Seeing farther: A new camera-based technique detects distant vehicles for safer roads

Researchers develop a simple, accurate method to detect distant vehicles by analyzing nearby motion, providing a new tool to reduce intersection accidents

Shibaura Institute of Technology

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

 

image: 

The proposed system identifies the far-road region by analyzing the motion of nearby vehicles. This region is then enlarged, allowing the system to detect distant vehicles and improve intersection safety for both drivers and pedestrians.

view more 

Credit: Professor Chinthaka Premachandra from Shibaura Institute of Technology, Japan

Road accidents often stem from failing to notice vehicles. A recent study introduced a method for detecting distant vehicles with over twice the accuracy of existing systems. Instead of using road markings, it analyzes the motion of nearby vehicles to estimate the road’s trajectory and vanishing point, capturing distant road areas. By enlarging distant regions, the system can more accurately detect faraway vehicles, thereby enhancing safety by providing distant-vehicle information to drivers and pedestrians.

Intersections are among the most unpredictable spots on city roads. Drivers may struggle to notice approaching vehicles, and pedestrians often misjudge when it is safe to cross. In Japan, nearly half of all road accidents occur at intersections, highlighting an urgent need for smarter systems that can improve visibility and safety.

Now, a team of researchers led by Professor Chinthaka Premachandra from the Advanced Electronic Engineering Course, the School of Engineering and the Graduate School of Engineering and Science at the Shibaura Institute of Technology (SIT), Tokyo, Japan, and Eigo Ito from the Department of Electronic Engineering, School of Engineering, SIT, Japan, has developed a new technique for accurately identifying distant vehicles. This research was published in Volume 6 of the journal IEEE Open Journal of Intelligent Transportation Systems, on September 26, 2025.

Most current vehicle detection methods depend on deep learning algorithms that require powerful computing systems. However, these methods often struggle when vehicles appear small or unclear in the distance. The newly developed method takes a simpler, more practical approach—it examines the movement of nearby vehicles to predict where the road extends far ahead.

“Most detection systems work well for vehicles close to the camera, but their accuracy drops sharply for those farther away,” explains Prof. Premachandra. “We wanted to overcome this limitation with a lightweight approach that doesn’t require large datasets or complex neural networks.”

The system analyzes continuous video footage of the road and tracks the motion of nearby vehicles from frame to frame. By mapping these movements—called trajectories—it estimates the road’s vanishing point, or the area where the road appears to narrow toward the horizon. Once this region is identified, the system digitally enlarges it, making distant vehicles more visible.

Next, the system uses a mathematical model known as a Gaussian Mixture Model to detect moving vehicles in the magnified image. This process helps identify even tiny, faraway vehicles that might otherwise go unnoticed—all while using a standard camera and modest computing power.

In tests conducted under both day and night conditions, the method achieved more than twice the detection accuracy of conventional systems. It even outperformed some deep learning-based techniques while running smoothly at 30 frames per second on small, low-cost devices, such as the Raspberry Pi and Jetson Nano.

“Detecting faraway vehicles earlier can significantly reduce intersection-related accidents,” says Prof. Premachandra. “Imagine a system that warns pedestrians before they step onto the crosswalk or alerts drivers to vehicles still hundreds of meters away—it could make daily commuting much safer.”

Prof. Premachandra believes that this innovation could form the foundation of next-generation guidance systems that communicate real-time traffic information to both drivers and pedestrians. For example, cameras installed at intersections could detect vehicles approaching from a distance in all directions and issue alerts to drivers and pedestrians near the intersection via connected devices or traffic signals.

Beyond safety, the technology could contribute to the development of intelligent transportation systems (ITS)—which are networks that integrate vehicles, sensors, and infrastructure to create safer and more efficient roads. Because the system operates without heavy computational resources, it can be easily deployed in urban and rural areas alike.

While the initial results are promising, Prof. Premachandra notes that further improvements are planned. The team aims to test the system under different weather conditions, such as fog, rain, and snow, since these conditions often obscure visibility. They also intend to expand the system’s capability to classify different vehicle types and integrate it into broader ITS frameworks.

“Our goal is to make roads safer for everyone,” says Prof. Premachandra. “Even a few extra seconds of early warning can make the difference between a safe journey and a serious accident.”

By combining smart observation with simple computation, this study demonstrates that innovation in traffic safety doesn’t always require complex artificial intelligence—sometimes, it just takes a sharper look at the road ahead.

 

A busy road intersection where vehicles approach from multiple directions. The new camera-based method helps detect distant vehicles earlier, reducing the risk of accidents for both drivers and pedestrians.

 

Credit

University Ave., Berkeley from Openverse Source Link: https://openverse.org/image/ffb34e0d-a4b7-429c-b148-717449d257bb?q=road+intersection+traffic&p=11

About Shibaura Institute of Technology (SIT), Japan

Shibaura Institute of Technology (SIT) is a private university with campuses in Tokyo and Saitama. Since the establishment of its predecessor, Tokyo Higher School of Industry and Commerce, in 1927, it has maintained “learning through practice” as its philosophy in the education of engineers. SIT was the only private science and engineering university selected for the Top Global University Project sponsored by the Ministry of Education, Culture, Sports, Science and Technology and had received support from the ministry for 10 years starting from the 2014 academic year. Its motto, “Nurturing engineers who learn from society and contribute to society,” reflects its mission of fostering scientists and engineers who can contribute to the sustainable growth of the world by exposing their over 9,500 students to culturally diverse environments, where they learn to cope, collaborate, and relate with fellow students from around the world.

Website: https://www.shibaura-it.ac.jp/en/

About Professor Chinthaka Premachandra from Shibaura Institute of Technology, Japan

Dr. Chinthaka Premachandra (Senior Member, IEEE) earned his BSc and MSc degrees from Mie University and his PhD in Engineering from Nagoya University, Japan. He is currently a Professor at the Shibaura Institute of Technology, Japan, where he heads the Image Processing & Robotics Laboratory. Over more than 15 years of academic experience, he has published extensively on AI, UAVs, image/audio processing, intelligent transport systems, and mobile robotics. His honors include the IEEE Sensors Letters Best Paper Award, the IEEE Japan Medal, and awards from IEICE/IPSJ. He also serves as Associate Editor for major IEEE journals.

---

Journal

IEEE Open Journal of Intelligent Transportation Systems

DOI

10.1109/OJITS.2025.3614862 

Method of Research

Computational simulation/modeling

Subject of Research

Not applicable

Article Title

Recognizing Distant Vehicles on GMM by Extracting Far Road Area Based on Analyzing Trajectories of Nearby Vehicles

£1.4 million grant for groundbreaking University of Stirling salmon health study

A major new University of Stirling led study aiming to address one of the biggest issues in UK aquaculture has been awarded more than £1.4 million in funding

University of Stirling

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

 

image: 

Dr Rose Ruiz Daniels of the University of Stirling.

view more 

Credit: University of Stirling

Scottish salmon farming generates around £750 million in exports annually, despite smolt (young fish) mortality rates of 15% to 20%, with gill and skin conditions being major contributors.

Dr Rose Ruiz Daniels, a Lecturer in Aquaculture Genomics, has now secured more than £1.4 million in funding from the UKRI Biotechnology and Biological Sciences Research Council (BBSRC) that she believes can transform understanding of salmon health and disease resilience.

The project will explore tissue remodelling processes in salmon, aiming to reduce mortality linked to gill and skin health issues - major ongoing challenges affecting the global aquaculture industry.

The research, hosted at the University’s world-renowned Institute of Aquaculture, also benefits from £120,000 of in-kind support from Benchmark Genetics - a global leader in aquaculture innovation.

Researchers will study salmon during smoltification - a key stage when young fish adapt from freshwater to seawater. This critical process involves major changes in the body, making it valuable for understanding how fish both repair and strengthen their tissues.

Dr Ruiz Daniels explained: “When smoltification fails to proceed normally, the fish become more vulnerable to stress and disease. By examining smoltification as a biological remodelling event, we can identify how salmon repair tissues, resist disease, and adapt to changing environments.

“The findings will help inform improved breeding and health management strategies that enhance resilience across the industry.”

The study has three core objectives:

• To develop phenotyping tools, tools that look at the visible traits or characteristics of a salmon and measure how effectively fish can repair and rebuild their body tissues during smoltification.
• To determine whether this healing capacity has a genetic basis and evaluate its potential to support better-informed future breeding strategies.
• To discover the key biological processes inside salmon cells that enable tissue repair, and link these processes to the fish’s ability to both heal and maintain long-term health.

By identifying key genes and biological processes involved in successful smoltification, the research team aims to deliver practical tools that support fish health and productivity, whilst also improving the sustainability of this vital contributor to the Scottish economy.

Dr Ruiz Daniels added: “This work will help transform how we understand salmon biology. Recognising remodelling as a selectable trait will support breeding strategies that enhance survival and welfare across aquaculture.”

The study builds on existing data and continues the University’s long-term collaboration with Benchmark Genetics and other aquaculture partners.

Andrew Preston, Lead Trait Development & Land Based, Benchmark Genetics, said: "Developing new health traits that complement existing gill health traits marks an important step toward improving salmon welfare at critical stages of production, including during smoltification.

“By broadening our understanding of the biological processes behind cell repair, our goal is to harness this knowledge to enhance robustness in salmon farming, supporting healthier fish at all stages during production."

The announcement comes with work on the University’s state-of-the-art National Aquaculture Technology and Innovation Hub (NATIH) nearing completion.

Funded by a £17million investment by the UK Government through the Stirling and Clackmannanshire City Region Deal, as well as a £1m Wolfson Foundation grant, NATIH will drive the UK’s ambition to be a world leader in modern aquaculture practice.

Machine learning offers growers a new tool for predicting crop water use

The Hebrew University of Jerusalem

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

A new study shows that machine-learning models can accurately predict daily crop transpiration using direct plant measurements and environmental data. By training models on seven years of high-resolution lysimeter data, the research demonstrates strong performance across tomatoes, wheat, and barley. The findings point toward future tools that may support both irrigation management and early detection of plant stress.

When it comes to irrigation, the difference between “just enough” and “too much” water can make or break a season. A new study from the Hebrew University of Jerusalem sheds light on a promising direction: a machine-learning method that predicts plant water use each day, using high-resolution data that captures how healthy plants naturally behave.

The research, jointly led by first authors Shani Friedman and Nir Averbuch under the supervision of Prof. Menachem Moshelion, brings together seven years of continuous monitoring from tomato, wheat, and barley plants grown in semi-commercial greenhouses. Using a high-precision load-cell lysimeter system—technology that records subtle changes in plant weight—the team generated highly accurate measurements of daily transpiration, the evaporation of water through leaves that reflects the plant’s water use.

By feeding these measurements into models such as Random Forest and XGBoost, the study showed that machine learning can reliably predict daily transpiration from environmental conditions and plant characteristics. In independent tests, the XGBoost model reached an R² of 0.82, closely matching measured transpiration even under differing climate conditions and in outside facilities. While the models currently rely on lysimeter-based weight data—technology that growers do not typically use in the field—they highlight an important conceptual step toward plant-driven prediction tools.

Two factors stood out as especially important: plant biomass and daily temperature. “These variables consistently shaped how much water plants consumed,” said Friedman. “Understanding how a healthy, well-irrigated plant is expected to behave on a given day also allows us to detect when something is off.”

Because the model predicts what a healthy plant should be doing, unexpected changes in transpiration may serve as early warning signs of stress, whether caused by drought, salinity, disease, root damage, or other environmental pressures. “If a plant behaves differently than the model predicts, that deviation can be an indicator of abnormal or unhealthy plant behavior,” Friedman added.

Averbuch, whose work focuses on precision irrigation, emphasized the long-term potential. “Today, many irrigation decisions still rely on indirect estimates,” he explained. “Although this model is not yet field-ready, the findings show how future systems could incorporate physiological predictions to support more accurate irrigation scheduling.”

The study comes at a time of rising interest in data-driven agriculture, especially as growers face increasing pressure from drought, heat waves, and fluctuating weather patterns. While the approach is not yet a practical farm-deployable solution, it offers a glimpse into how machine learning, environmental sensing, and plant physiology may eventually combine into tools that support both irrigation management and stress diagnostics.

Importantly, the model performed well when tested on plants grown in a different research greenhouse at Tel Aviv University, suggesting the approach could adapt across climates and production setups.

For growers, the message is clear: machine learning is becoming more than a buzzword. In the near future, predictive models based on real plant behavior may help identify stress earlier, support better water-use decisions, and improve crop resilience.

---

Journal

Plant Cell & Environment

DOI

10.1111/pce.70222 

Method of Research

Observational study

Subject of Research

Not applicable

Article Title

Integrating Load‐Cell Lysimetry and Machine Learning for Prediction of Daily Plant Transpiration

Surfing on the waves of the microcosm

Physics: Publication in Nature Communications

Heinrich-Heine University Duesseldorf

FacebookXLinkedInWeChatBlueskyMessageWhatsAppEmail

 

image: 

A particle (red sphere) is guided from left to its destination (right) using a laser trap (double-cone) by means of a protocol developed in the study, which is described by the parameter λ. A known time-dependent external force field F (t) acts on this environment. The optimised protocol exploits this force field in a way that extracts the maximum amount of work. This can be applied to various external fields, to active particles and to micro-robot transport problems. (Fig.: HHU/Kristian S. Olsen)

view more 

Credit: HHU/Kristian S. Olsen

Conditions can get rough in the micro- and nanoworld. To ensure that e.g. nutrients can still be optimally transported within cells, the minuscule transporters involved need to respond to the fluctuating environment. Physicists at Heinrich Heine University Düsseldorf (HHU) and Tel Aviv University in Israel have used model calculations to examine how this can succeed. They have now published their results – which could also be relevant for future microscopic machines – in the scientific journal Nature Communications.

When planning an ocean crossing, sailors seek a course, which makes optimum use of favourable wind and ocean currents, and manoeuvre in order to save time and energy. They also react to random fluctuations in wind and currents, and take advantage of fair winds and waves. Such considerations with regard to energy costs are also important for transport processes at the micro- and nanoscale. For example, molecular motors should use as little energy as possible when transporting nutrients from A to B between and within biological cells.

However, the conditions are much rougher in the highly dynamic environment of a living organism and the fluctuations to which the microtransporters need to respond are significantly larger. Large deterministic forces such as the periodicity of the heartbeat can however be harvested to realise optimum movement strategies; particles can surf on the waves of the microcosm, so to speak.

A German-Israeli team of physicists headed by Professor Dr Hartmut Löwen from the Institute for Theoretical Physics II at HHU and Professor Dr Yael Roichman from Tel Aviv University have now examined how to minimise the amount of work required to guide a particle to a specified destination within a specified time in a microscopic environment.

Professor Löwen, senior author of a study, which has now been published in Nature Communications: “In the best case scenario, this control problem can even be used to extract work, i.e. the fluctuations and external time-dependent forces are cleverly used to optimise the energy costs of the transport.”

Such nanomachines, which extract energy from fluctuations, are of great interest in the nanosciences and biology, while the underlying question is of fundamental physical importance as it relates to central aspects of thermodynamics. Dr Kristian Stølevik Olsen, Humboldt postdoctoral fellow at HHU and lead author: “The second law of thermodynamics defines how heat is converted into work in the macroscopic world. In the microscopic world, however, things can look very different and therefore cannot be described properly by the macroscopic theory.”

The authors investigated this control problem using model calculations in which colloidal particles – nano- to micrometre-sized particles in a medium – are transported using “optical tweezers” – structures, which can be used to manipulate microscopic objects by means of light. Olsen: “We have identified the maximum amount of work that can be extracted from such an optically driven non-equilibrium system. This is, so to speak, a generalisation of the second law of thermodynamics under the given constraints for very small fluctuating systems.”

“Given a known external force field, we developed an optimised protocol for guiding such a colloidal particle with optical tweezers in order to extract maximum work. This allows external forces to be cleverly used to perform work exactly when it is needed,” says Löwen. Olsen adds: “We need to know the acting external forces in advance, but our results are stable against small inaccuracies and are therefore practically relevant.”

While HHU was primarily responsible for completing the theoretical calculations, the authors from Tel Aviv University also considered application perspectives. Co-author Dr Rémi Goerlich, postdoc in Tel Aviv: “The activities we examined occur in precisely the same way in microscopic biological processes within cells. Learning optimal solutions helps understand the energetics of natural micro-systems, potentially enabling their use for synthetic systems.”

Professor Roichman: “In our laboratory, we can in principle confirm these statements on colloids in laser traps. The theory thus forms the basis for future nanomachines, which could for example be used to transport medication to the specific locations in the body where it is needed.”

Original publication

K. S. Olsen, R. Goerlich, Y. Roichman, H. Löwen, Harnessing non-equilibrium forces to optimize work extraction, Nature Communications (2025) 16:11031

DOI: 10.1038/s41467-025-67114-8

---

Journal

Nature Communications

DOI

10.1038/s41467-025-67114-8 

Article Title

Harnessing non-equilibrium forces to optimize work extraction