Development of an automated system for the aesthetic inspection of motorcycle silencers, based on artificial intelligence and computer vision technologies.
Officine Meccaniche Lafranconi
Officine Meccaniche Lafranconi produce silenziatori motociclistici resistenti a temperature elevate e gas corrosivi, garantendo qualità, durata, sicurezza e conformità a elevati standard estetici richiesti dai clienti. Attualmente, il controllo estetico è manuale, esposto a variabilità legata al giudizio umano.
Aesthetic conformity is essential for silencers to meet stringent customer specifications and avoid contractual or criminal penalties. However, the current control process, which relies solely on operators’ skills, is subject to variables such as fatigue, experience and human subjectivity. This lack of objectivity reduces the overall reliability of aesthetic control and makes automation difficult. Consequently, the company is exposed to potential economic risks arising from undetected defects, as well as operational challenges related to managing high-quality standards consistently and efficiently.
Lafranconi intends to develop an innovative system for the objective assessment of silencer aesthetics, exploiting artificial vision and artificial intelligence (AI) technologies. With the support of MADE and the Politecnico di Milano, a prototype test bench for image acquisition and processing will be built.
Hardware: Test bench with cylindrical structure equipped with programmable illuminators (directional/diffused light) and three industrial cameras for image acquisition.
Rapid prototyping: 3D-printed elements for silencer support.
Software: Artificial intelligence algorithms developed in Python using standard libraries (OpenCV, SkLearn, Tensorflow) for:
– Initial image segmentation
– Visual feature extraction
– Classification of defects using machine learning techniques.
Reduced variability: Eliminate dependence on human factors in aesthetic control, ensuring consistent and repeatable evaluation.
Automation: Integrate automated visual inspection into the production process, improving productivity and accuracy.
Reliability: Demonstrate superior sensitivity and specificity in defect identification, increasing overall product quality.
Quality improvement: Increased aesthetic and functional product conformity, reducing the risk of customer complaints or penalties.
Operational efficiency: Reduction in the time and resources required for aesthetic inspection.
Technological innovation: Enhancement of internal skills and increased market competitiveness through the introduction of advanced AI and artificial vision solutions.
