Moths are pests that pose a major threat to food production in China, and the monitoring and prevention of moth infestation is of great significance. To address the problems of a high diversity of moths with minor differences and difficult identification, a semantic segmentation network based on depthwise separable convolution, attention mechanism, pyramid pooling—Depthwise Squeeze-and-Excitation Pyramid network (DSEPNet)—was proposed. The network to extract texture features and wing edge information of moths was enhanced based on the optimization of the model of channel attention mechanism on UNet. The computational speed of the model was increased and the number of parameters of the model was reduced based on the improvement in depthwise separable convolution. A pyramid pooling module was added between the encoder and decoder so that the model could input images of an arbitrary size, while enhancing its ability to learn feature information of different dimensions. DSEPNet was evaluated by ablation and contrast experiments. Compared with UNet, the accuracy, mean intersection over union (mIoU), and F1-Score of DSEPNet were improved by 2.04%, 9.14%, and 4.08%, respectively. Based on the moth dataset, compared with R2AU-Net, the mIoU of DSEPNet was improved by 3.04%. To verify the generalization of the model, comparison experiments were done on the Pascal VOC 2012 dataset. The mIoU of DSEPNet was improved by 0.51% compared with PSPNet and by 0.18% compared with DeepLabv3. Meanwhile, an automatic annotation algorithm for data sets was proposed to solve the time-consuming and laborious process of manual annotation, which can automatically generate semantic segmentation annotation files. DSEPNet can be installed on moth traps to identify moths and monitor the number and species of moths in the area.
Through its association with proteins and plant pigments, leaf nitrogen (N) plays an important regulatory role in photosynthesis, leaf respiration, and net primary production. However, the traditional methods of measurement leaf N are rooted in sample-based spectroscopy in laboratory. There is a big challenge of deriving leaf N from the nondestructive field-measured leaf spectra. In this study, the original PROSPECT model was extended by replacing the absorption coefficient of chlorophyll in the original PROSPECT model with an equivalent N absorption coefficient to develop a nitrogen-based PROSPECT model (N-PROSPECT). N-PROSPECT was evaluated by comparing the model-simulated reflectance values with the measured leaf reflectance values. The validated results show that the correlation coefficient (R) was 0.98 for the wavelengths of 400 to 2500 nm. Finally, N-PROSPECT was used to simulate leaf reflectance using different combinations of input parameters, and partial least squares regression (PLSR) was used to establish the relationship between the N-PROSPECT simulated reflectance and the corresponding leaf nitrogen density (LND). The inverse of the PLSR-based N-PROSPECT model was used to retrieve LND from the measured reflectance with a relatively high accuracy (R2=0.77, RMSE=22.15 μg cm−2). This result demonstrates that the N-PROSPECT model established in this study can accurately simulate nitrogen spectral contributions and retrieve LND.
Dynamic mapping and monitoring of crop harvest on a large spatial scale will provide critical information for the formulation of optimal harvesting strategies. This study evaluates the feasibility of C-band polarimetric synthetic aperture radar (PolSAR) for monitoring the harvesting progress of oilseed rape (Brassica napus L.) fields. Five multitemporal, quad-pol Radarsat-2 images and one optical ZY-1 02C image were acquired over a farmland area in China during the 2013 growing season. Typical polarimetric signatures were obtained relying on polarimetric decomposition methods. Temporal evolutions of these signatures of harvested fields were compared with the ones of unharvested fields in the context of the entire growing cycle. Significant sensitivity was observed between the specific polarimetric parameters and the harvest status of oilseed rape fields. Based on this sensitivity, a new method that integrates two polarimetric features was devised to detect the harvest status of oilseed rape fields using a single image. The validation results are encouraging even for the harvested fields covered with high residues. This research demonstrates the capability of PolSAR remote sensing in crop harvest monitoring, which is a step toward more complex applications of PolSAR data in precision agriculture.
Early detection of bruises on apples is important for an automatic apple sorting system. A hyperspectral imaging system with the wavelength range of 400 to 1000nm was built for detecting bruises happened in an hour on ‘Fuji’ apples. Principal components analysis (PCA) was conducted on the hyperspecrtral images and the principal components (PC) images were compared. Three effective wavelengths 780, 850 and 960nm were determined using the weighing coefficients plot of the best PC image. Then, a multi-spectral imaging system with three bands 780, 850 and 960nm in the near-infrared range was developed. The system was consisted of two beamsplitters at 805 and 900nm, two bandpass filters and halogen tungsten lamp, and three CCD cameras. Images of 20 intact and 20 bruised apples were acquired. PCA was conducted on the three-band images of each apple and the best PC image was selected for bruise detection. A bruise detection algorithm based on the PC images and a global threshold method was developed. Results show that 90% of the bruised apples are correctly recognized.
Attribute of apple according to geographical origin is often recognized and appreciated by the consumers. It is usually an important factor to determine the price of a commercial product. Hyperspectral imaging technology and supervised pattern recognition was attempted to discriminate apple according to geographical origins in this work. Hyperspectral images of 207 Fuji apple samples were collected by hyperspectral camera (400-1000nm). Principal component analysis (PCA) was performed on hyperspectral imaging data to determine main efficient wavelength images, and then characteristic variables were extracted by texture analysis based on gray level co-occurrence matrix (GLCM) from dominant waveband image. All characteristic variables were obtained by fusing the data of images in efficient spectra. Support vector machine (SVM) was used to construct the classification model, and showed excellent performance in classification results. The total classification rate had the high classify accuracy of 92.75% in the training set and 89.86% in the prediction sets, respectively. The overall results demonstrated that the hyperspectral imaging technique coupled with SVM classifier can be efficiently utilized to discriminate Fuji apple according to geographical origins.
The application of eco-environment information management and the Land Use and Cover Change (LUCC) models in
system construction and data processing has formed a comparative matured system, but coupling using of them in the
information service system construction of eco-environment has not been thoroughly investigated. At present, the
management decision-making of the eco-environment urgently needs a kind of integrated, efficient and practical
technology to achieve intensive management. Because the eco-environment resources characterized by the broad
distribution and the complex structure, the 3S (GPS, GIS and RS) and other key technologies must be relied on to
achieve the targets of "automatic, efficient, informational and precise". In this paper, an information platform was
designed systematically according to the needs of dynamic monitoring and information management for ecoenvironment
using J2EE technology, WebGIS technology integrated with traditional MIS/OA seamlessly, by means of
spatial database, 3S integration technology, three-dimensional virtual simulation, computer network technology, etc. A
database of urban infrastructure was established, and the LUCC model service technology was embedded into the
platform for its significance on the eco-environment. This system can automatically analyze and classify different dates
of RS image data with the ability to dynamically export the LUCC maps, and to synchronously update the information
resources and network database. Results show that this system enhances the awareness as well as ability of analyzing and
forecasting the dynamic process of LUCC so as to provide a macro decision-making basis for the relevant departments.
Construction of network clusters and identifying hub nodes from networks has attracted more and more attentions in
spatial network analysis. In this paper, we proposed clustering algorithm and outlying node detection algorithm for
spatial road network analysis. Network clustering algorithm consists of constructing clusters and creating a simplified
structure of the network. When performing clustering on the network, we introduced the definitions of strong cluster and
weak cluster, where each node has more connections within the cluster than with the rest of the graph, for achieving
reliable and reasonable clusters. For users' understanding the structure of the network, we constructed a simplified graph
approximation of the network, whose nodes were representative nodes in clusters of the network, and edges were the
connections between those representative nodes. In outlying node detection algorithm, a node is identified as an outlier,
not because of its distribution different from that of other nodes but for its unexpected statistical information. Whether a
node is an outlier or not is examined with centrality index. The larger the node has centrality indexes, the more
probabilistically it may be identified as an outlier. The experimental results on artificial data sets demonstrated that two
algorithms are very efficient and effective.
Naked cropland elimination is an important part of Beijing Olympic ecological project. In this paper, Multi-temporal
satellite data were used to monitor and position the naked croplands. Three Landsat TM images and two "Beijing-
1"Micro-Satellite images were selected to calculate NDVI series according to crop phenological calendars and
investigated information of agricultural cropping structures in Beijing suburb. Based on the phenological spectral
characteristics of main agricultural land use types, a classification scheme was proposed to extract the naked croplands.
Considering the structural characteristic hierarchical classification and various demands of feature selection in different
periods, decision tree algorithm and a stepwise masking technology were employed to extract typical crops in each
season, and hence the naked croplands were left. Accuracy assessment of the naked croplands in winter and spring were
performed with comparison of the monitoring areas with statistical data. The results show that the area of the naked
croplands in winter and spring was 170368.1ha in Beijing. The areas of the top five districts (Yanqing, Shunyi, Daxing,
Miyun and Tongxian) were 17933.3ha, taking a percent of 69.2% of that of Beijing. The areas of the naked cropland
were 25719.6 ha, 4485.4 ha and 3325 ha in summer, autumn and all the year round respectively. Experimental results
demonstrated that our method would fast and simply monitor agricultural land use.
The Advanced technology in space-borne determination of grain crude protein content (CP) by remote sensing can help
optimize the strategies for buyers in aiding purchasing decisions, and help farmers to maximize the grain output by
adjusting field nitrogen (N) fertilizer inputs. We performed field experiments to study the relationship between grain
quality indicators and foliar nitrogen concentration (FNC). FNC at anthesis stage was significantly correlated with CP,
while spectral vegetation index was significantly correlated to FNC. Based on the relationships among nitrogen
reflectance index (NRI), FNC and CP, a model for CP prediction was developed. NRI was able to evaluate FNC with a
higher coefficient of determination of R2=0.7302. The method developed in this study could contribute towards
developing optimal procedures for evaluating wheat grain quality by ASTER image at anthesis stage. The RMSE was
0.893 % for ASTER image model, and the R2 was 0.7194. It is thus feasible to forecast grain quality by NRI derived
from ASTER image.
Advanced site-specific determination of grain protein content by remote sensing can provide opportunities to optimize the strategies for purchasing and pricing grain, and to maximize the grain output by adjusting field inputs. Field experiments were performed to study the relationship between grain quality indicators and foliar nitrogen concentration. Foliar nitrogen concentration at the anthesis stage is suggested to be significantly correlated with grain protein content, while spectral vegetation index is significantly correlated to foliar nitrogen concentration around the anthesis stage. Based on the relationships among nitrogen reflectance index (NRI), foliar nitrogen concentration, and grain protein content, a statistical evaluation model of grain protein content was developed. NRI proved to be able to evaluate foliar nitrogen concentration with a coefficient of determination of R2= 0.7302 in year 2002. The relationship between measured and remote sensing derived foliar nitrogen concentration had a coefficient of determination of R2=0.7279 in year 2003. The results mentioned above indicate that the inversion of foliar nitrogen concentration and the evaluation of grain protein content by NRI are surprisingly good.
Winter wheat canopy spectrum is dominated by wheat canopy closures, in this study. Our purpose is to study the quantitative influence of canopy closures on field canopy spectrum by quantitative reduced canopy stem densities. It indicated that canopy reflectance of winter wheat under different canopy stem densities has significant difference in near infrared bands. It has line relationship between spectral reflectance of 100% canopy stem densities and spectral reflectance under canopy stem densities, all the coefficients of determination (R2) for the equations are exceeding 0.8452, and all the results are surprised well. Canopy reflectance difference of winter under different stem densities were also studied, they all have line relationships between canopy reflectance of 100% canopy stem densities and quantitative reduced canopy stem densities, the simulation equations are different for the erective cultivars and loose cultivars. Relationship between canopy closures and canopy stem densities were also studied too, it has positive relationship between canopy closures and canopy stem densities, it reveals a very good agreement between canopy closures and canopy stem densities, with a coefficient of determination (R2) 0.8467, so the canopy stem densities can be simulated by canopy closures.
Investigations have been made on agronomy parameters as leaf area index (LAI), chlorophyll content (Chl), total Nitrogen (TN) and specific leaf weight (SLW) to describe growth status of winter wheat. More comprehensive parameters such as chlorophyll index (CI), projective chlorophyll index (CIp), Nitrogen index (NI) and projective Nitrogen index (NIp) have been developed to describe the dynamic growth information for foliage vertical layers by studying their vertical distribution characteristics along canopy and their spectral reflectance. Results are that from the canopy top to the ground surface, TN and Chl have shown an obvious gradient decreasing trend, while LAI and SLW have shown the ovate distribution. Compared with NI, CI and LAI, the absolute values of NIp, CIp and LAIp are more affected by canopy shape. The ratio of NIp to NI in different layers of erective varieties is significantly lower than that of loose varieties. Correlation analysis between canopy spectral reflectance and those developed parameters in different foliage layers at stage of anthesis shows that foliage Chl in upper layer is very sensitive to 400 nm-750 nm and 1300 nm-1750 nm while that in the middle layer is very sensitive to 750 nm -1300 nm. Higher correlation coefficient between spectral reflectance and TN is found in middle-under layer and it decreases upward.
A field trial was conduct to investigate the relationship between spectral feature of winter wheat canopy and LAI as well as leaf nitrogen (N) under different status of leaf water in field situation. The objective of this study is to investigate effect of water status in plants on the accuracy of estimating leaf area index (LAI) and plant nitrogen. The new defined spectral index, IAFC = (R2224-R2054)/ (R2224+R2054), where R is the reflectance at 2224nm or 2054nm, was significantly (α=0.05) or extremely significantly (α=0.01) correlated with LAI at all the six dates for water insufficient plants, but not significantly correlated for water sufficient plants at five of the six dates and the difference of leaf water content between the water insufficient plants and water sufficient plants was only about 2% at some dates. The study provided strong evidence that leaf water has a strong masking effect on the 2000-2300nm spectral feature, which could be strongly associated with LAI and leaf N even when the leaf water content was as high as about 80% if the water was insufficient for plant growth. The results indicated that the masking effect of leaf water on the 2000-2300nm spectral feature was not only dependent on the absolute plant water content but also on the water status and that remotely sensed data in the 2000-2300nm region could be of potential in monitoring plant canopy biophysics and biochemistry in drought condition.
This study was to develop the time-specific and time-critical method to overcome the limitations of traditional field sampling methods for variable rate fertilization. Farmers, agricultural managers and grain processing enterprises are interested in measuring and assessing soil and crop status in order to apply adequate fertilizer quantities to crop growth. This paper focused on studying the relationship between vegetation index (OSAVI) and nitrogen content to determine the amount of nitrogen fertilizer recommended for variable rate management in precision agriculture. The traditional even rate fertilizer management was chosen as the CK. The grain yield, ear numbers, 1000-grain weight and grain protein content were measured among the CK, uniform treatments and variable rate fertilizer treatments. It indicated that variable rate fertilization reduced the variability of wheat yield, ear numbers and dry biomass, but it didn't increased crop yield and grain protein content significantly and did not decrease the variety of 1000-grain weight, compared to traditional rate application. The nitrogen fertilizer use efficiency was improved, for this purpose, the variable rate technology based on vegetation index could be used to prevent under ground water pollution and environmental deterioration.
The recent study of agriculture has faced more and more time-consuming problems and digitalizing problems. In order to solve these problems, we raise our proposal of creating a new virtual agriculture system using the technologies of the virtual reality and artificial intelligence. In this paper, we will not only introduce the composition and the control mechanism of this new agriculture system, but also prove its promising future. We will also discuss the wide application fields of this new system.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.