scholarly journals A novel hyperspectral compressive sensing framework of plant leaves based on multiple arbitrary-shape regions of interest

2021 ◽  
Vol 7 ◽  
pp. e802
Author(s):  
Yuewei Jia ◽  
Lingyun Xue ◽  
Ping Xu ◽  
Bin Luo ◽  
Ke-nan Chen ◽  
...  

Massive plant hyperspectral images (HSIs) result in huge storage space and put a heavy burden for the traditional data acquisition and compression technology. For plant leaf HSIs, useful plant information is located in multiple arbitrary-shape regions of interest (MAROIs), while the background usually does not contain useful information, which wastes a lot of storage resources. In this paper, a novel hyperspectral compressive sensing framework for plant leaves with MAROIs (HCSMAROI) is proposed to alleviate these problems. HCSMAROI only compresses and reconstructs MAROIs by discarding the background to achieve good reconstructed performance. But for different plant leaf HSIs, HCSMAROI has the potential to be applied in other HSIs. Firstly, spatial spectral decorrelation criterion (SSDC) is used to obtain the optimal band of plant leaf HSIs; Secondly, different leaf regions and background are distinguished by the mask image of the optimal band; Finally, in order to improve the compression efficiency, after discarding the background region the compressed sensing technology based on blocking and expansion is used to compress and reconstruct the MAROIs of plant leaves one by one. Experimental results of soybean leaves and tea leaves show that HCSMAROI can achieve 3.08 and 5.05 dB higher PSNR than those of blocking compressive sensing (BCS) at the sampling rate of 5%, respectively. The reconstructed spectra of HCSMAROI are especially closer to the original ones than that of BCS. Therefore, HCSMAROI can achieve significantly higher reconstructed performance than that of BCS. Moreover, HCSMAROI can provide a flexible way to compress and reconstruct different MAROIs with different sampling rates, while achieving good reconstruction performance in the spatial and spectral domains.

2017 ◽  
Vol 17 (3) ◽  
pp. 434-449 ◽  
Author(s):  
Zhiliang Bai ◽  
Shili Chen ◽  
Qiyang Xiao ◽  
Lecheng Jia ◽  
Yanbo Zhao ◽  
...  

Ultrasonic phased array techniques are widely used for defect detection in structural health monitoring field. The increase in the element number, however, leads to larger amounts of data acquired and processed. Recently developed compressive sensing states that sparse signals may be accurately recovered from far fewer measurements, suggesting the possibility of breaking through the sampling limit of the Nyquist theorem. In light of this significant advantage, the novel use of the compressive sensing methodology for ultrasonic phased array in defect detection is proposed in this work. Based on CIVA software, we first present a simulated study on the effectiveness of the compressive sensing applied in ultrasonic phased array in defect detection through the average mean percent residual difference at varying compression rates. The results particularly show that the compressive sensing yields a breakthrough of the sampling limitation. We then experimentally demonstrate comparative analyses on the signals extracted from three types of artificial flaws (through-hole, flat-bottom hole, and electrical discharge machining notches) on two different specimens (made of aluminum and 20# steel). To find the optimal algorithm combination, the best sparse representation basis is chosen among fast Fourier transform, discrete cosine transform, and 34 wavelet kernels; the reconstruction performance is compared between five greedy algorithms; and the recovery accuracy is further improved via four sensing matrices selection. We also evaluate the influence of the sampling rate, and our results are comparable with the gold standard of signal compression, namely, the discrete wavelet transform.


2020 ◽  
Vol 2020 ◽  
pp. 1-8
Author(s):  
Thiago L. Marques ◽  
Milton K. Sasaki ◽  
Lidiane C. Nunes ◽  
Fábio R. P. Rocha

Salicylic acid (SA) is an important stress signaling phytohormone and plays an essential role in physiological processes in plants. SA fractionation has been carried out batchwise, which is not compatible with the high analytical demand in agronomical studies and increases susceptibility to analytical errors. In this context, a novel flow-batch sample preparation system for SA fractionation on fresh plant leaves was developed. It was based on microwave-assisted extraction with water and conversion of the conjugated species to free SA by alkaline hydrolysis. Free and total SA were quantified by fluorimetry after separation by sequential injection chromatography in a C18 monolithic column. The proposed procedure is directly applicable to plant leaves containing up 16 mg kg−1 SA, with a limit of detection of 0.1 mg kg−1 of SA, coefficient of variation of 3.0% (n = 10), and sampling rate of 4 samples h−1. The flow-batch sample preparation system was successfully applied to SA fractionation in sugarcane, corn, and soybean leaves without clogging or increasing in backpressure. The proposed approach is simple, less time-consuming, and more environmentally friendly in comparison to batchwise procedures.


Electronics ◽  
2021 ◽  
Vol 10 (4) ◽  
pp. 385
Author(s):  
Kunhao Zhang ◽  
Yali Qin ◽  
Huan Zheng ◽  
Hongliang Ren ◽  
Yingtian Hu

The use of non-local self-similarity prior between image blocks can improve image reconstruction performance significantly. We propose a compressive sensing image reconstruction algorithm that combines bilateral total variation and nonlocal low-rank regularization to overcome over-smoothing and degradation of edge information which result from the prior reconstructed image. The proposed algorithm makes use of the preservation of image edge information by bilateral total variation operator to enhance the edge details of the reconstructed image. In addition, we use weighted nuclear norm regularization as a low-rank constraint for similar blocks of the image. To solve this convex optimization problem, the Alternating Direction Method of Multipliers (ADMM) is employed to optimize and iterate the algorithm model effectively. Experimental results show that the proposed algorithm can obtain better image reconstruction quality than conventional algorithms with using total variation regularization or considering the nonlocal structure of the image only. At 10% sampling rate, the peak signal-to-noise ratio gain is up to 2.39 dB in noiseless measurements compared with Nonlocal Low-rank Regularization (NLR-CS). Reconstructed image comparison shows that the proposed algorithm retains more high frequency components. In noisy measurements, the proposed algorithm is robust to noise and the reconstructed image retains more detail information.


Author(s):  
Vivek K. Verma ◽  
Tarun Jain

The disease occurrence phenomena in plants are season-based which is dependent on the presence of the pathogen, crops, environmental conditions, and varieties grown. Some plant varieties are particularly subject to outbreaks of diseases; on the other hand, some are opposite to them. Huge numbers of diseases are seen on the plant leaves and stems. Diseases management is a challenging task. Generally, diseases are seen on the leaves or stems of the plant. Image processing is the best way for the detection of plant leaf diseases. Different kinds of diseases occur because of the attack of bacteria, fungi, and viruses. The monitoring of leaf area is an important tool in studying physiological capabilities associated with plant boom. Plant disorder is usually an unusual growth or dysfunction of a plant. Sometimes diseases damage the leaves of plants.


2010 ◽  
Vol 18 (3) ◽  
pp. 188-195 ◽  
Author(s):  
Algimantas Sirvydas ◽  
Vidmantas Kučinskas ◽  
Paulius Kerpauskas ◽  
Jūratė Nadzeikienė ◽  
Albinas Kusta

Solar radiation energy is used by vegetation, which predetermines the existence of biosphere. The plant uses 1–2% of the absorbed radiant energy for photosynthesis. All the remaining share of the absorbed energy, accounting for 99–98%, converts into thermal energy in the plant leaf. At the lowest wind under natural surrounding air conditions, plant leaves change their position with respect to the Sun. An oscillating plant leaf receives a variable amount of solar radiation energy, which causes changes in the balance of plant leaf energies and a changing emission of heat in the leaf. The analysis of solar radiation energy pulsations in the plant leaf shows that when the leaf is in the edge positions of angles 10°, 20° and 30° with respect to the Sun, 1.5%; 6% and 13% less of radiation energy reach the leaf, respectively. During periodic motion, when the amplitude of leaf oscillation is no bigger than 10°, the plant surface receives up to 1.6% less of solar radiation energy within a certain period of time, and when the amplitude of oscillation reaches 30° up to 14% less of solar radiation energy reach the leaf surface. The total amount of radiant energy received during pulsations of solar radiation energy is not dependent on the frequency of oscillation in the same interval of time. Temperature pulsations occur in the leaf due to solar radiation energy pulsations when the plant leaf naturally changes its position with respect to the Sun. Santrauka Saules spinduliuotes energija būtina augalijai, kuri lemia biosferos egzistavima. Augalas 1–2 % absorbuotos spinduliuotes energijos sunaudoja fotosintezei, o 99–98 % absorbuotos energijos augalo lape virsta šilumine energija. Natūraliomis aplinkos salygomis esant mažiausiam vejui augalo lapu padetis Saules atžvilgiu keičiasi. Taigi augalo svyruojančio lapo gaunamas Saules spinduliuotes energijos kiekis yra kintamas, tai sukelia pokyčius augalo lapo energiju balanse ir kintama šilumos išsiskyrima lape. Analizuojant Saules spinduliuotes energijos pulsacijas augalo lape, nustatyta, kad, lapui esant kraštinese 10°, 20° ir 30° kampu padetyse Saules atžvilgiu, i ji atitinkamai patenka 1,5 %; 6 % ir 13 % mažiau spinduliuotes energijos. Augalo lapui periodiškai svyruojant, kai svyravimo amplitude yra iki 10°, per tam tikra laika i lapo paviršiu patenka iki 1,6 % mažiau Saules spinduliuotes energijos, o kai svyravimo amplitu‐de siekia iki 30°, – iki 14 % mažiau. Saules spinduliuotes energijos pulsaciju metu gautas bendras spinduliuotes energijos kiekis nepriklauso nuo to paties laiko intervalo svyravimo dažnio. Del Saules spinduliuotes energijos pulsaciju, natūraliai keičiantis augalo lapo padečiai Saules atžvilgiu, lape kyla temperatūros pulsacijos. Резюме Растения потребляют солнечную лучевую энергию, которая является основой существования биосферы. 1–2% абсорбированной лучевой энергии они используют на фотосинтез. В натуральных условиях при малейшем дуновении ветра листья растений меняют свое положение относительно Солнца. Колеблющийся лист получает переменное количество лучевой энергии, которое вызывает изменения в энергетическом балансе листа растения, что сказывается на переменном выделении тепла в листе. Анализируя пульсации солнечной лучевой энергии в листе растения, установлено, что при крайних положениях листа относительно Солнца на 10, 20 и 30 градусов на лист попадает соответственно на 1,5%, 6% и 13% меньше лучевой энергии. При периодическом колебании листа, когда амплитуда его колебания составляет 10 градусов, за известный промежуток времени солнечная лучевая энергия, попадающая на поверхность листа, уменьшается до 1,6%, а при амплитуде колебания до 30 градусов соответственно количество лучевой энергии на поверхности листа растения уменьшается до 14%. Установлено, что суммарное количество солнечной лучевой энергии во время пульсации не зависит от частоты колебания листа за одинаковый промежуток времени. Пульсации солнечной лучевой энергии при изменении положения листа растения относительно Солнца вызывают температурные пульсации в листе.


Author(s):  
Malusi Sibiya ◽  
Mbuyu Sumbwanyambe

Machine learning systems use different algorithms to detect the diseases affecting the plant leaves. Nevertheless, selecting a suitable machine learning framework differs from study to study, depending on the features and complexity of the software packages. This paper introduces a taxonomic inspection of the literature in deep learning frameworks for the detection of plant leaf diseases. The objective of this study is to identify the dominating software frameworks in the literature for modelling machine learning plant leaf disease detecting systems.


2011 ◽  
Vol 301-303 ◽  
pp. 1612-1617
Author(s):  
Wei Shen ◽  
Jun Zheng Wang ◽  
Jing Li

The vibrator wireless monitoring system is designed based on Zigbee and the theory of compressive sensing. This system consists of the sensor wireless acquisition device, the vibrator wireless automation recorder and the vibrator remote management computer. Different modules of the system communicate via the Zigbee wireless network. The data to transmit is compressed by applying compressive sensing technology. As result, the quantity of the data transmition decreases sharply with the reliability and real–time operation being ensured. The system is already in use and the effectiveness is proved.


2013 ◽  
Vol 558 ◽  
pp. 561-566
Author(s):  
Yue Quan Bao ◽  
Hui Li ◽  
Jin Ping Ou

Compressive sampling also called compressive sensing (CS) is a emerging information theory proposed recently. CS provides a new sampling theory to reduce data acquisition, which says that sparse or compressible signals can be exactly reconstructed from highly incomplete random sets of measurements. CS broke through the restrictions of the Shannon theorem on the sampling frequency, which can use fewer sampling resources, higher sampling rate and lower hardware and software complexity to obtain the measurements. Not only for data acquisition, CS also can be used to find the sparse solutions for linear algebraic equation problem. In this paper, the applications of CS for SHM are presented including acceleration data acquisition, lost data recovery for wireless sensor and moving loads distribution identification. The investigation results show that CS has good application potential in SHM.


2016 ◽  
Vol 26 (11) ◽  
pp. 1650191 ◽  
Author(s):  
Yushu Zhang ◽  
Jiantao Zhou ◽  
Fei Chen ◽  
Leo Yu Zhang ◽  
Di Xiao ◽  
...  

The existing Block Compressive Sensing (BCS) based image ciphers adopted the same sampling rate for all the blocks, which may lead to the desirable result that after subsampling, significant blocks lose some more-useful information while insignificant blocks still retain some less-useful information. Motivated by this observation, we propose a scalable encryption framework (SEF) based on BCS together with a Sobel Edge Detector and Cascade Chaotic Maps. Our work is firstly dedicated to the design of two new fusion techniques, chaos-based structurally random matrices and chaos-based random convolution and subsampling. The basic idea is to divide an image into some blocks with an equal size and then diagnose their respective significance with the help of the Sobel Edge Detector. For significant block encryption, chaos-based structurally random matrix is applied to significant blocks whereas chaos-based random convolution and subsampling are responsible for the remaining insignificant ones. In comparison with the BCS based image ciphers, the SEF takes lightweight subsampling and severe sensitivity encryption for the significant blocks and severe subsampling and lightweight robustness encryption for the insignificant ones in parallel, thus better protecting significant image regions.


Sign in / Sign up

Export Citation Format

Share Document