A Deep Learning Time Series Approach for Leaf and Wood Classification from Terrestrial LiDAR Point Clouds

Han, Tao; Sánchez‐Azofeifa, Arturo

doi:10.3390/rs14133157

Cited by 14 publications

(11 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Han and Sanchez-Azofeifa [ 36 ] investigated the leaf and wood terrestrial laser scanning (TLS) time series classification with Fully Convolutional Neural Network (FCN), LSTM-FCN, and Residual Network (ResNet). CNN has been found to be capable of performing time series classification satisfactorily, as its multiple filters can produce multiple discrimination features for classification from the temporal inputs [ 35 ].…”

Section: Advances In Deep Learning Methods For Time Series Modelingmentioning

confidence: 99%

“… Tree diagram for grouping the popular Deep Learning methods for sensor time series classification and forecasting tasks covered in this survey [ 30 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 ] (note: if a paper uses two methods separately with similar satisfactory results, the paper will be listed under both groups). …”

Section: Figurementioning

confidence: 99%

See 1 more Smart Citation

Continual Deep Learning for Time Series Modeling

Ao¹,

Fayek

2023

Sensors

View full text Add to dashboard Cite

The multi-layer structures of Deep Learning facilitate the processing of higher-level abstractions from data, thus leading to improved generalization and widespread applications in diverse domains with various types of data. Each domain and data type presents its own set of challenges. Real-world time series data may have a non-stationary data distribution that may lead to Deep Learning models facing the problem of catastrophic forgetting, with the abrupt loss of previously learned knowledge. Continual learning is a paradigm of machine learning to handle situations when the stationarity of the datasets may no longer be true or required. This paper presents a systematic review of the recent Deep Learning applications of sensor time series, the need for advanced preprocessing techniques for some sensor environments, as well as the summaries of how to deploy Deep Learning in time series modeling while alleviating catastrophic forgetting with continual learning methods. The selected case studies cover a wide collection of various sensor time series applications and can illustrate how to deploy tailor-made Deep Learning, advanced preprocessing techniques, and continual learning algorithms from practical, real-world application aspects.

show abstract

Section: Advances In Deep Learning Methods For Time Series Modelingmentioning

confidence: 99%

Section: Figurementioning

confidence: 99%

Continual Deep Learning for Time Series Modeling

Ao¹,

Fayek

2023

Sensors

View full text Add to dashboard Cite

show abstract

“…The use of PlanetScope monthly basemap mosaics in the visual spectrum (R, G, and B bands) provided by Planet Labs at 4.77 m resolution could facilitate large-scale applications of imagery. By careful training of RGB imagery as input data, deep learning models have the capability of learning complex patterns and detect forest loss accurately over varied landscapes (Han and Sanchez-Azofeifa, 2022).…”

Section: Introductionmentioning

confidence: 99%

Detection of forest disturbance across California using deep-learning on PlanetScope imagery

Carter,

Wagner,

Dalagnol

et al. 2024

Front. Remote Sens.

View full text Add to dashboard Cite

California forests have recently experienced record breaking wildfires and tree mortality from droughts, However, there is inadequate monitoring, and limited data to inform policies and management strategies across the state. Although forest surveys and satellite observations of forest cover changes exist at medium to coarse resolutions (30–500 m) annually, they remain less effective in mapping small disturbances of forest patches (<5 m) occurring multiple times a year. We introduce a novel method of tracking California forest cover using a supervised U-Net deep learning architecture and PlanetScope’s Visual dataset which provides 3-band RGB (Red, Green, and Blue) mosaicked imagery. We created labels of forest and non-forest to train the U-Net model to map tree cover based on a semi-unsupervised classification method. We then detected changes of tree cover and disturbance with the U-Net model, achieving an overall accuracy of 98.97% over training data set, and 95.5% over an independent validation dataset, obtaining a precision of 82%, and a recall of 74%. With the predicted tree cover mask, we created wall to wall monthly tree cover maps over California at 4.77 m resolution for 2020, 2021, and 2022. These maps were then aggregated in a post-processing step to develop annual maps of disturbance, while accounting for the time of disturbance and other confounding factors such as topography, phenological and snow cover variability. We compared our high-resolution disturbance maps with wildfire GIS survey data from CALFIRE, and satellite-based forest cover changes and achieved an F-1 score of 54% and 88% respectively. The results suggest that high-resolution maps capture variability of forest disturbance and fire that wildfire surveys and medium resolution satellite products cannot. From 2020 to 2021, California maintained 30,923.5 sq km of forest while 5,994.9 sq km were disturbed. The highest observed forest loss rate was located at the Sierra Nevada mountains at 21.4% of the forested area being disturbed between 2020 and 2021. Our findings highlight the strong potential of deep learning and high-resolution RGB optical imagery for mapping complex forest ecosystems and their changes across California, as well as the application of these techniques on a national to global scale.

show abstract

“…In light of this, the field of machine learning offers a means of addressing and ultimately resolving these issues. Local wood industries can use machine learning to create uniform wood classes, allowing for reliable automatic classification [7].…”

Section: Introductionmentioning

confidence: 99%

Cedarwood Quality Classification using SVM Classifier and Convolutional Neural Network (CNN)

Murti¹,

Setianingsih²,

Kusumawardhani³

et al. 2022

IJACSA

View full text Add to dashboard Cite

Cedarwood is one of the most sought-after materials since it can be used to create a wide variety of household appliances. Other than its unique aroma, the product's quality is the most important selling attribute. Fiber patterns allow for a qualitative categorization of this wood. Traditionally, workers in the wood-processing business have relied solely on their eyesight to sort materials into several categories. As a result, there will be discrepancies in precision and efficiency, which will hurt the reputation of the regional wood sector. The answer to this issue is machine learning. In this study, we compare the performance of two different cedarwood quality classification systems where both systems use different machine learning methods namely Support Vector Machine (SVM) and Convolutional Neural Network (CNN). Each system will be sent images captured with a Logitech Brio 4K equipped with a joystick and ultrasonic sensors, labeled as belonging to one of five cedar classes (A, B, C, D, or E). In the initial method to learn the wood's pattern and texture, the Histogram of Oriented Gradient (HOG) will be used to identify the material. Meanwhile, the classification method uses a Support Vector Machine (SVM) which will be compared to find the best accuracy and time computation. The first system's experiment achieves 90 percent accuracy with a computation time of 1.40 seconds. For the second, we use a Convolutional Neural Network, a deep learning technique, to classify cedarwood (CNN). Extraction of features occurs in the convolution, activation, and pooling layers. Experimental results demonstrated a considerable enhancement, with an accuracy of 97% and a prediction speed of 0.56 seconds.

show abstract

A Deep Learning Time Series Approach for Leaf and Wood Classification from Terrestrial LiDAR Point Clouds

Cited by 14 publications

References 51 publications

Continual Deep Learning for Time Series Modeling

Continual Deep Learning for Time Series Modeling

Detection of forest disturbance across California using deep-learning on PlanetScope imagery

Cedarwood Quality Classification using SVM Classifier and Convolutional Neural Network (CNN)

Contact Info

Product

Resources

About