Search

Weeds have been one of the main factors to affect the growth of crops in the seedling stage. Timely weeding is a necessary measure to ensure crop yield. An intelligent field weeding equipment can also be a promising potential deployment in the unmanned farm system at the current stage of intelligent agriculture. Effective recognition of crops and weeds has been a high demand to promote the development of intelligent weeding equipment. Previous research was focused mainly on object detection and semantic segmentation using deep learning. A great challenge is still remained in the performance of target detection, in the case of overlap images between the crops and weeds under the complex field. The reason was that the different target areas cannot be further divided when the generated anchor box overlaps in a large area. The pixel level annotation can also be required to train the semantic segmentation, where the data samples cannot be easy to obtain. The weak real-time performance cannot be conducive to practical application. In this study, an improved model was proposed using shifted window Transformer (Swin Transformer) network, in order to enhance the accuracy and real-time performance of crop and weed recognition. The specific procedure was as follows. 1) A semantic segmentation model of corn was established for the real and complex field scene. The backbone of the model was the Swin Transformer architecture, which was denoted by Swin-Base. The full self-attention mechanism was also adopted to significantly enhance the modeling ability in the Swin Transformer using the shift window division configuration. Self-attention was then calculated locally in the non-overlapping window of the segmented image block, where the cross-window connection was allowed. The computational complexity of the backbone presented a linear relationship with the image size, thereby elevating the inference speed of the model. The hierarchical feature representation was constructed through the Swin Transformer for the dense prediction of the model at the pixel level. 2) The Unified perceptual parsing Network (UperNet) was used as an efficient semantic segmentation framework. Among them, the feature extractor was the Feature Pyramid Network (FPN) using the Swin Transformer backbone. The multi-level features obtained by Swin Transformer were used by the FPN to represent the corresponding pyramid level. An effective global prior feature expression was added in the Pyramid Pooling Module (PPM). Better performance of semantic segmentation was achieved using the fusion of the hierarchical semantic information. The Swing transformer backbone and UperNet framework were combined into one model through the Decoder-Encoder structure, denoted by Swin-Base-UN. 3) The structure of the Swin-Base backbone was improved to enhance the inference speed. The number of network parameters and calculation cost were reduced to decrease the number of hidden layer channels, headers, and Swin Transformer blocks. Therefore, four improved models were generated, including the Swin-Large-UN, Swin-Small-UN, Swin-Tiny-UN, and Swin-Nano-UN. The model size and computational complexity of improved models were about 2, 1/2, 1/4, and 1/8 times of Swin-Base-UN, respectively. 4) Taking the segmentation of corn morphological region as the case study, an improved image morphological processing combination was established to recognize and segment all the weed regions in real time. The segmentation of corn was also used to segment the weeds. The weed pixel annotation was removed from the training data of the model. As such, a large number of annotation data at the pixel level was obtained in the semantic segmentation of the improved model, compared with the original one. A comparison was made on the performance of all models in training, validation, and testing. Consequently, the Swin-Tiny-UN was determined as the best model to achieve the optimal balance between accuracy and speed. Specifically, the mean Intersection over Union (mIoU) and mean Pixel Accuracy (mPA) on the test set were 94.83% and 97.18%, respectively, which increased by 3.27 and 4.71 percentage points, respectively, compared with the RestNet-101-UN using traditional Convolutional Neural Networks (CNN) backbone. The inference speed of the model was achieved by 18.94 frames/s. The best model of semantic segmentation was superior to the traditional one, in terms of the region segmentation accuracy, pixel recognition accuracy, and inference speed. The image segmentation showed that the improved model can be expected to accurately recognize and segment maize and weeds in complex field scenes. The average correct detection rate of the improved model was 95.04% for the video stream data in the process of field work, whereas, the average detection time per frame was 5.51×10-2 s. Consequently, the improved model can be expected to detect the corn and weeds in the process of field work, indicating higher accuracy and real-time performance under practical application conditions. The findings can provide a strong reference for the development of intelligent weeding equipment. [ABSTRACT FROM AUTHOR]

Weed control is an inevitably necessary task in field management. Effective recognition of crops and weeds has therefore been an essential basis to promote the development of intelligent weeding equipment. Nevertheless, the recognition targets are not fixed in images except for crops, due mainly to the variety of weeds and random distribution of their positions. It is highly demanded for better recognition performance to detect the crops from all categories of weeds in the image. All weed targets are required to be labeled in the dataset, where there are comprehensive all-inclusive weed species. However, human vision can only identify the target crops from the weeds. The species and quantity of weeds are still lacking in the identification. Moreover, the crops and weeds are usually overlapped in the field images with complex scenes. It is also difficult to accurately segment the boundary of various objects, especially when the generated anchor box was superimposed by a large area in deep overlapping. In this study, a recognition and semantic segmentation of maize at the seeding stage was proposed to identify the weeds on the premise of maize recognition using a dual attention network. Fine segmentation of morphological boundary was obtained. The main contents were as follows. 1) The original architecture of the model was determined to compare 6 state-of-the-art semantic segmentation networks. It was found that the architecture of the dual attention network presented the best performance for the training, validation, and testing dataset, thereby realizing the pixel-wise recognition and segmentation of maize field images. In the validation set, the mean intersection over union (mIoU) and mean pixel accuracy (mPA) at the end of iteration were 92.73% and 96.88%, respectively. In the test set, the mIoU and mPA were 92.8% and 94.66%, respectively, and the speed of segmentation was 15.2 frames/s. 2) The semantic segmentation model of maize at the seeding stage was established using the improved network architecture. The function of the model was a binary classification of maize pixels and all of the other pixels, suitable for the recognition and morphological segmentation of maize in complex field scenes at the seeding stage. The improved backbone was used to enhance the feature representation. More details of features were retained, while the amount of computation was reduced. Recurrent criss-cross and channel attention modules were combined to compose a dual attention mechanism, in order to synchronously construct long-range contextual dependencies in spatial and channel dimensions of the feature map. The discriminability of feature representation was improved significantly. The encoder-decoder structure was used to build the model, and then the auxiliary head was attached to optimize the underlying features. The loss function was improved, while the transfer learning strategy was formulated. 3) The segmentation map of weeds was obtained via image morphological processing on the segmentation map of maize at the seeding stage. The regions of weed were identified by the segmentation of maize, particularly without considering the pixel-wise prediction of the weed region. The results showed that the performance of the model was better than the original network in the whole training process. At the end of the iteration, the mIoU and mPA were 93.98% and 97.48%, increasing 1.35% and 0.62%, respectively, compared with the original network. There was an obvious increase in the accuracy of region segmentation, the accuracy of pixel recognition, and segmentation speed, indicating better comprehensive performance of the model. The mIoU and mPA of the test set were 94.16% and 95.68%, exceeding the baseline by 1.47% and 1.08%, respectively. The speed of segmentation achieved 15.9 frames/s, which was increased 4.61% compared with the original network. The finding can provide a promising reference for the development of intelligent weeding equipment, thereby accurately recognizing and segment the maize and weeds at the seeding stage in complex field scenes. [ABSTRACT FROM AUTHOR]

Effective recognition method of crop and weed is the basis for promoting the development of intelligent mechanization weeding pattern. Summarizing the previous research, we found that the accuracy and stability of the recognition model could be improved by natural and diversified feature presentation, but there are still 2 main problems. On the one hand, feature presentation of the natural property of target was difficult to be obtained by the hand-engineered feature extractor. The spatial consistency of the obtained features was bad, and the real-time performance of recognition system was reduced for the complex feature extraction algorithm. On the other hand, the effect of image preprocessing has important influence on recognition results, especially the overlapping segmentation of crop and weed. For overlapped objects, it is usually difficult to segment them without affecting their respective feature presentations, resulting in low recognition accuracy and stability. In order to solve the main problems in the current research, we explored the way to improve the recognition accuracy, stability and real-time performance, and a recognition method of crop and weed based on multiscale hierarchical feature learning combined with superpixels segmentation was proposed. The main research contents of this paper were as follows: 1) Excellent internal features of image are hierarchical. In this research, the multiscale hierarchical feature is a scene level feature with invariance and consistency in scale space. Multiscale convolutional neural network was built to extract multiscale hierarchical feature. Multiscale convolutional neural network contains multiple copies of a single CNN (convolutional neural network) that are applied to multi-scale Gaussian pyramid of the input image. The CNN model as feature extractor in this paper includes 3 stages. In the first 2 stages, it contains a bank of filters (convolution kernels) to produce dense feature maps, a point to point nonlinear mapping activation function, and a spatial pooling module for sub-sampling of each feature map, While the last stage only contains a bank of filters. Each filter is applied to the input feature maps through 2-dimensional convolution operation, in which local feature presentations are detected at all pixel locations on the input image. For each pixel, the CNN model is used to collectively encode the internal information in a large sense region around the location of given pixel. The CNN is fed with raw pixels and trained with back propagation method. With complete training, this CNN model can automatically extract hierarchical feature representations from the input image, thereby decreasing the need for hand-engineered features extracting. A series of feature maps for multiscale regions centered on each pixel of the image are produced by the multiscale convolutional neural network; these representations contain shape, texture and sense information. Therefore, the multiscale hierarchical feature is learned to allow the recognition of the class of all pixels in the image. The average pixel recognition rate is 93.41%. 2) In this research, multiscale hierarchical feature is used to produce the class distribution for every pixel through a 2-layer MLP (multi-layer perceptron). But recognizing the class of pixel independently from its surrounding regions could produce some interference at the boundary of targets. Accurate boundary segmentation for each target in the image is not provided. In this paper, the superpixel method is used for generating an over segmentation of the image, where each segmentation component (superpixel) is an irregular pixel block composed of similar pixels in texture, color and brightness characteristics. Object in the image is composed of these superpixels, and has the exact original boundary. The superpixels segmentation is computed following the simple linear iterative clustering method described in this paper. 3) We proposed the recognition strategy of multiscale hierarchical feature learning combined with the superpixels segmentation. Firstly, an over segmentation of the original image was produced through the superpixels segmentation. At the same time, each pixel location of the image was classified densely based on the multiscale hierarchical features. These predictions of pixels in each superpixel were aggregated to produce the class prediction of superpixel, through computing the average class distribution within the superpixel. Adjacent superpixels with the same class were merged to obtain the final target class prediction and image segmentation. The accurate image segmentation was achieved while recognizing the target in the image by this recognition method, which effectively avoided the problems caused by targets overlapping, and the recognition results were more stable and accurate. Maize seedlings at 2-5 leaves stage and weed during the same stage were used as research object, and the recognition method of multiscale hierarchical feature learning combined with superpixels segmentation was tested. The results showed that the average target recognition rate with this method was 98.92%, and the standard deviation was 0.55%. The average target recognition rate with the method in previous research was 98.36%, and the standard deviation was 1.05%. So, the accuracy and stability of recognition results by this method in this paper were improved with different degrees. In aspect of real-time performance, the average time to recognize a single image was 1.68 s with this method, and compared with the method proposed in previous research, it was decreased by 1.58 s. The real-time performance of this method can be further improved by GPU (graphics processing unit) hardware acceleration, and the average time to recognize a single image was only 0.72 s. Therefore, the recognition method based on multiscale hierarchical feature learning combined with superpixels segmentation can effectively achieve accurate, stable and efficient recognition of maize and weed. The research results provide reference for the development of precision weeding. [ABSTRACT FROM AUTHOR]

High pulsed electric field (HPEF) pretreatment is helpful to induce breakdown and displacement effect in cell of fruits and vegetables, which is useful to improve drying speed. We should study the effect of HPEF on the physical properties of fruits and vegetables, which puts the base for optimizing the parameters and mechanism analysis. In this study, 2 kinds of fruits and 2 kinds of vegetables, apple, pear, potato and radish were chosen as raw material. The moisture contents of apple, pear, potato and radish were approximately 86.89%, 89.46%, 78.35% and 93.61%, respectively. Immediately before measuring, the fruits and vegetables were washed and peeled by hand, and then cut into 13 mm diameter and 5 mm thickness using a cork borer. High pulsed electric field generator was used for the HPEF pretreatment of the fruits and vegetables, and the parameters were set as follows: 1) 1 000 V/cm pulse strength, 60 ?s pulse length, pulse number of 15, 2) 1 250 V/cm pulse strength, 90 ?s pulse length, pulse number of 30, and 3) 1 500 V/cm pulse strength, 120 ?s pulse length, pulse number of 45. After HPEF pretreatment, the dynamic viscoelasticity behavior was performed with DMA testing machine using compression platen (diameter of 15 mm). The linear viscoelastic range (LVR) was determined through stress sweeps. A constant amplitude of 10 μm was then chosen (common to the LVR of all samples) to obtain the mechanical spectra as frequency sweeps from 1 to 100 Hz at 25 ℃ and temperature sweeps from 25 to 90 ℃ at 10 Hz. The test and analysis of dynamic viscoelastic of fruits and vegetables were performed by TA Universal Analysis software. Analysis of variance was applied to the data. In order to study the differences among samples, all the data were analyzed with the use of SAS (statistical analysis system) software (version of 9.2). The microstructure analysis was performed by scanning electron microscopy (SEM) on apple samples with the 3 sets of pretreatment parameters, and untreated samples. The storage modulus, loss modulus and loss tangent on different HPEF pretreatment parameters and different oscillation frequency were plotted to characterize the viscoelastic behavior. The theoretical analysis about dynamic viscoelastic mechanics was presented. Results showed that all samples showed a slight linear increase in storage modulus, loss modulus and loss tangent as frequency diminished. A statistically significant difference in storage modulus, loss modulus and loss tangent between different varieties was confirmed. For apple samples, the HPEF pretreatment led to the decrease of storage modulus and loss modulus by 72.36% and 78.09%, respectively. It was primarily attributed to the changes of the cell structure and the intracellular space of the cell as revealed by the SEM. The same decreasing tendency of storage modulus and loss modulus with the increasing of HPEF parameters was observed for potatoes. However, the radish showed a contrary tendency that the storage modulus and loss modulus of radish tissue increased with the increasing of HPEF parameters. High cell turgor pressure may be the main reason for the increase in radish dynamic viscoelasticity after HPEF pretreatment. To pears, as HPEF pretreatment parameters were 1 000 V/cm pulse strength, 60 ?s pulse length, and pulse number of 15, the storage modulus and loss modulus of pear tissue reached the maximum. The reason may be the changes in pear cell turgor pressure caused by low HPEF pretreatment, and then with the further increasing of the HPEF parameters, pear cell membrane could be damaged and the dynamic viscoelasticity of pear tissue decreased. With the temperature increasing, the storage modulus of apple, pear, potato, and radish decreased by 34.44%±3.47%, 49.04%±10.64%, 28.12%±10%, and 55.25%±5.61%, respectively. The loss modulus of them decreased by 66.49%±8.05%, 82.12%±15.12%, 44.66%±7.79%, and 67.84%±2.54%, respectively. These results can provide useful information for the optimization of drying process parameters and the improvement of the industrial pretreatment technique in vacuum freeze-drying processing. [ABSTRACT FROM AUTHOR]

In order to clarify the rheological mechanical properties of millet grain group during mechanical harvesting and processing, this paper studied the effects of different varieties and different moisture contents on the creep characteristics of millet grain group. In this experiment, different varieties (Jingu 21, Zhangza 10), and different moisture contents (12.10%, 16.05%, 20.00%) of millet grain group were selected as creep experimental factors. The creep characteristics (instantaneous elastic modulus, delayed elastic modulus, relaxation time and viscosity coefficient) of millet seed group were used as experimental indicators, and one-way analysis of variance and multi-factor analysis of variance were performed. Based on the original contents of millet seed group, the target moisture contents of the millet grain group were formulated to be 12.10%, 16.05%, 20.00%. The experiment was conducted on the CMT-6104 universal testing machine, and the special fixture for testing the creep characteristics of millet grain group was designed. Combining the Burgers model, nonlinear programming method was used to optimize search regression parameters, and creep parameters of millet grain group were obtained. The creep curve of millet grain group of different varieties showed that, as time going on, the amount of deformation caused by creep tended to be constant, which reaching the strain saturation. There were differences in the millet grain group creep characteristic curves of the different varieties. The significance test standard was 0.05, and the variance analysis was performed. The results showed that the creep characteristics of millet grain group could be described by the four-element Burgers model, and the creep parameters of millet grain group with different moisture content and different varieties were different. The variety had a significant effect on the delayed elastic modulus of the millet grain group. The delayed elastic modulus of the Jingu 21 millet grain group was higher than that of Zhangza 10. The moisture content had significant effects on the instantaneous elastic modulus, delayed elastic modulus and viscosity coefficient of the millet grain group, which was decreasing while the moisture increasing. The instantaneous elastic modulus of the millet grain group with the water content of 12.10% was significantly higher than the water content of 16.05% and 20.00%. The delayed elastic modulus of the millet grain group with the moisture content of 12.10% was significantly higher than the water content of 20.00%. The instantaneous elastic modulus and the delayed elastic modulus represented elasticity and hardness of the millet grain group, respectively. With the increasing of the moisture content of the millet grain group, the instantaneous elastic modulus and the delayed elastic modulus of millet grain group decreased, resulting in a high increasing of the creep curve. The strain in the screw conveyor of combine harvester was correspondingly larger. In this paper, the creep properties of millet grain group with different varieties and different moisture contents were studied experimentally, which provided theoretical support for low-loss harvesting, processing and storage and parameter optimization of millet. There are many factors affecting the mechanical properties of the material, in addition to the influence of moisture contents and varieties, it is also related to many factors such as temperature, internal molecular structure and chemical composition of the material, et al. The rheological properties of agricultural materials need to be further studied in conjunction with the principles of life sciences. [ABSTRACT FROM AUTHOR]

Millet contains protein, vitamin B1 and other mineral elements needed for human health, which has functions of enhancing blood circulation, brain tonic, and sleep quality. In recent years, the market demand for millet is rising. However, the mechanization level of millet harvesting is low. Manual harvesting is very common, which requires high labor intensity, has low efficiency and results in insufficient supply of millet. In this paper, we systematically investigated effects of moisture content and varieties of millet on its physical parameters, effects of millet varieties and compression azimuth on the compression mechanical properties (including yield load, deformation, failure energy and apparent elastic modulus) of millet, and effects of compression azimuth on crack formation of millet. The study results are expected to provide the basis for the design and optimization of millet harvester and related processing machinery. In this paper, the 1000-grain weight was measured by an electronic balance, the triaxial sizes by a vernier caliper, and the moisture content by fast grain moisture analyzer for four millet varieties: Zhangza-10, Dungu, Jingu-21, Changgu. The sphericity of the millet, the arithmetic mean diameter and the geometric diameter were calculated using the triaxial sizes. The compression tests were carried out along X, Y, Z axis using a texture analyzer and the data were analyzed by SAS statistics software. The yield load, deformation, and failure energy were obtained from the force-distance curve measured by a texture analyzer. At the same time, the apparent elastic modulus was calculated according to the Hertz formula. The experimental results were summarized as follows in six categories. 1) For the physical parameters of the millet, with a moisture content between 15.2% and 25.1%，the length, width, height, 1 000-grain weight, average diameter and geometric mean diameter of millet grain decreased as the moisture content decreased. The millet length, width and thickness range from 2.028 to 2.112 mm, 1.818 to 1.931 mm, 1.413 to 1.513 mm, respectively. The sphericity of millet was more than 84%; the arithmetic mean diameter varied from 1.750 mm to 1.852 mm; and the geometric mean diameter 1.732 mm to 1.834 mm. Therefore, the shape of millet can be regarded as ellipsoid. There were significant differences (P<0.001) in the triaxial sizes, 1000-grain weight, sphericity, arithmetic mean diameter and geometric mean diameter among different varieties. When the moisture content of millet was about 25%, the 1000-grain weight of Jingu-21 was 3.473 g, which was the heaviest millet, while Zhangza-10 was 3.264 g, which was the lightest millet. The moisture content has a negligible effect on the sphericity. The effect of varieties on sphericity was significant (P<0.001). 2) For factors on the compression mechanical properties, the influence of moisture content, compression azimuth, and variety on the compression mechanical properties were significant (P<0.01). The significant factors affecting the compression mechanical properties were compression azimuth, moisture content, and variety in a sequence from the most significant to the least significant. 3) For the effects of the moisture content, for the same kind of millet, the yield load and apparent elastic modulus decreased as the moisture content increased during compression in X, Y, Z axis. The yield load decreased slowly when moisture content ranged from 20.9% to 25.1%. The deformation and failure energy first decreased and then increased as the moisture content increased. For the same azimuth compression, the better the quality of the millet, the more compact the internal structure, and the smaller the deformation. 4) For crack formation of the millet: on the X axial, a crack started from the crest of the grain along the junction of embryo and endosperm, and then extended throughout the X axis. On the Y axis, a crack started from the connection point of grain and probe and the connection point of grain and base throughout the Y axis because of stress concentration. On the Z axis, a crack started from the junction of embryo and endosperm on the back to the ventral hilum along the weakest part of endosperm resistance extending throughout half of the grain. 5) For the effects of compression azimuth, the yield load increased in turn on the X, Y, and Z axis and the deformation and failure energy decreased on the X, Z, and Y axis in turn. The apparent elastic modulus of the millet increased in turn on the X, Y, and Z axis. This result indicated that millet was not uniform and it was an anisotropic body. 6) For the effects of millet varieties, with the same moisture content, the compression mechanical properties of Jingu-21, Dungu, Changgu and Zhangza-10 varied from large to small. This indicated that Jingu-21 had a relatively good compression resistance and quality. [ABSTRACT FROM AUTHOR]