Design of Visual Online Spectrometer Based on Support Vector Machine

Abstract

Using the ARM and DSP dual core processor and H.264 algorithm to study the iron spectrum image classification technology, the typical oil sample image is scaled, stable and rotating, and a complete grinding grain sample library is established. By using a Vinci processor, H.264 encoding method, the Visual on-line iron spectrometer that can be used to monitor the visual line of the wear particles in the oil and meet the requirements of on-line oil monitoring is produced. The iron spectrometer is supporting vector machine and recognition accuracy are ideal.