Computers and Electronics in Agriculture最新文献

Rice panicle density is one of the essential bases for the automatic speed regulation of unmanned harvesters, making density detection crucial for intelligent upgrades. Currently, existing methods for detecting rice panicle density do not meet actual harvesting scenarios and struggle to meet real-time requirements. To address this, we developed a real-time rice panicle density detection method for unmanned harvesters. This method includes a panicle detection model based on YOLOv5n (RP-YOLO) and a rice panicle density calculation based on coordinate transformations. RP-YOLO was optimized through various techniques, such as enhancing the target detection head, reconfiguring the backbone network and downsampling module, introducing an attention mechanism, and refining the loss function. Based on coordinate conversion, we converted the world coordinates of the detection frame vertex to image coordinates and calculated the panicle density. We established the RP-1668 dataset for japonica rice and trained and tested the model. Compared to the original YOLOv5n model, our modifications reduced floating-point operations per second (FLOPs) by 33.33 %, decreased model size by 31.90 %, increased detection speed by 12.63 %, and improved accuracy (AP0.5) by 3.82 % (AP0.5:0.95, 6.96 %). RP-YOLO achieved superior accuracy and detection speed compared to both conventional lightweight and non-lightweight models. In field applications, the error in density detection was less than 10 % compared to manual counting, and the results clearly reflected changes in rice panicle density. For a 1.4 m × 1.0 m rice field imaging area (with a resolution of 2560 × 1280), the method detects at 15 fps on an on-board industrial computer, providing reliable data support for adjusting the operating speed of driverless harvesters.

A novel design methodology for tubular photobioreactors (TPBRs) for the mass culture of microalgae that can account for mixing intensity is presented. To date, TPBRs have been mainly designed and operated under the assumption of perfect mixing with regard to photosynthesis performance (light integration regime). In this work we show that this simplification has been leads to significant errors in the prediction of the optimal dilution rate and biomass productivity. To this end, computational fluid dynamics and light distribution model have been employed to calculate the trajectories and light histories I(t) of a microalgal cell population represented by 50 particles of 5 μm diameter. The density of the microalgal cells was set at 1000 kg m⁻³ and the tube diameters (D) were 14, 24, 44, 64 and 84 mm, with the circulation velocities (v) ranging from 0.4 to 1 m s^-1. This has been coupled to a dynamic photosynthesis model in order to calculate the average photosynthetic response and hence the integration factors in TPBRs. It has been demonstrated that for a generic microalgal strain, the use of the light integration simplification (Γ = 1) would result in the prediction of an optimal dilution rate of 0.0315 h⁻¹ (for D = 14 mm and v = 0.4 m/s as an example), which would lead to an actual biomass productivity of 182.5 g biomass m⁻³h⁻¹ if the predicted integration factor (Γ = 0.578) is used whereas the newly proposed method predicts an optimal dilution rate of 0.0125 h⁻¹ and a biomass productivity of 362.3 g biomass m⁻³h⁻¹. This demonstrates that simplifying the light integration regime is inadequate for TPBRs design and operation, resulting in significant inaccuracies. The estimation charts and regressions proposed in this work to estimate actual integration factors will enable the development of an optimization method for TPBRs based on mixing intensity.

Hyperspectral satellite imagery has significant advantages in rapidly monitoring soil organic matter (SOM) content over a large area. However, limitations in the timely acquisition of site-specific data may affect its effectiveness due to weather influences and revisit cycles. This paper proposes a novel method for estimating SOM content that combines the high temporal resolution of the Zhuhai-1 hyperspectral satellite image and Sentinel-2A multispectral satellite image to broaden the spectral range of Zhuhai-1 images. Multisource features, including spectral bands, topographic features, textural features, and spectral indexes, are extracted from the combined image and digital elevation model. An improved genetic algorithm (IGA) is proposed to optimize feature selection and extreme gradient boosting (XGBoost) is then applied to estimate SOM content. The proposed method was validated using 197 topsoil samples and satellite images collected from a demonstration area in Lishu County, Jilin Province, China. The results indicate that the estimation accuracy using the combined image was greater than using one single image. Compared with only using the Sentinel-2A and Zhuhai-1 images, the coefficient of determination (R²) values improved from 0.61 and 0.73 to 0.82, the ratio of the prediction to the deviation (RPD) values improved from 1.62 and 1.93 to 2.35, and the ratio of performance to the interquartile distance (RPIQ) values improved from 2.16 and 2.58 to 3.15, respectively. Furthermore, the XGBoost algorithm outperformed the random forest algorithm in terms of model accuracy and mapping reliability. The use of multisource features improved the R² value from 0.45 to 0.82 compared to only using 31 spectral bands of Zhuhai-1 and Sentinel-2A image, with contribution rates in descending order of spectral indexes (48.2%), topographic features (24.7%), spectral bands (19.9%), and textural features (7.2%). This paper thus presents a promising method for efficient periodic mapping of the SOM content by combining data from a hyperspectral satellite constellation and a multispectral satellite image.

This study demonstrates how the availability of frequent Synthetic Aperture Radar (SAR) observations has transformed crop yield prediction, a critical component of food security and agricultural practices. SAR observations along with the climatic variables are integrated into an advanced deep learning technique for predicting crop yield. Capitalizing on the unique advantages of high-resolution SAR images, including consistent acquisition schedule, and not being affected by cloud cover and variations between day and night, this research explores new potential in agriculture. Deep learning due to its ability to discern both spatial and temporal relationships within SAR data, captures the salient features from SAR observations to predict the yield of Michigan’s non-irrigated Corn, Soybean, and Winter Wheat.

We employed advanced deep learning and established machine learning techniques including patch-based 3D Convolutional Neural Networks (3D-CNNs), Random Forest, Support Vector Machine, and XGBoost to significantly improve the accuracy of yield estimation.

Spanning eight years from 2016 to 2023, our analysis underscores the exceptional potential of VH channel of Sentinel-1 SAR data for near accurate yield prediction. Among the methods tested, XGBoost consistently surpassed others in crop yield estimating accuracy, particularly in scenarios with limited reference data. Patch-based 3D-CNNs also demonstrated a remarkable ability to approximate XGBoost’s performance, albeit with a streamlined set of input features. Our study further illuminates the delicate balance required in selecting SAR data resolution, demonstrating the need for careful compromise between reducing noise and preserving crucial data intricacies. Notably, our predictive models showcased formidable precision, predicting yields with a mere 7.5% margin of error a full month prior to harvest. These compelling findings signal the need for continued innovation and integration of deep learning technologies, calling for the enrichment of yield datasets to realize more comprehensive and pinpoint-accurate yield predictions.

Sensors implemented in agriculture play a significant role in soil and plant growth and enable real-time physical and chemical interactions in the environment, such as temperature, moisture/humidity, pH, and contaminant levels. Additionally, these sensors provide essential data that can enhance crop growth scenarios, resist biotic and abiotic stresses, and improve crop production. This article provides a thorough examination of the evolving landscape of agricultural sensor technologies, perspectives, and challenges in the field. Currently, some of the key soil sensors used in agricultural programs, include those that measure moisture, temperatures, pH, organic matter components, insects, and soil pollutants. On the other hand, nanobiotechnology sensors implement optical, wireless, or electrical signals to provide information about plant signaling molecules related to the conditions of agronomic equipment. We shed more light on the use of nanomaterial-facilitated transport of genetically encoded sensors as devices for the investigation and advancement of advanced plant sensors. Innovative technologies, including wireless sensor networks and plant wearables, are also addressed with regard to their potential for precision agriculture. The paper concludes by presenting future perspectives and difficulties in the fields of soil sensors and intelligent agriculture. In summary, we provide a comprehensive and forward-looking perspective on the potential of nanotechnology to facilitate the development of intelligent plant sensors. These sensors are capable of communicating with and controlling electrical equipment, with the aim of tracking and improving the output and resources applied to individual plants.

Cardiovascular diseases (CVD) in animals can severely impact the heart and circulatory systems, like those in humans. Early diagnosis and treatment are crucial for improving animal welfare and lifespan. Traditional diagnostic methods face challenges such as insufficient anamnesis information, high costs of biochemical and hematological tests, and increasing data complexity. This study aims to address these issues by developing AI-based diagnostic systems for fast and accurate CVD diagnosis in cattle using retinal images. A total of 1118 retinal images from 100 cattle were collected, with 52 diagnosed with CVD and 48 as non-CVD. The dataset is publicly available on Kaggle. We evaluated three machine learning methods (Extreme Learning Machine, K-Nearest Neighbors, and Support Vector Machine) and four deep learning models (DenseNet201, ResNet101, SqueezeNet, and InceptionV3) for their diagnostic capabilities. ResNet101 emerged as the most effective model, achieving an accuracy of 96.1 ± 3.15 %, sensitivity of 97.3 ± 2.96 %, specificity of 94.9 ± 4.07 %, and an F1-score of 96.4 ± 0.03. This study demonstrates that AI-based systems, particularly deep learning models, can significantly improve the accuracy of CVD diagnosis in animals, marking a critical advancement in veterinary healthcare.

Apple is one of the most popular fruits. In-field grading and sorting of apples would enhance growers’ economic benefits by lowering production costs. This study reviews the key components and progress of the quality inspection algorithm for in-field grading and sorting of apples. Four key components (e.g., conveyor, imaging chamber, sorting actuator, and bin filler) are presented in detail, followed by summarizing the shortcomings of these components. The apple’s external (color, size, and defects) and internal quality inspection technologies, such as optical technologies of visible light, near-infrared (NIR), hyperspectral/multispectral imaging (HSI/MSI), and structured illumination (SI) were presented. Despite the excellent detection performance of emerging technologies (e.g., HSI, MSI, and SI), visible light is still dominantly used for in-filed grading. Challenges in getting information on the whole surface area of the apple, uneven lighting, machine size, throughput, and associated costs hamper the commercialization of apple in-field grading and sorting equipment. At present, more efforts should be devoted to internal quality inspection, by developing reliable, fast, and accurate detection equipment and algorithms. With the advancement of sensors and automation algorithms, as well as the emergence of mechanical systems that are suitable for in-field use, it is anticipated that the apple in-field grading and sorting equipment that inspects both external and internal quality will be realized and commercialized in the near future.

Deep learning approaches have provided state-of-the-art performance in crop mapping. Recently, several studies have combined the strengths of two dominant deep learning architectures, Convolutional Neural Networks (CNNs) and Transformers, to classify crops using remote sensing images. Despite their success, many of these models utilize patch-based methods that require extensive data labeling, as each sample contains multiple pixels with corresponding labels. This leads to higher costs in data preparation and processing. Moreover, previous methods rarely considered the impact of missing values caused by clouds and no-observations in remote sensing data. Therefore, this study proposes a lightweight multi-stage CNN-Transformer network (MCTNet) for pixel-based crop mapping using time-series Sentinel-2 imagery. MCTNet consists of several successive modules, each containing a CNN sub-module and a Transformer sub-module to extract important features from the images, respectively. An attention-based learnable positional encoding (ALPE) module is designed in the Transformer sub-module to capture the complex temporal relations in the time-series data with different missing rates. Arkansas and California in the U.S. are selected to evaluate the model. Experimental results show that the MCTNet has a lightweight advantage with the fewest parameters and memory usage while achieving the superior performance compared to eight advanced models. Specifically, MCTNet obtained an overall accuracy (OA) of 0.968, a kappa coefficient (Kappa) of 0.951, and a macro-averaged F1 score (F1) of 0.933 in Arkansas, and an OA of 0.852, a Kappa of 0.806, and an F1 score of 0.829 in California. The results highlight the importance of each component of the model, particularly the ALPE module, which enhanced the Kappa of MCTNet by 4.2% in Arkansas and improved the model’s robustness to missing values in remote sensing data. Additionally, visualization results demonstrated that the features extracted from CNN and Transformer sub-modules are complementary, explaining the effectiveness of the MCTNet.

Effectively improving the detection accuracy of wheat chlorophyll content is of great significance for the detection of photosynthetic capacity and growth status of wheat canopy. However, due to inhomogeneous light distribution issues in canopy, the existence of shaded and sun wheat leaves in wheat canopy has influence for spectral-based detection of chlorophyll content. Therefore, in order to improve detection of wheat canopy chlorophyll content, a light distribution correction method was proposed to correct intact leaves’ light distribution based on shaded and sun leaves in multispectral images. Firstly, R-G-NIR images were reconstructed to segment and analyze shaded and sun leaves of wheat. Secondly, Homomorphic Filter (HF) and Gamma light correction method was used to optimize shaded leaves’ light distribution. Then, the differential responses of 10 different types of vegetation indices in shaded, sun, original intact and corrected intact leaves were analyzed to screen chlorophyll-sensitive parameters based on Random Frog Method (RFM). Finally, Random Forest (RF), Support Vector Regression (SVR), and Partial Least Squares Regression (PLSR) were used to establish models for the detection of chlorophyll content in shaded, sun, original intact and corrected intact leaves of wheat, respectively. The results showed that the proposed light correction method reduced the inhomogeneous light and kept more uniform. Normalized difference vegetation index (NDVI), green normalized difference vegetation index (GNDVI), difference vegetation index (DVI), normalized redness intensity (NRI), and optimized soil-adjusted vegetation index (OSAVI) were selected as the optimal spectral variables. And the models after correction had higher accuracy than the models before correction. The wheat chlorophyll content model of corrected intact leaves based on RF had the highest accuracy, with a calibration set $R_c^2$ of 0.816, ${RMSE}_c$ of 3.702, a validation set $R_v^2$ of 0.804, ${RMSE}_v$ of 3.958, respectively. The research integrates the above results to improve detection of wheat canopy chlorophyll content, which provides technical support for the light distribution correction in multispectral images.

Fresh fruits and vegetables are characterized by high water content, perishability and strong seasonality. In the process of supply chain, it is easy to lead to the decrease of freshness. Artificial intelligence (AI) technology is considered as an effective new means to assist the preservation of fruits and vegetables. It is used in the post-harvest, storage and cold chain transport stages of the fruits and vegetables supply chain. This paper reviews several advanced AI technologies and systematically introduces the application of these technologies in fruits and vegetables preservation. AI can achieve quality assessment, cold chain environment monitoring, and shelf-life prediction for fruits and vegetables, optimize supply chain management, and improve product traceability. The new fruits and vegetables preservation strategy based on AI has greatly maintained the product quality of fruits and vegetables. In addition, the current limitations and future development trends of AI technology in fruit and vegetable preservation are also discussed. This paper aims to provide guidelines for applying AI technology in fruit and vegetable preservation and provide new ideas for the future fruit and vegetable preservation process to make it more efficient and intelligent.