Our Products

We develop and manufacture Hyper-spectral Laser Induced Fluorescence (HLIF) based LiDAR (LiDAR i.e. Light Detection and Ranging) sensors for our monitoring systems. As oil is intrinsically fluorescent and emits light in the blue-green spectral range when illuminated by UV light it is easily distinguished from other substances in the water.

Using the Hyper-spectral curve of the emitted light, we can “fingerprint” the sample, and thus receive a real-time identification of oil category and type (crude oil, hydraulic oil etc) on the spot, with no samples or lab analysis required.

Our Oil in Water Locator (OWL™) sensors detect and differentiate between various organic matter in the sea, such as hydrocarbons, Colored Dissolved Organic Matter (CDOM), etc. It is the particular wavelength we use in our lasers that define what substances we are able to detect. For oil in water, we typically use a LiDAR configured at 308 nanometer in order to be able to differentiate and classify various types of oil.

Data are presented in the OWL™MAP online presentation module. All our OWL™ sensors use Global Positioning System (GPS) to mark in OWL™MAP the exact location of a particular sample made. OWL™MAP is a map layer-based, real-time user subscriber-module which easily provides an early warning and a live situation / incident report when a sensor detects an anomaly at sea. With its sensor database and oil sample database, OWL™MAP quickly presents findings of oil in water; the position, volume (ppm) and type of oil.

OWL™MAP can be connected to additional online services, such as Automatic Identification System for ships (AIS). AIS makes it possible to display what vessels were present at a detection point in the hours before a detection was made. Further, OWL™MAP can integrate with Meteorological data, presenting wind direction and speed, wave height and direction, and most importantly for submerged oil; sea current and direction.

When correlating data and with machine learning building prediction models and other Artificial Intelligence (AI) systems it is crucial to have access to relevant and high quality sensor data.
That's what Ocean Visuals is all about. 

Enabling machine learning, data correlation, and prediction modeling

Machine learning can have several applications in the area of oil spill detection and recovery. Here's how it can be helpful:

• Oil Spill Detection: Machine learning algorithms, especially convolutional neural networks (CNNs) and other deep learning models, can analyze satellite images to identify oil spills. These models can be trained on large datasets to understand different patterns and indicators of oil spills. They can even detect minor changes that might be missed by the human eye. The benefit of using such algorithms is that they can monitor large areas in a short amount of time, provide real-time updates, and aid in immediate action. Once an observation is made, we provide the sensor system that can measure and classify the oil, in real-time.
   
• Oil Spill Characterization: Machine learning can also assist in characterizing the oil spill, such as estimating the size, shape, and spread direction of the oil spill. This information is crucial for coordinating the recovery efforts effectively.
   
• Prediction Models: Machine learning can be used to create models that predict the movement and behavior of an oil spill based on a variety of factors such as ocean currents, wind direction, temperature, and the properties of the oil itself. This predictive capability can be of great help in taking proactive measures and minimizing the environmental impact of the spill.
   
• Recovery and Cleanup: Machine learning can optimize the oil spill recovery process by predicting the most effective cleanup methods based on the nature of the spill, environmental conditions, and available resources. Machine learning algorithms can also analyze past recovery operations and suggest improvements for future operations.
   
• Wildlife Impact Assessment: Machine learning can also help assess the potential impact on wildlife and the environment. By analyzing past incidents and environmental data, machine learning models can predict which areas and species might be most affected and help authorities take preventive measures.
   
• Oil Spill Prevention: Finally, machine learning can also contribute to preventing oil spills in the first place. It can be used to monitor the condition and performance of oil transportation systems (like pipelines and tankers) and predict potential failures or leaks.

Correlating data in the context of oil spill detection and recovery can help streamline the identification process, enhance predictive models, and optimize cleanup strategies. Here's a more detailed view:

• Enhancing Detection: Correlation between different types of data such as satellite imagery, weather data, and sensor data can help identify patterns that are indicative of an oil spill. For instance, correlating the visual patterns in satellite imagery with sensor data on oil presence can help enhance the reliability of oil spill detection algorithms.
   
• Improving Predictive Models: Data correlation can also enhance the accuracy of predictive models that forecast the spread of an oil spill. By correlating variables such as wind speed, ocean currents, oil viscosity, and temperature, these models can provide more accurate predictions about the spill's spread, which can inform the emergency response.
   
• Optimizing Cleanup Efforts: By correlating data from past oil spills, including the techniques used, their effectiveness, the type of oil, the surrounding environment, and the weather conditions, machine learning models can suggest the most effective cleanup strategies for future spills. These correlations can lead to more informed and efficient decision-making.
   
• Validating Cleanup Efforts: Validation of cleanup efforts: Once the cleanup operations are initiated, hyper-spectral LIDAR sensor systems can be used to assess their effectiveness. By comparing pre- and post-cleanup data, the systems can detect changes in the oil slick's distribution and characteristics. Reduced oil thickness, diminished fluorescence emissions, or the absence of certain spectral signatures indicate successful cleanup efforts.
   
• Monitoring Residual Contamination: Even after the primary cleanup is completed, residual oil may still be present in the environment. Hyper-spectral LIDAR sensors can continue to monitor the area to detect any remaining oil patches or sheens. This ongoing monitoring helps ensure that all necessary measures are taken to mitigate the long-term impact of the spill.
   
• Environmental Impact Assessment: Correlation of data on the location and severity of oil spills with data on local ecosystems can help predict the potential environmental impact more accurately. This could include impacts on marine life, coastal regions, and even human populations.
   
• Preventive Measures and Maintenance: Correlating data from equipment monitoring with instances of spills can help identify potential signs of failure or weakness in systems that transport or store oil. This predictive maintenance can prevent some spills from occurring in the first place.

It's important to note that while correlating data can provide many insights, care must be taken to ensure that the correlations are meaningful. It's also important to verify the quality and reliability of the data used, as inaccurate data can lead to incorrect conclusions. As with all data analysis, it's essential to remember that correlation does not always imply causation.

OWL™MAP presents both lidar measurement data and camera data in an easy-to-use graphical user interface

Predicting oil spill occurrence and the trajectory of an existing oil spill using meteorological data can be a complex task, due to the variety of factors that contribute to the behavior of oil in the environment. Here is a general overview of the process:

• Data Collection: The first step is gathering and understanding the data you have. Data on oil spill incidents should include factors like the time, location, volume of the spill, type of oil, and the cause if known. Meteorological data may include wind speed and direction, precipitation, air temperature, and possibly other factors depending on the region and specifics of the data you're working with.
   
• Understanding the Relationship: Once you have the data, it's important to study the relationship between oil spills and various meteorological factors. For example, wind can influence both the likelihood of a spill (in the case of shipping accidents) and the spread of a spill once it occurs. In most cases, this would involve statistical analysis to understand the relationships between various factors.
   
• Building the Model: Based on the relationships you've identified, you can begin to build a predictive model. This might be a statistical model, a machine learning model, or some combination of the two. The model should take into account the various factors you've identified and provide predictions about either the likelihood of a spill or the trajectory of a spill given certain meteorological conditions.
   
• Validation: Once the model is built, it should be validated with independent data to ensure its accuracy. This typically involves using a portion of your data for testing the model after it has been trained.
   
• Implementation and Iteration: The model can then be used to make predictions and guide decision-making. Over time, as more data is gathered and the model's performance is assessed, it can be further refined and improved.

Keep in mind that oil spill prediction is a complex task and these models are generally probabilistic, providing a likelihood rather than a certainty. This process requires expertise in several fields, including meteorology, oceanography, data analysis, and computer science, among others. Also, the behavior of oil spills can be influenced by a range of factors beyond just meteorological conditions, including ocean currents, temperature of the water, biological factors, and the specific characteristics of the oil involved. For this reason, these models often incorporate a range of data sources beyond just meteorology.

In Norway, all aquaculture licenses are marked with their GPS spot in the OWL™MAP as points-of-interest (POIs), making it possible to warn the salmon farming community of acute oil or organics in the water: