This thesis presents the development of a system for measuring the dimensions and predicting the frequencies of accordion reeds. A device was designed to capture images and detect reed contours using machine vision, alongside a system for measuring acoustic frequencies. Based on the acquired data, various neural network models were trained, with the MLP model yielding the most accurate results. The approach enables precise, repeatable, and partially automated frequency estimation, laying the groundwork for automating the traditionally manual and expertise-intensive process of reed manufacturing.
|
