Crop and weed discrimination in natural field environments is still challenging for implementing automatic agricultural practices, such as weed control. Some weed control methods have been proposed. However, these methods are still restricted as they are implemented under controlled conditions. The development of a sound weed control system begins by recognizing the crop and the different weed plants presented in the field. In this work, a classification approach of Zea mays L. (Crop), narrow-leaf weeds (NLW), and broadleaf weeds (BLW) from multi-plant images are presented. Moreover, a large image dataset was generated. Images were captured in natural field conditions, in different locations, and growing stages of the plants. The extraction of regions of interest (ROI) is carried out employing connected component analysis (CCA), whereas the classification of ROIs is based on Convolutional Neural Networks (CNN) and compared with a shallow learning approach. To measure the classification performance of both methods, accuracy, precision, recall, and F1-score metrics were used. The best alternative for the weed classification task at early stages of growth and in natural corn field environments was the CNN-based approach, as indicated by the 97% accuracy value obtained.

The implementation of a machine learning algorithm for measuring refractive index of liquid samples using Fresnel reflection at the tip of a fiber is proposed in order to overcome the measurement ambiguity between samples having refractive index values below and above the effective refractive index of the fiber fundamental mode. This is the first time that a machine learning algorithm is implemented in a fiber refractometer. The algorithm, used for pattern classification, is the Support Vector Machine (SVM). The sensing head is formed by two-cascaded cavities that generate an interference pattern that changes each time the fiber is immersed in a different sample. The changes in the interference pattern are classified by the proposed algorithm, which extends the sensing range and eliminates any ambiguity in the obtained RI values. The proposed system is also self-referenced, and therefore it is unaffected by any intensity change of the optical source. A theoretical model and experimental results are presented in detail to demonstrate the effectiveness of the proposed system.

A fiber sensor for directional bending monitoring is presented. The sensor is made of a single piece of a multimode elliptical-core fiber, and works in reflection regime. The interference spectrum generated by the proposed system has a non-uniform amplitude and non-periodic shape that makes the sensing signal adequate for a pattern recognition algorithm to be implemented. The elliptical-core fiber is fixed to a flexible rod and each bending position of the rod generates a specific optical spectrum, which is then identified by a pattern recognition algorithm. Using the algorithm identification, the trajectory of the flexible rod can be calculated. Experimental results demonstrating directional bending monitoring, based on pattern recognition and estimation of the flexible rod trajectory, are presented.

Activity recognition is one of the most active areas of research in ubiquitous computing. In particular, gait activity recognition is useful to identify various risk factors in people’s health that are directly related to their physical activity. One of the issues in activity recognition, and gait in particular, is that often datasets are unbalanced (i.e., the distribution of classes is not uniform), and due to this disparity, the models tend to categorize into the class with more instances. In the present study, two methods for classifying gait activities using accelerometer and gyroscope data from a large-scale public dataset were evaluated and compared. The gait activities in this dataset are: (i) going down an incline, (ii) going up an incline, (iii) walking on level ground, (iv) going down stairs, and (v) going up stairs. The proposed methods are based on conventional (shallow) and deep learning techniques. In addition, data were evaluated from three data treatments: original unbalanced data, sampled data, and augmented data. The latter was based on the generation of synthetic data according to segmented gait data. The best results were obtained with classifiers built with augmented data, with F-measure results of 0.812 ($\sigma$ = 0.078) for the shallow learning approach, and of 0.927 ($\sigma$ = 0.033) for the deep learning approach. In addition, the data augmentation strategy proposed to deal with the unbalanced problem resulted in increased classification performance using both techniques.