There is a growing trend to use a digital camera as an instrument to measure velocity instead of a regular sensor approach. This paper introduces a new proposal for estimating kinematic quantities, namely, the angle and the relative speed, from a single motion blur image using the discrete cosine transform (DCT). Motion blur is a common phenomenon present in images due to the relative movement between the camera and the objects, during sensor exposure to light. Today, this source of kinematic data is mostly dismissed. The introduced technique focuses on cases where the camera moves at a constant linear velocity while the background remains unchanged. 2250 motion blur pictures were shot for the angle experiments and 500 for the speed estimation experiments, in a light and distance controlled environment, using a belt motor slider driven at angles between 0° and 90° and 10 preset speeds. The DCT Hough and DCT Radon results were compared to discrete Fourier transform (DFT) Hough and DFT Radon algorithms for angle estimation. The mean absolute error of the DCT Radon method for direction estimation was 4.66°. In addition, the mean relative error for speed estimation of the DCT Pseudocepstrum was 5.15%. The innovative DCT frequency analysis proposals were more accurate than all competitors evaluated for the reconstruction of the point spread function that enables calculation of relative velocity and motion direction. These results demonstrate that cameras as an instrument can be used to measure velocity even using a single linear motion blur degraded image.

Automatic crop identification and monitoring is a key element in enhancing food production processes as well as diminishing the related environmental impact. Although several efficient deep learning techniques have emerged in the field of multispectral imagery analysis, the crop classification problem still needs more accurate solutions. This work introduces a competitive methodology for crop classification from multispectral satellite imagery mainly using an enhanced 2D convolutional neural network (2D-CNN) designed at a smaller-scale architecture, as well as a novel post-processing step. The proposed methodology contains four steps: image stacking, patch extraction, classification model design (based on a 2D-CNN architecture), and post-processing. First, the images are stacked to increase the number of features. Second, the input images are split into patches and fed into the 2D-CNN model. Then, the 2D-CNN model is constructed within a small-scale framework, and properly trained to recognize 10 different types of crops. Finally, a post-processing step is performed in order to reduce the classification error caused by lower-spatial-resolution images. Experiments were carried over the so-named Campo Verde database, which consists of a set of satellite images captured by Landsat and Sentinel satellites from the municipality of Campo Verde, Brazil. In contrast to the maximum accuracy values reached by remarkable works reported in the literature (amounting to an overall accuracy of about $81\%$, a $f_1$ score of $75.89\%$, and average accuracy of $73.35\%$), the proposed methodology achieves a competitive overall accuracy of $81.20\%$, a $f_1$ score of $75.89\%$, and an average accuracy of $88.72\%$ when classifying 10 different crops, while ensuring an adequate trade-off between the number of multiply-accumulate operations (MACs) and accuracy. Furthermore, given its ability to effectively classify patches from two image sequences, this methodology may result appealing for other real-world applications, such as the classification of urban materials.

Objective: This study aims to introduce and assess a novel AI-driven tool developed for the classification of orthodontic arch shapes into square, ovoid, and tapered categories. Methods: Between 2016 and 2019, we collected 450 digital dental models. Applying our inclusion and exclusion criteria, we refined our dataset to 50 models, ensuring a focused and detailed analysis. Plaster casts were digitized into 3D models with AutoScan-DS-EX. Three trained evaluators then measured mesiodistal and arch widths using MeshLab. The development of DentalArch was undertaken in two versions: the first version incorporates 18 input parameters, including mesiodistal widths (from the first molar to the first molar, totaling 14) and arch widths (1 intercanine, 2 interpremolar, and 1 intermolar, totaling 4); the second version uses only 4 parameters related to arch widths. Both versions aim to predict the arch shape. An evaluation of 28 machine learning methods through a k = 5-fold cross-validation was conducted to determine the most effective techniques. Results: In the tests, the performance evaluation of the DentalArch software in detecting arch shapes revealed that version 1, which analyzes 18 parameters, achieved an accuracy of 94.7% for the lower arch and 93% for the upper arch. The more streamlined version 2, which assesses only four parameters, also showed high precision with an accuracy of 93.0% for the lower arch and 92.7% for the upper arch. Conclusions: DentalArch provides a tool with potential use in orthodontic diagnostics, particularly in the task of arch shape classification. The software offers a less subjective and data-driven approach to arch shape determination. Moreover, the open-source nature of DentalArch ensures its global availability and encourages contributions from the orthodontic community.

Cable-driven transmissions are used widely in robotic applications. However, design variables and parameters of this kind of transmission remain under study, both analytically and experimentally. In this paper, an experimental test bench to evaluate the behavior of medium-low power pulley-cable transmissions is presented. The design of the test bench allows manipulating variables such as dimensions, external load, speed, and cable tension. The system consists mainly of a brushless direct current (DC) motor, two load cells to measure the mechanical reactive force in the motor, two dismountable pulleys, two drums, a perforated disk, and several masses that provide the load and the inertial load, and electronic modules to control the speed and position of the pulley. Special attention was paid to the calibration of the load cells, focused in compensating the effect of creep. Validation tests were carried out in order to evaluate the device design. Next, pilot experiments were performed to estimate the friction behavior in the transmission. Preliminary results suggest that the friction in the transmission is largely governed by the friction behavior of the bearings.

Although many friction models have been developed by various researchers over the years, a general model does not exist yet. In this article, we present the work that was developed to identify and propose an empirical model that represents the behavior of friction in a cable-driven transmission. We designed and built a test bench that allowed us to study the incidence of friction as evidenced by the type of cable thread, the dimensions of the pulley and the capstan and the application of an external load for pulley-cable type transmissions in a variety of configurations. The results obtained show a marked influence of the external load on friction behavior. Based on these results, we propose a friction model that involves the external load and the transmission ratio. During the validation of the model, it was found that the proposed model reproduced the friction behavior better than the LuGre model, obtaining errors up to 44% lower, mainly for high loads and low speeds.