Search for a command to run...
1. Description This dataset contains 11,593 high-resolution images capturing common physical defects in Fused Filament Fabrication (FFF) 3D printing. Unlike generic datasets, these images were collected systematically using a Raspberry Pi HQ Camera integrated into a Delta-type 3D printer environment. The data was generated through a structured Design of Experiments (DoE), varying critical process parameters (e.g., nozzle temperature, bed temperature, print speed, and cooling rates) to induce and capture various failure modes. The study utilizes both PLA and ABS materials, specifically focusing on the challenges of printing ABS in open-chamber environments. Specific geometries were considered for each of the defect types. 2. Dataset Composition The data is organized into six specialized zipped folders, representing binary classifications for three major defect types. Each category uses optimized geometries specifically selected to highlight the initiation and progression of that particular defect. Folder Name Description Image Count Warping.zip Evidence of base layer lifting/detachment 1,707 No_Warping.zip Successful bed adhesion/stable base 1,815 Stringing.zip Unwanted plastic threads during travel moves 2,555 No_Stringing.zip Clean retraction and travel moves 1,811 Cracking.zip Inter-layer delamination and structural splitting 1,663 No_Cracking.zip High structural integrity and layer bonding 2,042 Total 11,593 3. Technical Specifications Hardware: Delta 3D Printer equipped with a Raspberry Pi HQ Camera (12.3 MP Sony IMX477 sensor). Materials: Polylactic Acid (PLA) and Acrylonitrile Butadiene Styrene (ABS). Methodology: Layerwise image acquisition during active printing stages. Experimental Design: Parameters varied as per a structured Design of Experiments (DoE) to ensure statistical diversity in defect intensity and lighting conditions. 4. Potential Applications Computer Vision: Training Convolutional Neural Networks (CNNs) for real-time defect classification. Quality Assurance: Developing closed-loop feedback systems for automated print abortion. Process Optimization: Analyzing the correlation between printing parameters and physical defect initiation.