Effects of Image Dataset Configuration on the Accuracy of Rice Disease Recognition Based on Convolution Neural Network.

In recent years, the convolution neural network has been the most widely used deep learning algorithm in the field of plant disease diagnosis and has performed well in classification. However, in practice, there are still some specific issues that have not been paid adequate attention to. For instance, the same pathogen may cause similar or different symptoms when infecting plant leaves, while the same pathogen may cause similar or disparate symptoms on different parts of the plant. Therefore, questions come up naturally: should the images showing different symptoms of the same disease be in one class or two separate classes in the image database? Also, how will the different classification methods affect the results of image recognition? In this study, taking rice leaf blast and neck blast caused by Magnaporthe oryzae, and rice sheath blight caused by Rhizoctonia solani as examples, three experiments were designed to explore how database configuration affects recognition accuracy in recognizing different symptoms of the same disease on the same plant part, similar symptoms of the same disease on different parts, and different symptoms on different parts. The results suggested that when the symptoms of the same disease were the same or similar, no matter whether they were on the same plant part or not, training combined classes of these images can get better performance than training them separately. When the difference between symptoms was obvious, the classification was relatively easy, and both separate training and combined training could achieve relatively high recognition accuracy. The results also, to a certain extent, indicated that the greater the number of images in the training data set, the higher the average classification accuracy.

Influence of Ethanol on Pitting Corrosion Behavior of Stainless Steel for Bioethanol Fermentation Tanks.

The role of ethanol (C2H5OH) in pitting corrosion behavior of AISI 316L austenitic stainless steel was investigated in aqueous ethanolic solution with chloride. The pitting susceptibility and surface morphology of 316L in a series of ethanol-containing solutions were examined using X-ray photoelectron spectroscopy (XPS), optical microscopy with 3D stitching, immersion tests, and potentiodynamic polarization measurements. Results demonstrated that the ethanol concentration impacted little on the passive film stability while it dramatically influenced the pitting corrosion susceptibility. Corrosion rate of 316L after immersion tests first increased and then decreased as the concentration of ethanol increased from 0 to 10 M in ferric chloride solution. This, however, did not correspond to the breakdown potential which directly decreased from 489 to 249 mV as the water concentration decreased in ethanolic NaCl solutions. The pits density after both immersion and electrochemical tests showed that the initiation of pitting in ethanolic solution tended to occur at multiple points at the same time. The synergy effect on pitting behavior of hydrolysis enhancement and solubility reduction of metal cations due to the introduction of ethanol has also been discussed.

Influence of macroscopic defects on the corrosion behavior of U-0.79 wt%Ti alloy in sodium chloride solution.

Uranium alloys containing a low concentration of titanium have received wide attention due to their greatly enhanced corrosion resistance and outstanding mechanical performances. Herein, we investigated the effect of macroscopic defects on the corrosion behavior of U-0.79 wt%Ti (denoted as U-Ti) alloy in 0.01 M NaCl solution using traditional electrochemical testing technologies and a novel scanning electrochemical composite probe (SECP). The results demonstrate that pitting corrosion occurs rapidly on the alloy surface due to macroscopic defects. Moreover, macroscopic defects led to a decrease in corrosion potential and polarization resistance, and an increase in corrosion current density. Furthermore, the potential and pH value distributions were detected in the same region using the composite probe. The results show that the region around the macroscopic defects become corrosion-active positions and the potential difference (vs. the average potential of the alloy surface) in this area is significantly higher than that at positions without macroscopic defects, while the opposite was observed for the pH value distribution. In addition, the distribution of the vertical direction (Z) potential at the active point was clearly different from that at the inactive point. A possible reason for this could lie in the difference in the electric field distribution and electrode reaction type between the active point and inactive point on the alloy surface.

Enzyme-free and label-free ultra-sensitive colorimetric detection of Pb(2+) using molecular beacon and DNAzyme based amplification strategy.

An enzyme-free and label-free colorimetric Pb(2+) sensor based on DNAzyme and molecular beacon (MB) has been developed and demonstrated by recycle using enzyme strand for signal amplification. The substrate strand DNA (S-DNA) of DNAzyme could be converted into MB structure with base pairs of stem part at the both ends. The MB could hybridize with enzyme strand DNA (E-DNA) to form DNAzyme, and be activated and cleaved in the presence of Pb(2+). The cleaved MB is much less stable, releasing from the DNAzyme as two product pieces. The product pieces of MB are flexible and could bind to unmodified AuNPs to effectively stabilize them against salt-induced aggregation. Then, the E-DNA is liberated to catalyze the next reaction and amplify the response signal. By taking advantage of repeated using of E-DNA, our proposed method exhibited high sensitive for Pb(2+) detection in a linear range from 0.05 to 5 nM with detection limit of 20 pM by UV-vis spectrometer. Moreover, this method was also used for determination of Pb(2+) in river water samples with satisfying results. Importantly, this strategy could reach high sensitivity without any modification and complex enzymatic or hairpins based amplification procedures.

Ultrasensitive visual detection of DNA with tunable dynamic range by using unmodified gold nanoparticles and target catalyzed hairpin assembly amplification.

A simple and novel strategy for enzyme-free ultrasensitive DNA detection platform has been present here based on gold nanoparticles (AuNPs) colorimetry and target catalyzed hairpin assembly amplification. Three hairpin auxiliary probes (H1, H2, and H3) are designed with signal-stranded DNA (ssDNA) sticky ends which could effectively stabilize AuNPs against salt-induced aggregation. However, a cascade of assembly steps with H1, H2, and H3 are activated in the presence of the target DNA, followed by a disassembly step in which H3 displaces the target DNA from the complex, freeing the target DNA to catalyze the self-assembly of additional branched junctions. The formed branched junction consisted with dsDNA is stiffer, and cannot prevent salt-induced AuNPs aggregation, corresponding to a red-to-blue color change. The result can be read out by naked eyes or UV-vis spectrometer. The detection limit of this method is 0.1 pM by naked eyes, and this result is comparable or even better than enzyme or hybridization chain reaction (HCR) based amplification AuNPs colorimetric assays. Moreover, the dynamic range of sensor could be tuned by using different concentration of hairpins. Importantly, this strategy provides a versatile ultrasensitive detection platform for the DNA and related filed targets including metal ions, small molecules, proteins, cells et al. by combining with specific DNAzymes and aptamers.