首页期刊导航|Computers and Electronics in Agriculture
期刊信息/Journal information
Computers and Electronics in Agriculture
Elsevier Science Publishers
Computers and Electronics in Agriculture

Elsevier Science Publishers

0168-1699

Computers and Electronics in Agriculture/Journal Computers and Electronics in AgricultureSCIEIISTP
正式出版
收录年代

    TomatoScan: An Android-based application for quality evaluation and ripening determination of tomato fruit

    Sherafati A.Mollazade K.Koushesh Saba M.Vesali F....
    8页
    查看更多>>摘要：? 2022 Elsevier B.V.In this study, a new method consisting contact imaging and concentrated light beam injection was used to predict the indices related to the quality of tomato fruit and also to determine the ripening stage. In total 220 tomato samples were used belonging to six stages of ripening and two stages of storage. Contact images were taken by the RGB smartphone camera. After selecting the superior features of contact images by stepwise regression method, multilayer perceptron artificial neural networks were used to create prediction and classification models. The best prediction performance was obtained using white light for a* (CIELAB color space), titratable acidity, and soluble solid content, 650 nm laser light for carotenoid, combination of 532 and 650 nm lasers for L* (CIELAB color space), elasticity, and lycopene, and combination of 650 and 780 nm wavelengths for total chlorophyll. For classification of tomatoes based on their ripening stage, the white light was also found to be the best light source. Based on the architecture and the bias and weight values of neurons of created prediction/classification models in MATLAB, an application called TomatoScan was developed for Android smartphones. The results of evaluating the TomatoScan app were almost similar to the results obtained in the test stage of neural network models using MATLAB software. Based on the results obtained for the testing dataset, the correlation coefficient (R) values for estimating L*, a*, elasticity, total chlorophyll, carotenoid, lycopene, titratable acidity, and soluble solid content were 0.901, 0.964, 0.856, 0.664, 0.824, 0.923, 0.816, and 0.792, respectively, while the corresponding values for the mean square error were 3.549, 13.485, 0.000, 14.070, 0.065, 39.198, 0.058, and 0.259, respectively. TomatoScan was also able to determine the ripening stage of tomatoes with overall accuracy of 75.00 %.
      原文链接:
    • NSTL
    • Elsevier

    UAV-based multispectral and thermal cameras to predict soil water content – A machine learning approach

    Bertalan L.Pataki A.Szabo G.Szabo S....
    11页
    查看更多>>摘要：? 2022 The Author(s)Soil water content (SWC) estimation is a crucial issue of agricultural production, and its mapping is an important task. We aimed to study the efficacy of UAV-based thermal (TH) and multispectral (MS) cameras in SWC mapping. Soil samples were collected and the SWC content was determined in a laboratory as reference data and four machine learning regression algorithms (Random Forest [RF], Elastic Net [ENR], General Linear Model [GLM], Robust Linear Model [RLM]) were tested for the prediction efficacy, combined with three pixel value extraction methods (single pixel, mean of 20 and 30 cm radius buffer). We found that MS cameras ensured better input data than TH cameras: R2s were 0.97 vs 0.71, mean-normalized root mean square errors (nRMSE) were 10 vs 25 %, respectively. Best models were obtained by the RF (0.97 R2) and ENR (0.88 R2) in case of MS camera. Relationship between SWC and thermal data was exponential, which was incorrectly handled by the GLM (>40 % nRMSE; furthermore, RLM and ENR was not working with only one variable), thus, TH data was acceptable only with the RF (24.4 % nRMSE). Single pixel extraction provided the best input for the estimations, mean of buffered areas did not perform better in the models. Maps provided appropriate SWC estimations according to the nRMSEs, with high spatial resolution. In spite of potential inaccuracies, visualizing the spatial heterogeneities can be a great help to farmers to increase the efficacy of planning irrigation in precision agriculture.
      原文链接:
    • NSTL
    • Elsevier

    Adjacent age classification algorithm of yellow-feathered chickens based on multi-scale feature fusion

    Xia Y.Cui D.Liu Y.Li L....
    11页
    查看更多>>摘要：? 2022 Elsevier B.V.The age of yellow-feathered chicken is important to distinguish the freshness of meat quality in the trade of yellow-feathered chicken. To investigate whether the Convolutional Neural Network (CNN) model can be applied to the instar classification of yellow-feathered chickens, a multi-scale feature fusion model called Chicken_Age_Network (CANet) was proposed. The model uses Inception to construct a feature extraction layer and extract the feature information of each chicken face image to improve the classification accuracy of the model. First, the self-developed yellow-feathered chicken facial image collection application was used to collect images, and the yellow-feathered chicken image database was constructed by using day-age classification. Second, the standard face image of yellow-feathered chicken was obtained by using Structural Similarity Index Method detection(SSIM), image segmentation, background removal, and normalization. The adjacent age classification needs to extract more features, and CANet's feature extraction layer based on multi-scale feature fusion can extract features of different sizes. Another advantage of CANet is that the GMP (Global Max Pooling) replaces the final fully connected layer of the general CNN to reduce parameters and optimize the network model. Finally, chicken face images of adjacent days of age were tested and compared on VGG13, VGG19, DenseNet121, DenseNet161, SE-ResNet-20, MobileNet V1, ShuffleNet G2, ResNet50, ResNet34 and CSPDenseNet121. Test results show that CANet can quickly and accurately identify the age of yellow-feathered chickens. The classification accuracy on CANet model is 96.29 %, which is better than VGG13 (93.09 %), VGG19 (95.08 %), ResNet50 (85.53 %), ResNet34 (91.14 %), DenseNet121 (93.03 %), DenseNet161 (93.33 %), MobileNet V1(86.06 %), ShuffleNet G2(93.18 %), CSPDenseNet121(84.76 %) and SE-ResNet-20 (83.86 %). In order to verify the generalization of the model, experiments were carried out on the public data set CIFAR-10. Results show that CANet has good generalization and can be applied to other classification problems.
      原文链接:
    • NSTL
    • Elsevier

    A new criterion based on estimator variance for model sampling in precision agriculture

    Oger B.Tisseyre B.Vismara P.Le Moguedec G....
    12页
    查看更多>>摘要：? 2022Model sampling has proven to be an interesting approach to optimize the sampling of an agronomic variable of interest at the field level. The use of a model improves the quality of the estimates by making it possible to integrate the information provided by one or more auxiliary data. It has been shown that such an approach gives better estimations compared to more traditional approaches. Through a statistical work describing the properties of model sampling variance, this paper details how the different factors either related to sample characteristics or to the correlation between the auxiliary data and the variable of interest, affect estimation error. The resulting equations show that the use of samples with a mean close to the field mean and with a substantial dispersion reduces the estimation variance. On the basis of these statistical considerations, a variance criterion is defined to compare sample properties. The lower the value of the criterion of a sample, the lower the variance of the estimate and the expected errors. These theoretical insights were applied to real commercial vine fields in order to validate the demonstration. Nine vine fields were considered with the objective to provide the best yield estimation. High resolution vegetative index derived from airborne multispectral image was used to drive the sampling and the estimation. The theoretical considerations were verified on the nine fields; as the observed estimation errors correspond quite well to the values predicted by the equations. The selection of a large number of random samples from these fields confirms that samples associated with higher values of the chosen criterion result, on average, in larger yield estimation errors. Samples with the highest criterion values are associated with mean estimation errors up to two times larger than those of average samples. Random sampling is also compared to two target sampling approaches (Clustering based on quantiles or on k-means algorithm) commonly considered in the literature, whose characteristics improve the value of the proposed criterion. It is shown that these sampling strategies produce samples associated with criterion values up to 100 times smaller than random sampling. The use of these easy-to-implement methods thus guarantees to reduce the variance of the estimation and the estimation errors.
      原文链接:
    • NSTL
    • Elsevier

    Detection of pears with moldy core using online full-transmittance spectroscopy combined with supervised classifier comparison and variable optimization

    Zhang Q.Huang W.Wang Q.Li J....
    11页
    查看更多>>摘要：? 2022 Elsevier B.V.Moldy core is a serious disease that influences the quality of pears. There is no apparent difference between the diseased and sound fruit because this kind of disease mainly occurs in core of pears. In automatic detection and grading of pear quality, it is very desirable to detect pears with moldy core, especially at the early stage. This study used the visible and near infrared (Vis/NIR) full-transmittance spectroscopy, combined with wavelength selection algorithms and supervised classifiers, to discriminate 'Ya' pears with moldy core. The spectra of pears were collected using an online measurement system. Savitzky-Golay smoothing and standard normal variables (SGS-SNV) were used to preprocess the spectra. Four variable selection algorithms, including Monte Carlo-uninformative variable elimination (MC-UVE), bootstrapping soft shrinkage (BOSS), combination algorithm MC-UVE-SPA (successive projections algorithm), and combination algorithm BOSS-SPA, were used to extract the effective wavelengths. Four supervised classifiers, including support vector machine (SVM), least squares-support vector machine (LS-SVM), random forest (RF), and partial least square discriminant analysis (PLS-DA), were used for modeling. The two-class classification (i.e., sound and diseased) and the three-class classification (i.e., sound, slightly moldy, and severely moldy) models were established based on full wavelengths and selected wavelengths, respectively. The performance of all models was evaluated by considering some indicators such as 'Tolerability', ‘Stability’, ‘Accuracy’ and ‘Complexity’. It indicated that BOSS-SPA-PLS-DA and BOSS-SPA-LS-SVM were the optimal models for two-class and three-class classification with the overall accuracy of 99.76 % and 94.71 %, respectively. This study indicated that the online detection of 'Ya' pears with moldy core using the Vis/NIR full-transmittance spectra technology was feasible. Moreover, the proposed method has the potential to be used for online detection of early moldy core.
      原文链接:
    • NSTL
    • Elsevier

    Land suitability assessment for second cropping in terms of low temperature stresses using landsat TIRS sensor

    Rahimi-Ajdadi F.
    16页
    查看更多>>摘要：? 2022 Elsevier B.V.The second cropping which is done in the autumn after the main cultivation is one of the ways to increase food security, farmers' income, and thus prevent migration of villagers. The main limiting factor in autumn cultivation is the reduction in ambient and soil temperatures, which causes a severe drop in crop yield if it reaches below the physiological base temperature and freezing temperature. In Gilan, one of the northern provinces of Iran more than 90 % of agricultural land is cultivated once a year and has no production in half of the year. Accordingly, the purpose of the present study is to evaluate the feasibility of autumn cultivation of rice and hull-less seed pumpkin (Cucurbita pepo L. var. pepo subsp. styriaca) in terms low temperature stresses. In the first step, the land use map was prepared using the fusion of images with panchromatic band of Landsat 8 Maximum Likelihood classification method and then agricultural lands were separated. The land surface temperature (LST) maps were extracted on two dates, October 30, 2018 and December 17, 2019 using the TIRS sensor. Comparison of the estimated data with the meteorological data showed that LST was 0.6 °C higher than the meteorological data on October and it was 1.1 °C lower on December. Then, the LST map was validated with meteorological data in the last 10 years and a modified map was obtained to measure the cultivation potential of rice and hull-less seed pumpkin. Comparison of LST maps obtained on the two dates showed that the temperature variation was greater in the colder date. Considering base temperatures of 12 and 10 °C for rice and pumpkin, respectively, as well as the freezing temperature of 5 °C for these two crops, land suitability was identified. Having the minimum temperature of 15 °C for almost all lands, there will be no temperature stress for the crops until October 30. But, if the crop calendar lasts until December 17, only 2.62 % of agricultural lands have the potential to cultivate rice without temperature restrictions. Meanwhile, lands with cultivation potential for pumpkin were estimated 6983.3 ha (83.5 %) on December. Therefore, pumpkin cultivation has a lower temperature risk compared to rice in Rezvanshahr. The results of this study, in addition to being the required source for any planning in the field of autumn cultivation of rice and pumpkin, is one of the main sources for selecting different crops of the second cropping and their crop calendar.
      原文链接:
    • NSTL
    • Elsevier

    Predicting grain yield and protein content of winter wheat at different growth stages by hyperspectral data integrated with growth monitor index

    Liu S.Hu Z.Han J.Li Y....
    11页
    查看更多>>摘要：? 2022Hyperspectral reflectance data can detect a great detail for predicting wheat's grain yield (GY) and protein content (GPC). Based on different water and nitrogen rates, canopy spectral data and LAI and SPAD values were collected at four growth stages of wheat. The growth monitor index (GMI) was formulated by combining LAI and SPAD of wheat canopy with the variation coefficient method. Univariate models were constructed to predict GY and GPC in three forms, direct models predicted by spectral parameters, indirect models predicted with intermediate coefficient transmitted by GMI, and models predicted by GMI-treated spectral parameters. Multivariate models used single characteristic band and spectral reflectance indices (SRIs) of wheat canopy spectral reflectance. The results indicate that integrating GMI with canopy spectral data improves the prediction of spectral parameters for GY and GPC. The correlation coefficient between SAVIGMI and GY at the filling stage increased by 0.775 compared to SAVI; the correlation coefficient between SAVIGMI and GPC at the jointing stage is 0.645 higher than SAVI. By integrating GMI, the prediction model can explain over 90% of the variance in GY and 60% of the variance in GPC at the highest, the predictive performance is enhanced at some growth stages, maximally, R2 of the prediction based on GNDVIGMI for GY is 0.927 increased by 128%, and R2 of the prediction based on NDVIGMI for GPC is 0.532 increased by 130% relative to the original prediction at the filling stage. This study proposed the exploratory and experimental application of GMI to characterize canopy spectra.
      原文链接:
    • NSTL
    • Elsevier

    Multi-view density-based field-road classification for agricultural machinery: DBSCAN and object detection

    Zhang X.Jia J.Kuang K.Wu C....
    9页
    查看更多>>摘要：? 2022 Elsevier B.V.Field-road classification that automatically identifies the operation modes (either in-field or on-road) of GNSS (Global Navigation Satellite System) points plays an important role for the operational performance analysis of agricultural vehicles. Intuitively, a field often has high point density because in-field driving speed is rather low and the distance between consecutive strips is closed. In this paper, two methods were used to capture the in-field high-density characteristic: DBSCAN (Density-Based Spatial Clustering of Applications with Noise) and an object detection model. DBSCAN is a widely-used density-based clustering algorithm, which clusters the points with high point density into a cluster. Alternatively, a trajectory can be transformed into an image, and an object detection model can be applied to detect objects in the image, where an object is a set of pixels with high pixel density (i.e., a set of points with high point density). Finally, the two field-road classification results are combined using DBI (Davis Bouldin index), a metric which can evaluate the quality of either classification result. The developed method was validated by the harvesting trajectories of two crops (wheat and paddy), and the density-based field-road classification achieved 85.97% and 73.34% accuracy on the wheat data and the paddy data, respectively.
      原文链接:
    • NSTL
    • Elsevier

    Theoretical and experimental analyses of high-speed seed filling in limited gear-shaped side space of cotton precision dibbler

    Cai Y.Luo X.Hu B.Mao Z....
    14页
    查看更多>>摘要：? 2022 The AuthorsA novel state-of-motion adjustment method of seeds for high-speed seed filling in a cotton precision dibbler is proposed to solve the problems of arch lifting (multiple particles squeezing each other and causing congestion) and leakage (caused by exceedingly high relative linear velocity between the seed and seed tray during the seed-filling process). This method is based on the stress of the seed group and the limited gear-shaped side space (seed holding space). Kinematic and dynamic analyses of the main states of motion of the seed in the holding space were conducted, and the effects of the stress and movement states on seed-filling performance and automatic seed-cleaning performance were investigated. Through simulation, the velocity fluctuation timings of transverse and longitudinal rotation fillings in the holding space were compared. It was found that sufficient arrangement and a full state-of-motion adjustment on the seed-arranging surface was the key condition to ensuring that seeds fill the type hole by transverse rotation of the seed-filling surface. When seeds were not fully arranged and their state of motion not fully adjusted on the seed-arranging surface, the structural design of the type hole, seed-cleaning surface, and seed-filling surface affected the filling result of longitudinal rotation. The Box–Behnken centre combination method was implemented by setting the friction coefficient μ1 (X1), seed height z (X2), and rotation speed n (X3) as the factors with the qualified index A, leakage index M, and velocity fluctuation ΔH as the evaluation indexes. Multiple regression models and response surface optimisation analyses were performed to obtain the optimal combination of parameters affecting the dibbler seed-filling performance. The results indicated that the most significant factors affecting the A, M, and ΔH, were: z >n > μ1, z > n > μ1, and n >z > μ1. The optimal combination of these parameters was μ1 = 0.463, z = 0.4 kg, and n = 1.88 rps (approximately 9.78 km/h), providing a maximum A = 98.837 %, minimum M = 0.199 %, and ΔH = 0.111 m/s. When ΔH was in the interval of 0.09–0.18 m/s, A was high and stable, whereas M was small, indicating that the seed-filling performance was effective. When μ1 = 0.48, n = 1.5 rps (approximately 7.8 km/h), and z = 0.3 kg, the average values of the bench test were 91.07 % and 6.07 %, consistent with the simulation results. This study contributes significantly toward the research and development of the type hole of precise dibblers, high-speed seed selection method, orderly arrangement, migration, and quantitative separation method of irregular rotary material groups.
      原文链接:
    • NSTL
    • Elsevier

    Large scale pest classification using efficient Convolutional Neural Network with augmentation and regularizers

    Setiawan A.Yudistira N.Wihandika R.C.
    9页
    查看更多>>摘要：? 2022 Elsevier B.V.Insects are the most common type of animal on earth. In agriculture, insects are often referred to as pests because they are predators and parasitoids of plants. However, due to the large number of insects, it is challenging to identify the types of insects in order to determine the proper prevention. An artificial intelligence system via pest image recognition using CNN is expected to overcome this. Hence, prevention can be carried out more precisely for each insect type and thus saving costs and time. However, an efficient model with few parameters yet robust detection is required. Thus, it can be applied on mobile devices like smartphones or drones. Furthermore, a training procedure for maximizing the optimization results is required to achieve robust detection in insect classification. Insects usually appear in arbitrary positions, scales, and intensities, either with occlusion or without. The IP102 dataset is used as a benchmark for conducting training on the CNN algorithm. In this study, we propose an efficient training framework for optimizing small-sized models of MobileNetV2 using dynamic learning rate, exploiting CutMix augmentation, freezing layers, and sparse regularization. Combining those methods during training achieved the highest accuracy of 0.7132. The best hyper-parameters used during training are the AdamW optimizer configuration with the initial learning rate of 0.0001 and dropout of 0.2 on the linear layer. It outperforms several baseline models, which use a higher number of parameters.
      原文链接:
    • NSTL
    • Elsevier