Methods and Challenges Using Multispectral and Hyperspectral Images for Practical Change Detection Applications

Kwan, Chiman

doi:10.3390/info10110353

Cited by 38 publications

(18 citation statements)

References 137 publications

(206 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Landsat-8 satellite is an example of a multispectral imager consisting of 11 bands with a high spatial resolution of 30 m in most bands. NASA's Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) is a hyperspectral imager with 224 bands with 0.4-2.5 µm (Kwan, 2019).…”

Section: Data Capturing Using Different Sensing Technologymentioning

confidence: 99%

Intelligent Fruit Yield Estimation for Orchards Using Deep Learning Based Semantic Segmentation Techniques—A Review

et al. 2021

View full text Add to dashboard Cite

Smart farming employs intelligent systems for every domain of agriculture to obtain sustainable economic growth with the available resources using advanced technologies. Deep Learning (DL) is a sophisticated artificial neural network architecture that provides state-of-the-art results in smart farming applications. One of the main tasks in this domain is yield estimation. Manual yield estimation undergoes many hurdles such as labor-intensive, time-consuming, imprecise results, etc. These issues motivate the development of an intelligent fruit yield estimation system that offers more benefits to the farmers in deciding harvesting, marketing, etc. Semantic segmentation combined with DL adds promising results in fruit detection and localization by performing pixel-based prediction. This paper reviews the different literature employing various techniques for fruit yield estimation using DL-based semantic segmentation architectures. It also discusses the challenging issues that occur during intelligent fruit yield estimation such as sampling, collection, annotation and data augmentation, fruit detection, and counting. Results show that the fruit yield estimation employing DL-based semantic segmentation techniques yields better performance than earlier techniques because of human cognition incorporated into the architecture. Future directions like customization of DL architecture for smart-phone applications to predict the yield, development of more comprehensive model encompassing challenging situations like occlusion, overlapping and illumination variation, etc., were also discussed.

show abstract

Section: Data Capturing Using Different Sensing Technologymentioning

confidence: 99%

Intelligent Fruit Yield Estimation for Orchards Using Deep Learning Based Semantic Segmentation Techniques—A Review

et al. 2021

View full text Add to dashboard Cite

show abstract

“…In order to validate the proposed TDRD effectively, results derived by the proposed method are compared with those derived by eight other methods, including absolute distance (AD) [5], absolute average difference (AAD) [43], [44], subspacebased change detection (SCD) [45], local SCD (LSCD) [45], adaptive SCD (ASCD) [45], ED [5], 4D-HOSVD [30], and M 2 C 2 VA [11].…”

Section: Parameters Settingmentioning

confidence: 99%

Three-Order Tucker Decomposition and Reconstruction Detector for Unsupervised Hyperspectral Change Detection

Hou

Tao

et al. 2021

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Change detection from multi-temporal hyperspectral images has attracted great attention. Most of traditional methods using spectral information for change detection treat a hyperspectral image as a 2D matrix, and do not take into account inherently structure information of spectrum, which leads to limited detection accuracy. To better approximate both spectral and spatial information, a novel three-order Tucker decomposition and reconstruction detector (TDRD) is proposed for hyperspectral change detection. Initially, Tucker decomposition and reconstruction strategies are used to eliminate influence of various factors in a multi-temporal dataset. Specifically, a singular value accumulation strategy is used to determine principal components in factor matrixs. Meanwhile, spectral angle is used to analyze spectral change after tensor processing in different domains. Finally, a new detector is designed to further improve the detection accuracy. Experiments conducted on five real hyperspectral datasets demonstrate that the proposed detector achieves better detection performance.

show abstract

“…This helps to enhance reflectance estimates in situations where the surrounding illumination conditions are changing. This situation requires a conversion to normalized reflectance [81]. During data collection, the two sensors were not synchronized for every frame, meaning images may not be perfectly matched.…”

Section: Dataset Collectionmentioning

confidence: 99%

Spatial Super Resolution of Real-World Aerial Images for Image-Based Plant Phenotyping

et al. 2021

View full text Add to dashboard Cite

Unmanned aerial vehicle (UAV) imaging is a promising data acquisition technique for image-based plant phenotyping. However, UAV images have a lower spatial resolution than similarly equipped in field ground-based vehicle systems, such as carts, because of their distance from the crop canopy, which can be particularly problematic for measuring small-sized plant features. In this study, the performance of three deep learning-based super resolution models, employed as a pre-processing tool to enhance the spatial resolution of low resolution images of three different kinds of crops were evaluated. To train a super resolution model, aerial images employing two separate sensors co-mounted on a UAV flown over lentil, wheat and canola breeding trials were collected. A software workflow to pre-process and align real-world low resolution and high-resolution images and use them as inputs and targets for training super resolution models was created. To demonstrate the effectiveness of real-world images, three different experiments employing synthetic images, manually downsampled high resolution images, or real-world low resolution images as input to the models were conducted. The performance of the super resolution models demonstrates that the models trained with synthetic images cannot generalize to real-world images and fail to reproduce comparable images with the targets. However, the same models trained with real-world datasets can reconstruct higher-fidelity outputs, which are better suited for measuring plant phenotypes.

show abstract

Methods and Challenges Using Multispectral and Hyperspectral Images for Practical Change Detection Applications

Cited by 38 publications

References 137 publications

Intelligent Fruit Yield Estimation for Orchards Using Deep Learning Based Semantic Segmentation Techniques—A Review

Intelligent Fruit Yield Estimation for Orchards Using Deep Learning Based Semantic Segmentation Techniques—A Review

Three-Order Tucker Decomposition and Reconstruction Detector for Unsupervised Hyperspectral Change Detection

Spatial Super Resolution of Real-World Aerial Images for Image-Based Plant Phenotyping

Contact Info

Product

Resources

About