FIN validation rules OR do not include a valid type plate image will NOT be accepted under any circumstances. Please follow these requirements exactly. Non-compliant work will be rejected.

Project ID: 40422950