RESUMO
The power sector is one of the most important engineering sectors, with a lot of equipment that needs to be appropriately maintained, often spread over large areas. With the recent advances in deep learning techniques, many applications can be developed that could be used to automate the power line inspection process, replacing previously manual activities. However, in addition to these novel algorithms, this approach requires specialized datasets, collections that have been properly curated and labeled with the help of experts in the field. When it comes to visual inspection processes, these data are mainly images of various types. This paper consists of two main parts. The first one presents information about datasets used in machine learning, especially deep learning. The need to create domain datasets is justified using the example of the collection of data on power infrastructure objects, and the selected repositories of different collections are compared. In addition, selected collections of digital image data are characterized in more detail. The latter part of the review also discusses the use of an original dataset containing 2630 high-resolution labeled images of power line insulators and comments on the potential applications of this collection.
RESUMO
Some plant diseases can significantly reduce harvest, but their early detection in cultivation may prevent those consequential losses. Conventional methods of diagnosing plant diseases are based on visual observation of crops, but the symptoms of various diseases may be similar. It increases the difficulty of this task even for an experienced farmer and requires detailed examination based on invasive methods conducted in laboratory settings by qualified personnel. Therefore, modern agronomy requires the development of non-destructive crop diagnosis methods to accelerate the process of detecting plant infections with various pathogens. This research pathway is followed in this paper, and an approach for classifying selected Solanum lycopersicum diseases (anthracnose, bacterial speck, early blight, late blight and septoria leaf) from hyperspectral data captured on consecutive days post inoculation (DPI) is presented. The objective of that approach was to develop a technique for detecting infection in less than seven days after inoculation. The dataset used in this study included hyperspectral measurements of plants of two cultivars of S. lycopersicum: Benito and Polfast, which were infected with five different pathogens. Hyperspectral reflectance measurements were performed using a high-spectral-resolution field spectroradiometer (350-2500 nm range) and they were acquired for 63 days after inoculation, with particular emphasis put on the first 17 day-by-day measurements. Due to a significant data imbalance and low representation of measurements on some days, the collective datasets were elaborated by combining measurements from several days. The experimental results showed that machine learning techniques can offer accurate classification, and they indicated the practical utility of our approaches.
Assuntos
Solanum lycopersicum , Aprendizado de Máquina , Diagnóstico Precoce , Doenças das Plantas/microbiologia , Folhas de Planta/microbiologiaRESUMO
This work aimed to assess the recalibration and accurate characterization of commonly used smart soil-moisture sensors using computational methods. The paper describes an ensemble learning algorithm that boosts the performance of potato root moisture estimation and increases the simple moisture sensors' performance. It was prepared using several month-long everyday actual outdoor data and validated on the separated part of that dataset. To obtain conclusive results, two different potato varieties were grown on 24 separate plots on two distinct soil profiles and, besides natural precipitation, several different watering strategies were applied, and the experiment was monitored during the whole season. The acquisitions on every plot were performed using simple moisture sensors and were supplemented with reference manual gravimetric measurements and meteorological data. Next, a group of machine learning algorithms was tested to extract the information from this measurements dataset. The study showed the possibility of decreasing the median moisture estimation error from 2.035% for the baseline model to 0.808%, which was achieved using the Extra Trees algorithm.
Assuntos
Solo , Água , Aprendizado de Máquina , Meteorologia , Água/análiseRESUMO
The datasets provide hyperspectral imagery of potato fields with referencing agronomic measurements of several parameters. It contains also meteorological data collected on the place at the same time and some additional data on the variety of potatoes and the experiment. The experiment has been conducted in 2020 and two different potato varieties (Lady Claire and Markies) on the different soil profiles were planted and observed. During that time, on 4 different days, to provide a detailed picture of the experiment the hyperspectral imagery has been taken using a UAV and 150-band hyperspectral camera. The collected material has been later processed into 8 georeferenced ortophotomaps. To provide precise reference information that could be later used for modeling purpose the measurements of plants from each field has been performed. The registered data contains data on plant height, number of stems, stem fresh and dry mass, leaf fresh and dry mass, leaf assimilation area and LAI, number of tubers, tuber fresh and dry mass, starch content, RWC, chlorophyll a fluorescence index, the maximum quantum yield of PSII photochemistry, and the performance of electron flux.
RESUMO
Electrical insulators are elements of power lines that require periodical diagnostics. Due to their location on the components of high-voltage power lines, their imaging can be cumbersome and time-consuming, especially under varying lighting conditions. Insulator diagnostics with the use of visual methods may require localizing insulators in the scene. Studies focused on insulator localization in the scene apply a number of methods, including: texture analysis, MRF (Markov Random Field), Gabor filters or GLCM (Gray Level Co-Occurrence Matrix) [1], [2]. Some methods, e.g. those which localize insulators based on colour analysis [3], rely on object and scene illumination, which is why the images from the dataset are taken under varying lighting conditions. The dataset may also be used to compare the effectiveness of different methods of localizing insulators in images. This article presents high-resolution images depicting a long rod electrical insulator under varying lighting conditions and against different backgrounds: crops, forest and grass. The dataset contains images with visible laser spots (generated by a device emitting light at the wavelength of 532â¯nm) and images without such spots, as well as complementary data concerning the illumination level and insulator position in the scene, the number of registered laser spots, and their coordinates in the image. The laser spots may be used to support object-localizing algorithms, while the images without spots may serve as a source of information for those algorithms which do not need spots to localize an insulator.