Maurício Fernando Lima Pereira

A large volume of data is being produced by agrometeorological stations, satellites, Unmanned Aerial Vehicles (UAV), agricultural machines, among other equipment's, all used for agricultural management in food production. Related to that, actually, there are demands for the organization of new computational techniques for agricultural risk management. Therefore, the data volume is immense, and there are challenges to execute them based on both conventional algorithms and architectures. This paper presents the organization of a new parallel model based on semantics and the Apache Spark framework for agricultural risk management related to soil quality. Based in such concept, a parallel algorithm is proposed to distribute tomographic images of soil to several nodes, as well as, to ask them for not only pore recognition, but also for soil density and soil compaction measurements from those images. Besides, the developed parallel algorithm allowed to obtain a better performance gains and reduction of the time required to generate maps for agricultural soil quality, i.e., to support decision making in such field of interest.

This paper presents the use of a multiprocessor architecture for the performance improvement of tomographic image reconstruction. Image reconstruction in computed tomography (CT) is an intensive task for single-processor systems. We investigate the filtered image reconstruction suitability based on DSPs organized for parallel processing and its comparison with the Message Passing Interface (MPI) library. The experimental results show that the speedups observed for both platforms were increased in the same direction of the image resolution. In addition, the execution time to communication time ratios (Rt/Rc) as a function of the sample size have shown a narrow variation for the DSP platform in comparison with the MPI platform, which indicates its better performance for parallel image reconstruction.

This paper presents a novel model for combining projection and image filtering in computerized tomography. First, it is used an a priori one-dimensional projection filtering, through an extended Kalman filter with joint estimation. Then, the reconstructed images, obtained filtered backprojection algorithms (including the use of Hamming windows), are filtered using the two-dimensional DWT and wavelet thresholding, a non-linear technique. Experiments considering only one filtering stage (a priori 1-D filtering or 2-D DWT image filtering) show images with significant higher noise levels and the combination showed great noise reduction. The obtained results lead to the conclusion that the proposed combining model is a valid and interesting tool for tomographic image analysis.