In the process of X-ray transmission imaging, the mutual occlusion between structures will lead to the image information overlap, and the computed tomography (CT) method is often required to obtain the structure information at different depths, but with low efficiency. To address these problems, an X-ray focused on imaging algorithm based on multi-line scanning is proposed, which only requires the scene target to pass through the detection area along a straight line to extract multi-view information, and uses the optical field reconstruction theory to achieve the de-obscured reconstruction of the structure at a specified depth with high real-time. The results of multi-line scan and X-ray reconstruction of the target show that the proposed method can reconstruct the information of any specified depth layer, and it can perform fast imaging detection of the mutually occluded target structures and improve the recognition of the occluded targets, which has a good application prospect.

The range migration algorithm (RMA) is an accurate imaging method for processing synthetic aperture radar (SAR) signals. However, this algorithm requires a big amount of computation when performing Stolt mapping. In high squint and wide beamwidth imaging, this operation also requires big memory size to store the result spectrum after Stolt mapping because the spectrum will be significantly expanded. A modified Stolt mapping that does not expand the signal spectrum while still maintains the processing accuracy is proposed in this paper to improve the efficiency of the RMA when processing frequency modulated continuous wave (FMCW) SAR signals. The modified RMA has roughly the same computational load and required the same memory size as the range Doppler algorithm (RDA) when processing FMCW SAR data. In extreme cases when the original spectrum is significantly modified by the Stolt mapping, the modified RMA achieves better focusing quality than the traditional RMA. Simulation and real data is used to verify the performance of the proposed RMA.

In order to strengthen the study of soil-landscape relationships in mountain areas, a digital soil mapping approach based on fuzzy set theory was applied. Initially, soil properties were estimated with the regression kriging (RK) method, combining soil data and auxiliary information derived from a digital elevation model (DEM) and satellite images. Subsequently, the grouping of soil properties in raster format was performed with the fuzzy c-means (FCM) algorithm, whose final product resulted in a fuzzy soil class variation model at a semi-detailed scale. The validation of the model showed an overall reliability of 88% and a Kappa index of 84%, which shows the usefulness of fuzzy clustering in the evaluation of soil-landscape relationships and in the correlation with soil taxonomic categories.

The purpose of this paper is to explore the performance of ridge regression and the random forest model improved by genetic algorithm in predicting the Boston house price data set and conduct a comparative analysis. To achieve it, the data is divided into training set and test set according to the ratio of 70-30. The RidgeCV library is used to select the best regularization parameter for the Ridge regression model, and for the random forest model, the genetic algorithm is used to optimize the model's hyperparameters. The result shows that compared with ridge regression, the random forest model improved by genetic algorithm can perform better in the regression problem of Boston house prices.