Processamento digital de sinais de ultrassom embarcado no Raspberry Pi
The B-mode ultrasound technique represents one of the main modalities of imaging to aid medical diagnosis. The great advantage of using ultrasound in medicine is the non-invasive technique and non-ionizing radiation, which make the method painless and safe. To improve the quality of the generated im...
| Autor principal: | Medeiros, Renan Antonio Corrêa |
|---|---|
| Formato: | Trabalho de Conclusão de Curso (Graduação) |
| Idioma: | Português |
| Publicado em: |
Universidade Tecnológica Federal do Paraná
2022
|
| Assuntos: | |
| Acesso em linha: |
http://repositorio.utfpr.edu.br/jspui/handle/1/28093 |
| Tags: |
Adicionar Tag
Sem tags, seja o primeiro a adicionar uma tag!
|
| Resumo: |
The B-mode ultrasound technique represents one of the main modalities of imaging to aid medical diagnosis. The great advantage of using ultrasound in medicine is the non-invasive technique and non-ionizing radiation, which make the method painless and safe. To improve the quality of the generated image, new approaches and techniques for digital signal processing based on hardware and software platforms are expected. This work deals with the study and implementation of ultrasound signal processing algorithms in the Raspberry Pi embedded system. The developed steps include: digital filtering, coherent summation, demodulation and envelope detection, and logarithmic compression, where the scan conversion step was performed in the Matlab software after the data has been processed in Raspberry Pi. To validate the implemented algorithm, we used data sampled with frequency of 40 MHz, obtained by simulation and through the ULTRA-ORS research platform. Qualitative and quantitative analysis using the cost functions of the Normalized Root Mean Squared Error (NRMSE) and the Normalized Residual Sum of Squares (NRSS) show that the Python algorithm implemented in Raspberry Pi presents results compatible with the reference adopted in Matlab and validated in previous studies. All NRMSE results were less than 10% and NRSS results were close to zero, indicating excellent agreement with the Matlab model. |
|---|