Rapid industrialization, urbanization, and growth in agricultural activities worldwide have significantly increased environmental pollution, leading to adverse effects on human health, the economy, and the environment. Thus, effective environmental monitoring is essential to achieve sustainable growth and maintain a healthy society. Artificial intelligence (AI) technologies, such as machine learning (ML), can increasingly play a crucial role in environmental monitoring by effectively assessing the water and air quality. This article discusses the role and applications of AI in environmental monitoring.
Image credit: LeoWolfert /Shutterstock
Importance of AI in Environmental Monitoring
Environmental monitoring primarily involves systematic analysis and collection of environmental data to identify potential environmental problems, develop solutions for the identified problems, and track changes in the environment over time. The monitoring can be performed at both local and global scales to monitor soil composition, water contamination, air quality, and noise levels. Specific assessment techniques and tools are used to determine the soil/water/air quality and the impact of human activities on the environment.
Climate change has significantly increased the importance of environmental monitoring to respond more effectively to sudden, severe climate events. The application of AI can transform environmental monitoring as the technology can detect and analyze subtle changes in the environment or soil/water/air quality in real time and provide more timely and accurate data. This data can be used to make more informed decisions on environmental management, leading to greater environmental management efficiency compared to traditional environmental management practices. AI-based environmental monitoring can significantly reduce the cost and time required while using conventional monitoring methods.
Potential environmental threats, such as hazardous waste, water contamination, and air pollution, can be effectively detected by AI-powered environmental monitoring systems, such as systems comprising ML forecasting algorithms and Internet of Things (IoT) monitoring devices, to reduce risks of environmental damage.
These systems can provide more accurate climate change predictions by analyzing data from weather stations and satellites, which can be used to make informed decisions on climate change adaptation and mitigation strategies. Currently, specialized observational tools, sensors, and AI, specifically ML algorithms, are paired with IoT devices for environmental monitoring.
Data captured using IoT tools/device sensors from different environmental conditions can be integrated using a wireless sensor network (WSN), a cloud-based environmental system, to enable the devices to characterize, record, analyze, and monitor elements in a specific environment.
AI Techniques in Environmental Monitoring
Smart water pollution monitoring (SWPM) systems and smart air quality monitoring (SAQM) systems use AI and sensors to analyze and collect water and air quality data, respectively. The data can be utilized to track and identify water and air pollution and implement actions to mitigate/prevent the pollution.
SWPM systems can be employed to monitor water quality in streams, lakes, and rivers, while SAQM systems can be used to monitor air quality in rural areas, towns, and cities. Real-time and more accurate data, cost-effectiveness, and wider coverage are the major advantages of these systems over traditional alternatives.
Several AI techniques can be used in IoT-based SWPM systems to assess water quality. For instance, linear regression (LR), stochastic gradient descent (SGD), and ridge regression can be utilized to control agricultural water pollution using remote sensing.
Similarly, support vector machine (SVM) and color layout descriptors can be used for water contamination assessments. Pollutants in water can be investigated using a model based on a dolphin swarm algorithm and extreme learning machine (DSA-ELM).
Water contamination analysis and surveillance can be performed with high accuracy using SVM. The fast learning technique, decision tree (DT), k-nearest neighbors (KNN), and SVM can be employed for drinking water analysis. Moreover, the concentration of chlorophyll-A in lake water can be quantified using SVM and backpropagation neural network (BPNN).
In IoT-based SAQM systems, ML is used extensively for air quality monitoring, organic compound detection, and estimation of particulate matter 10 (PM10), PM2.5, sulfur dioxide, nitrogen oxides, ozone, lead, carbon monoxide, and benzene. Similarly, ML-based predictive models and extreme machine learning can be used for air quality characterization and monitoring based on estimating PM2.5 concentration levels, respectively. Additionally, forecasting models and deep learning can be utilized for monitoring urban air pollution based on ozone, nitrogen dioxide, and sulfur dioxide concentrations and detecting unusual ozone levels, respectively.
Other AI techniques used for air quality monitoring include multiple-layered perceptrons (MLP), 3-layered feed-forward neural network (FFNN), general regression neural network (GRNN), radial basis function network (RBF), and Elman recurrent network (EN).
Benefits and Challenges of AI in Environmental Monitoring
Benefits: AI-driven environmental monitoring systems are a cost-effective and efficient method to protect and monitor the environment. These systems can assist researchers in identifying regions at risk of experiencing severe weather events, such as floods and droughts, by collecting data on climate-related factors, including precipitation and temperature.
Additionally, these systems can identify areas at risk of experiencing environmental degradation due to human activities by collecting data on water and air pollution and deforestation.
AI-based monitoring systems can be utilized to analyze substantial amounts of data to identify trends and patterns in environmental regulation compliance to develop more effective environmental policies and regulations.
Moreover, future environmental trends can also be predicted, and strategies can be developed to mitigate their impacts using AI-powered systems, enabling organizations and governments to better enforce and monitor environmental regulations.
Challenges: The high cost of AI technology implementation is one of the key challenges of using AI for environmental monitoring. AI systems require substantial investments in software, hardware, and personnel, which can be prohibitively expensive for several organizations.
AI systems must be updated regularly to ensure their reliability and accuracy, which can increase their overall costs. The lack of data required for AI systems is another significant challenge, as these systems depend on large datasets to analyze environmental conditions accurately.
However, such datasets are typically unavailable or incomplete, which hinders the effectiveness of AI-powered environmental monitoring systems. Moreover, the efficiency of AI systems is also influenced by the reliability of the AI algorithms.
Improperly designed and untested algorithms can lead to unreliable/inaccurate insights, resulting in significant financial losses and environmental damage. Thus, AI algorithms must be designed and tested thoroughly to ensure accurate predictions of environmental conditions.
Recent Developments
In a paper recently published in the journal Environmental Impact Assessment Review, researchers proposed an AI-assisted semantic IoT (AISIoT) using a WSN for the environmental monitoring system and the real economy. Researchers used the proposed architecture to remotely measure parameters like gas leakage, fire, rainfall, and sunlight. The data can be stored online to evaluate and forecast climate patterns.
The study described environmental monitoring and quality analysis systems based on a combination of IoT and AI methods with WSN models. Researchers provided a mathematical framework for analyzing the WSN protocol's interdependent aspects for communication and design of signal processing.
Sensor systems were used to identify parameters, such as noise and carbon monoxide levels, while collecting data, calculating, and regulating actions, such as noise variations. The smart environmental monitoring system monitors and regulates environmental change-induced impacts on humans, plants, and animals.
Additionally, the IoT-based framework comprised the complete information system about the environment from the sensor level to data management. Experimental results demonstrated that the environmental monitoring system can precisely collect data from a set time interval and calculate the acquisition phase location accurately.
Moreover, the results also demonstrated that the calculated parameter variability was distributed at a proper rate in the experiment through multiple network nodes. Thus, the transmission level obtained was extremely high and sufficient for the monitoring task.
The proposed AI-SIoT method using the WSN method demonstrated 95.6% accuracy, significantly improved performance, 93.7% efficiency, and 97.4% reliability, which was higher compared to the existing methods, which displayed the effectiveness of the AI-SIoT for the long-term analysis of environmental data.
References and Further Reading
Zhang, X., Shu, K., Rajkumar, S., Sivakumar, V. (2021). Research on the deep integration of application of artificial intelligence in environmental monitoring systems and the real economy. Environmental Impact Assessment Review, 86, 106499. https://doi.org/10.1016/j.eiar.2020.106499
Ullo, S.L., Sinha, G.R. (2020) Advances in Smart Environment Monitoring Systems Using IoT and Sensors. Sensors, 20, 3113. https://doi.org/10.3390/s20113113
Frąckiewicz, M. (2023) The Use of Artificial Intelligence in Environmental Monitoring [Online] Available at https://ts2.space/en/the-use-of-artificial-intelligence-in-environmental-monitoring/ (Accessed on 13 August 2023)
Diaz, M. (2022). Machine Learning in Environmental Monitoring [Online] Available at https://gemmo.ai/machine-learning-in-environmental-monitoring/ (Accessed on 13 August 2023)