Vehicle Speed Estimation Based on 3D ConvNets and Non-Local Blocks

Abstract— The positive impact of video and image processing in traffic monitoring, analysis, and traffic condition monitoring in various cities and urban vicinities cannot be overstated. This paper is another approach considering estimating vehicles' speed. Basically, traffic videos are collected from a stationary camera mounted on freeways. On the one hand, camera calibration with respect to accurate measure could mainly solve the geometrical equations conveniently supported directly using references. Camera calibration could yield precise measurements, but rather difficult in achieving the accurate speed. The system is also adaptable to extend into other traffic-related applications. The average errors in detected vehicle speed were ± 7 km/h, and the experiments were conducted under various resolutions and different video sequences. The accelerating advancement of computing power of commonplace computers has now enabled a broader reach in the integration of deep learning techniques in traffic surveillance video comparisons. It constitutes the core functions of traffic analysis, such as prediction of traffic flows, anomaly detection, vehicle re-identification, and vehicle tracking. Among these applications, traffic flow prediction, or vehicle speed estimation, is one of the most serious research subjects during the recent years. It can be a good solution to this problem in preventing road accidents and enhancing road engineering by almost checking the traffic demand. Vehicle speed prediction is effectively proposed through integration of state-of-the-art deep learning models and classical computer vision techniques. Some present-and-state-of-the-art efforts concerning estimation of vehicle speeds, detection of vehicles, and tracking of objects are discussed in this article. Optical flow contains the speed and direction information of pixel displacements in an image. Our final approach deals with collecting multi-scale convolutional network. This extracts parameter data of vehicles. Keywords: Speed detection in vehicles, video sequences, computer vision, background modeling, traffic monitoring, OpenCV

Accurate prediction of agricultural yield is extremely important to ensure food security and cope with the challenges created by climate change and natural disasters. Forecasting agricultural yield is a challenging task due to the complex nature of variables (fertiliser, rainfall, temperature and others) that affect agricultural production. This study employs six supervised machine learning algorithms: Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), Multi-Layer Perceptron (MLP), Recurrent Neural Network (RNN), and Convolutional Neural Network (CNN) to build a predictive model using 49 years of historical data (1973-2021) on paddy, wheat, and maize. Model performance was evaluated using Mean Squared Error (MSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Root Mean Squared Error (rMSE). Results show that DT and RF models are the most precise with MSE 1% to 5%, MAE 8% to 21%, followed by SVM and CNN. Key predictors of crop yield include area cultivated, capital expenditure, banking expansion, rainfall, temperature, and fertilizers, while irrigation and road network were less significant. The study recommends that farmers prioritize commercial farming, agricultural equipment, and timely availability of fertilizer for application. The Government of Nepal (GoN) should redirect subsidies towards agricultural mechanization, ensure timely supply of fertilizer, and expand banking services in agricultural areas.

A CNN encoder decoder LSTM model for sustainable wind power predictive analytics

Forecasting stock prices has always been a complex and challenging task due to its high volatility, noise in data, and abrupt price fluctuations caused by different factors. Traditional approaches, such as quantitative, fundamental, and technical analysis, often struggle to accurately predict stock price movements due to their difficulty in processing non-linear data. The evolution of machine learning, particularly neural network models, opened new opportunities for improving prediction accuracy by handling non-linear and multidimensional data with high effectiveness. This study focuses on evaluating and comparing three neural network architectures, in which the last is a combination of the first two - Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), and a hybrid CNN-LSTM model, for forecasting stock prices. Using ten years of historical stock data, taken for Apple Inc. (AAPL) stocks as an example, ranging from 2015 to 2025, the models were trained and tested on key performance metrics such as Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and R². Results show that while both CNN and LSTM models show good performance individually, the hybrid CNN-LSTM model can significantly outperform them, achieving the highest accuracy with the lowest error rates. By combining the strengths of CNN for feature extraction and LSTM for temporal sequence processing, the hybrid model effectively captures complex patterns in stock price movements. By iterating over the types of layers to use, we came to the conclusion, that by combining both models into one hybrid model, some layers, like one CNN layer, and one LSTM layer should be removed. This research demonstrates the potential of hybrid neural network models in financial time-series forecasting and comprehensively analyzes their advantages over non-hybrid models.

An accurate vertical wind speed (WS) data estimation is required to determine the potential for wind farm installation. In general, the vertical extrapolation of WS at different heights must consider different parameters from different locations, such as wind shear coefficient, roughness length, and atmospheric conditions. The novelty presented in this article is the introduction of two steps optimization for the Recurrent Neural Networks (RNN) model to estimate WS at different heights using measurements from lower heights. The first optimization of the RNN is performed to minimize a differentiable cost function, namely, mean squared error (MSE), using the Broyden-Fletcher-Goldfarb-Shanno algorithm. Secondly, the RNN is optimized to reduce a non-differentiable cost function using simulated annealing (RNN-SA), namely mean absolute error (MAE). Estimation of WS vertically at 50 m height is done by training RNN-SA with the actual WS data a 10–40 m heights. The estimated WS at height of 50 m and the measured WS at 10–40 heights are further used to train RNN-SA to obtain WS at 60 m height. This procedure is repeated continuously until the WS is estimated at a height of 180 m. The RNN-SA performance is compared with the standard RNN, Multilayer Perceptron (MLP), Support Vector Machine (SVM), and state of the art methods like convolutional neural networks (CNN) and long short-term memory (LSTM) networks to extrapolate the WS vertically. The estimated values are also compared with real WS dataset acquired using LiDAR and tested using four error metrics namely, mean squared error (MSE), mean absolute percentage error (MAPE), mean bias error (MBE), and coefficient of determination (). The numerical experimental results show that the MSE values between the estimated and actual WS at 180 m height for the RNN-SA, RNN, MLP, and SVM methods are found to be 2.09, 2.12, 2.37, and 2.63, respectively.

Accurate prediction of energy consumption in smart homes is essential for optimizing energy usage, reducing operational costs, and enhancing occupant comfort. Traditional statistical and machine learning models often struggle to capture the complex spatiotemporal patterns inherent in energy usage data. To address this limitation, this study proposes a hybrid Convolutional Neural Network–Long Short-Term Memory (CNN-LSTM) model for forecasting energy consumption in smart homes, combining the spatial feature extraction strengths of Convolutional Neural Networks (CNNs) with the temporal learning capabilities of Long Short-Term Memory (LSTM) networks. The model demonstrated high predictive accuracy and generalization, as reflected by the evaluation metrics: Mean Absolute Error (MAE) = 0.0325, Mean Squared Error (MSE) = 0.0031, Root Mean Squared Error (RMSE) = 0.0557, and Mean Absolute Percentage Error (MAPE) = 0.2054. However, as the model was trained on synthetic data, further validation using real-world energy consumption datasets is necessary to assess its practical applicability and robustness.

Increased industrialization, urbanization, and population rise have caused considerably poor air quality, resulting in serious health complaints and environmental damage. Precise forecasting of Air Quality Index (AQI) is crucial in order to address problems and facilitate proactive health responses. This research seeks to compare and assess the predictive capability of four models – Levenberg-Marquardt artificial neural network (LM-ANN), Bayesian regularization artificial neural network (BR-ANN), multiple linear regression (MLR), and convolutional neural network (CNN). Air quality observations from five Bhubaneswar monitoring stations, recorded between January 2020 and December 2023, were compared against the chosen models. The performance of the models was evaluated through mean squared error (MSE), root mean squared error (RMSE), mean absolute error (MAE), and the coefficient of determination (R2 ) and t-statistics analysis. The CNN model performed the best with a predictive accuracy of MSE 42.81, RMSE 6.54, MAE 2.93, and R 2 0.97. The t-statistic test also proved that CNN had the best proximity with actual values (t=0.212) and BR-ANN had the maximum deviation (t=0.983). The research validates that CNN performs better than conventional models in AQI forecasting because it can retain spatial information and learn hierarchical features.

Globally, wind energy if properly harnessed, could serve as a source of energy generation in Africa. This study compared the performance of two Machine Learning (ML) algorithms (Linear regression and Random Forest) in predicting wind speed in five major cities in Africa (Yaoundé, Pretoria, Nairobi, Cairo and Abuja). Wind data were collected between January 1, 2000, and December 31, 2022, using the Solar Radiation Data Archive. The data preprocessing was carried out with 80% of the data used for training and 20% for validation. The performance of these ML algorithms was evaluated using Mean Square Error (MSE), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) and coefficient of determination (R2). The result shows that Nairobi (3.814795 m/s) closely followed by Cairo (3.606453 m/s) has the highest mean wind speed while Yaoundé (1.090512 m/s) has the lowest. Based on the performance metrics used, the two Machine Learning algorithms were competitive. Still, the Linear Regression (LR) algorithm outperformed the Random Forest Algorithm in predicting wind speed in all the selected major African cities. In Yaoundé (RMSE = 0.3892, MAE= 0.3001, MAPE =0.5030), Pretoria (RMSE=1.2339, MAE=0.9480, MAPE=0.7450) Nairobi (RMSE= 0.4223, MAE =0.6499, MAPE =0.1872), Nairobi (RMSE=0.6499, MAE=0.5171, MAPE =0.1872), Cairo (RMSE =1.0909, MAE =0.8544, MAPE =0.3541) and Abuja (RMSE = 0.70245, MAE =0.5441, MAPE= 0.4515) the Linear regression algorithms was found to outperformed Random Forest Regression. Therefore, the Linear regression algorithm is more reliable in predicting wind speed compared with the Random Forest regression.

This article uses machine learning techniques as fuzzy and neuro-fuzzy ANFISs, to develop and compare prediction models capable of relating pregnant women’s exposure to air pollutants, such as Nitrogen Dioxide and Particulate Matter, the mother’s age, and the number of prenatal consultations to the incidence of premature birth. In the current literature, studies can be found that relate prematurity to the exposure of pregnant women to NO2, O3, and PM10; to Toluene and benzene, mainly in the window 5 to 10 days before birth; and to PM10 in the week before birth. Both models used logistic regression to quantify the effects of pollutants as a result of premature birth. Datasets from Brazil—Departamento de Informatica do Sistema Único de Saúde (DATASUS) and Companhia Ambiental do Estado de São Paulo (CETESB)—were used, covering the period from 2016 to 2018 and comprising women living in the city of São José dos Campos (SP), Brazil. In order to evaluate and compare the different techniques used, evaluation metrics were calculated, such as correlation (r), coefficient of determination (R2), Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), Mean Square Error (MSE), and Mean Absolute Error (MAE). These metrics are widely used in the literature due to their ability to evaluate the robustness and efficiency of prediction models. For the RMSE, MAPE, MSE, and MAE metrics, lower values indicate that prediction errors are smaller, demonstrating better model accuracy and confidence. In the case of (r) and R2, a positive and strong result indicates alignment and better performance between the real and predicted data. The neuro-fuzzy ANFIS model showed superior performance, with a correlation (r) of 0.59, R2 = 0.35, RMSE = 2.83, MAPE = 5.35%, MSE = 8.00, and MAE = 1.70, while the fuzzy model returned results of r = 0.20, R2 = 0.04, RMSE = 3.29, MSE = 10.81, MAPE = 6.67%, and MAE = 2.01. Therefore, the results from the ANFIS neuro-fuzzy system indicate greater prediction capacity and precision in relation to the fuzzy system. This superiority can be explained by integration with neural networks, allowing data learning and, consequently, more efficient modeling. In addition, the findings obtained in this study have potential for the formulation of public health policies aimed at reducing the number of premature births and promoting improvements in maternal and neonatal health.

Accurate electrical load forecasting is of great significance to help power companies in better scheduling and efficient management. Since high levels of uncertainties exist in the load time series, it is a challenging task to make accurate short-term load forecast (STLF). In recent years, deep learning approaches provide better performance to predict electrical load in real world cases. The convolutional neural network (CNN) can extract the local trend and capture the same pattern, and the long short-term memory (LSTM) is proposed to learn the relationship in time steps. In this paper, a new deep neural network framework that integrates the hidden feature of the CNN model and the LSTM model is proposed to improve the forecasting accuracy. The proposed model was tested in a real-world case, and detailed experiments were conducted to validate its practicality and stability. The forecasting performance of the proposed model was compared with the LSTM model and the CNN model. The Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE) were used as the evaluation indexes. The experimental results demonstrate that the proposed model can achieve better and stable performance in STLF.

Three dimensional tracking of rigid objects in motion using 2D optical flows

The local climate of cities is changing, and one of the primary reasons for this change is rapid urbanization. The Lahore district is situated in the Punjab province of Pakistan and is mainly comprised of Lahore city. This city is among the fastest expanding cities in Pakistan. Due to this rapid urbanization, the natural land surfaces are being altered, harming the local environment and thus causing the urban heat island (UHI) effect. For the analysis of the UHI effect, the fundamental and essential step is assessing the land surface temperature (LST). Therefore, the current investigation assessed LST to evaluate the UHI effect of the Lahore district. This study used the remote sensing data retrieved from the Advanced Spaceborne Thermal Emission and Reflection Radiometer Global Digital Elevation Model (ASTER GDEM) and Moderate-Resolution Imaging Spectroradiometer (MODIS) sensor. Different new generation algorithms were initially used, but a convolutional neural network (CNN) model was used based on the accuracy. The model was developed by utilizing the past 19 years’ LST values along with elevation, road density (RD), and enhanced vegetation index (EVI) as input parameters for analyzing and predicting the LST. The LST data of the year 2020 was used for the validation of the outcomes of the CNN model. Among the model predicted LST and observed LST, a high correlation was noticed. The mean absolute percentage error (MAPE), mean absolute error (MAE), and mean squared error (MSE) for the considered two different periods (January and May) were also computed for both the training and validation processes. The prediction error for most parts of the district was within 0.1 K of the observed values. Hence, the formulated CNN model can be utilized as an essential tool for analyzing and predicting LST and thus for the evaluation of the UHI effect at any location.

We propose an open-ended multimodal video question answering (VideoQA) method that predicts textual answers by referring to multimodal information derived from videos. Most current open-ended VideoQA methods focus on motion and appearance features from videos and ignore the audio features that are useful for understanding video content in more detail. A few prior works that use motion, appearance, and audio features showed poor results on public benchmarks since they failed to (e.g., region or grid-level) multimodal features effectively fuse the features with details for video reasoning. We overcame these limitations with multi-stream 3-dimensional convolutional networks (3D ConvNets) and a transformer-based modulator for VideoQA. Our network represents detailed motion and appearance features as well as an audio feature on multiple 3D ConvNets and modulates each intermediate representation with question information to extract their relevant spatiotemporal features over the frames. Based on the question content, our network fuses the multimodal information of 3D ConvNets and predicts the final answers. Our VideoQA method, which effectively combined multimodal data yields, outperformed both a previous multimodal VideoQA method and a state-of-the-art method on standard benchmarks. Visualization suggests that our method can predict the correct answers by listening to the audio information, even when the motion and appearance features are inadequate for understanding the video constant.

In order to achieve real-time prediction of the battery internal temperature via the external temperature measured, a method for predicting internal temperature of a ternary polymer lithium-ion battery pack based on convolutional neural networks (CNN) and virtual thermal sensor (VTS) was proposed in this paper. A 128-channel thermometer was used to measure the internal (64 uniformly distributed points) and external (64 uniformly distributed points) temperature of the lithium-ion battery pack during seven discharge cycles for a total of 81 376 sets of data. The external temperature measured was used as the input of CNN and the internal temperature predicted as the output of CNN. CNN compared with linear regression (LR) to verify the difference of prediction accuracy. Mean square error (MSE), mean absolute error (MAE), max-error (MAXE), and goodness of fit (R2-score) were used to evaluate the prediction accuracy. The results showed that the proposed method can accurately predict real-time temperature with the MSE as 0.047. In addition, this method does not require any knowledge of battery thermal properties, heat generation, or thermal boundary conditions.

Traffic surveillance is an important aspect of road safety these days as the number of vehicles is increasing rapidly. In this paper, a system is generated for automatic vehicle detection and speed estimation. The vehicle detection and tracking are done through optical flow and centroid tracking using frames of the low-resolution video. The vehicle speed is detected using the relation between the pixel motion and the actual distance. The dataset is taken from the system (Luvizon et al. in A video-based system for vehicle speed measurement in urban roadways. IEEE Trans Intell Transp Syst 1–12, 2016 [1]). The measured speeds have an average error of +0.63 km/h, staying inside [−6, +7] km/h in over 90% of the cases. The license plate localization is done on the vehicle detected by extracting the high-frequency information and using the morphological operations. This algorithm can be used to detect the speed of multiple vehicles in the frame. The measured speeds have a standard deviation error of 4.5 km/h, which is higher than that in the system (Luvizon et al. in A video-based system for vehicle speed measurement in urban roadways. IEEE Trans Intell Transp Syst 1–12, 2016 [1]) by 2 km/h. But our algorithm uses a low-resolution video which reduces the processing time of the system.