CN117743975A - Methods for improving soil environment on hillside cultivated land - Google Patents
Methods for improving soil environment on hillside cultivated land Download PDFInfo
- Publication number
- CN117743975A CN117743975A CN202410190945.8A CN202410190945A CN117743975A CN 117743975 A CN117743975 A CN 117743975A CN 202410190945 A CN202410190945 A CN 202410190945A CN 117743975 A CN117743975 A CN 117743975A
- Authority
- CN
- China
- Prior art keywords
- crop growth
- soil
- feature
- growth state
- vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 239000002689 soil Substances 0.000 title claims abstract description 270
- 238000000034 method Methods 0.000 title claims abstract description 50
- 239000013598 vector Substances 0.000 claims abstract description 214
- 230000004927 fusion Effects 0.000 claims abstract description 54
- 230000003993 interaction Effects 0.000 claims abstract description 53
- 238000003973 irrigation Methods 0.000 claims abstract description 28
- 230000002262 irrigation Effects 0.000 claims abstract description 28
- 238000010219 correlation analysis Methods 0.000 claims abstract description 11
- 238000005728 strengthening Methods 0.000 claims abstract 12
- 239000011159 matrix material Substances 0.000 claims description 35
- 238000013527 convolutional neural network Methods 0.000 claims description 12
- 238000000605 extraction Methods 0.000 claims description 10
- 238000011176 pooling Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 claims 1
- 230000017105 transposition Effects 0.000 claims 1
- 230000006872 improvement Effects 0.000 abstract description 13
- 230000008859 change Effects 0.000 abstract description 7
- 230000002452 interceptive effect Effects 0.000 description 19
- 238000009826 distribution Methods 0.000 description 9
- 238000012545 processing Methods 0.000 description 7
- 238000012937 correction Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012271 agricultural production Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 231100000502 fertility decrease Toxicity 0.000 description 1
- 230000003862 health status Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000004162 soil erosion Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Description
技术领域Technical field
本发明涉及农业管理技术领域，且更为具体地，涉及一种山坡耕地土壤环境改良方法。The present invention relates to the technical field of agricultural management, and more specifically, to a method for improving the soil environment of cultivated land on a hillside.
背景技术Background technique
山坡耕地是指在一定坡度的山地上开垦的农田，通常分为梯田和坡耕地两种。山坡耕地是农业生产的重要组成部分，但由于地形、气候等多种因素的影响，其土壤环境往往面临诸多问题，如土壤侵蚀、肥力下降、水分不足等，这些问题直接影响到农作物的生长和产量。因此，改良山坡耕地的土壤环境对提高农业生产效益具有重要意义。Hillside farmland refers to farmland cultivated on mountains with a certain slope, and is usually divided into two types: terraced fields and slope farmland. Hillside cultivated land is an important part of agricultural production. However, due to the influence of terrain, climate and other factors, its soil environment often faces many problems, such as soil erosion, reduced fertility, insufficient water, etc. These problems directly affect the growth and development of crops. Yield. Therefore, improving the soil environment of hillside cultivated land is of great significance to improving agricultural production efficiency.
在改良山坡耕地土壤环境的过程中，合理灌溉是关键环节。然而，传统的灌溉方法往往依赖于人工经验判断，缺乏科学依据，难以实现精准灌溉。同时，由于地形、气候等条件的复杂性和多变性，人工判断的准确性和效率往往受到限制。因此，期待一种优化的山坡耕地土壤环境改良方法。In the process of improving the soil environment of hillside cultivated land, rational irrigation is a key link. However, traditional irrigation methods often rely on artificial experience judgment and lack scientific basis, making it difficult to achieve precise irrigation. At the same time, due to the complexity and variability of terrain, climate and other conditions, the accuracy and efficiency of manual judgment are often limited. Therefore, an optimized soil environment improvement method for hillside cultivated land is expected.
发明内容Contents of the invention
为了解决上述技术问题，本发明实施例提供了一种山坡耕地土壤环境改良方法。本发明采用的技术方案如下：In order to solve the above technical problems, embodiments of the present invention provide a method for improving the soil environment of hillside cultivated land. The technical solutions adopted by the present invention are as follows:
一种山坡耕地土壤环境改良方法，其包括：A method for improving the soil environment of cultivated land on a hillside, which includes:
获取由摄像头采集的作物生长状态图像；Obtain images of crop growth status collected by cameras;
获取由温度传感器和湿度传感器采集的预定时间段内多个预定时间点的土壤温度值和土壤湿度值；Obtain soil temperature values and soil moisture values at multiple predetermined time points within a predetermined time period collected by the temperature sensor and the humidity sensor;
对所述作物生长状态图像进行生长状态特征提取和局部特征强化，以得到局部特征强化作物生长状态特征向量；Perform growth state feature extraction and local feature enhancement on the crop growth state image to obtain a local feature enhanced crop growth state feature vector;
对所述多个预定时间点的土壤温度值和土壤湿度值进行关联分析，以得到土壤温度-湿度时序特征向量；Perform correlation analysis on the soil temperature values and soil moisture values at the multiple predetermined time points to obtain a soil temperature-humidity time series feature vector;
基于所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量之间的语义交互融合特征，确定是否进行灌溉。Based on the semantic interaction fusion feature between the local feature-enhanced crop growth state feature vector and the soil temperature-humidity time series feature vector, it is determined whether to irrigate.
在上述山坡耕地土壤环境改良方法中，对所述作物生长状态图像进行生长状态特征提取和局部特征强化，以得到局部特征强化作物生长状态特征向量，包括：将所述作物生长状态图像通过基于卷积神经网络模型的作物生长状态特征提取器，以得到作物生长状态特征图；对所述作物生长状态特征图进行局部特征强化，以得到所述局部特征强化作物生长状态特征向量。In the above soil environment improvement method for cultivated land on hillside, performing growth state feature extraction and local feature enhancement on the crop growth state image to obtain a local feature enhanced crop growth state feature vector includes: converting the crop growth state image through a volume-based The crop growth state feature extractor of the cumulative neural network model is used to obtain a crop growth state feature map; local feature enhancement is performed on the crop growth state feature map to obtain the local feature enhanced crop growth state feature vector.
在上述山坡耕地土壤环境改良方法中，对所述作物生长状态特征图进行局部特征强化，以得到局部特征强化作物生长状态特征向量，包括：将所述作物生长状态特征图通过局部信息高效建模模块，以得到所述局部特征强化作物生长状态特征向量。In the above soil environment improvement method for cultivated land on hillside, local feature enhancement is performed on the crop growth state feature map to obtain a local feature enhanced crop growth state feature vector, which includes: efficiently modeling the crop growth state feature map through local information module to obtain the local feature-enhanced crop growth state feature vector.
在上述山坡耕地土壤环境改良方法中，将所述作物生长状态特征图通过局部信息高效建模模块，以得到所述局部特征强化作物生长状态特征向量，包括：以局部特征强化公式对所述作物生长状态特征图进行局部信息高效建模，以得到所述局部特征强化作物生长状态特征向量；其中，所述局部特征强化公式为：In the above soil environment improvement method for cultivated land on hillside, the crop growth state characteristic map is passed through the local information efficient modeling module to obtain the local feature enhanced crop growth state feature vector, which includes: using the local feature enhancement formula to calculate the crop growth state feature vector. The growth state feature map performs efficient modeling of local information to obtain the local feature enhanced crop growth state feature vector; where the local feature enhancement formula is:
其中，Foutput为所述局部特征强化作物生长状态特征向量，Finput为所述作物生长状态特征图，GAP表示进行池化操作，ReLU表示进行ReLU激活处理，Conv1×1(·)表示进行基于1×1卷积核的卷积操作，Conv3×3(·)表示进行基于3×3卷积核的卷积操作。 Among them, F output is the local feature-enhanced crop growth state feature vector, F input is the crop growth state feature map, GAP means performing pooling operation, ReLU means performing ReLU activation processing, and Conv 1×1 (·) means performing Convolution operation based on 1×1 convolution kernel, Conv 3×3 (·) indicates convolution operation based on 3×3 convolution kernel.
在上述山坡耕地土壤环境改良方法中，对所述多个预定时间点的土壤温度值和土壤湿度值进行关联分析，以得到土壤温度-湿度时序特征向量，包括：将所述多个预定时间点的土壤温度值和土壤湿度值按照时间维度分别排列为土壤温度时序输入向量和土壤湿度时序输入向量；对所述土壤温度时序输入向量和所述土壤湿度时序输入向量进行时序关联编码，以得到土壤温度-土壤湿度时序关联矩阵；将所述土壤温度-土壤湿度时序关联矩阵通过基于卷积神经网络模型的土壤温度-湿度时序特征提取器，以得到所述土壤温度-湿度时序特征向量。In the above method for improving the soil environment of cultivated land on a hillside, correlation analysis is performed on the soil temperature values and soil moisture values at the plurality of predetermined time points to obtain a soil temperature-humidity time series feature vector, which includes: converting the plurality of predetermined time points The soil temperature values and soil moisture values are respectively arranged into soil temperature time series input vectors and soil moisture time series input vectors according to the time dimension; perform time series correlation coding on the soil temperature time series input vector and the soil moisture time series input vector to obtain the soil Temperature-soil moisture time series correlation matrix; pass the soil temperature-soil moisture time series correlation matrix through the soil temperature-humidity time series feature extractor based on the convolutional neural network model to obtain the soil temperature-humidity time series feature vector.
在上述山坡耕地土壤环境改良方法中，对所述土壤温度时序输入向量和所述土壤湿度时序输入向量进行时序关联编码，以得到土壤温度-土壤湿度时序关联矩阵，包括：计算所述土壤温度时序输入向量和所述土壤湿度时序输入向量之间的样本协方差关联矩阵，以得到所述土壤温度-土壤湿度时序关联矩阵。In the above soil environment improvement method for cultivated land on hillside, time series correlation coding is performed on the soil temperature time series input vector and the soil moisture time series input vector to obtain a soil temperature-soil moisture time series correlation matrix, including: calculating the soil temperature time series The sample covariance correlation matrix between the input vector and the soil moisture time series input vector is used to obtain the soil temperature-soil moisture time series correlation matrix.
在上述山坡耕地土壤环境改良方法中，计算所述土壤温度时序输入向量和所述土壤湿度时序输入向量之间的样本协方差关联矩阵，以得到所述土壤温度-土壤湿度时序关联矩阵，包括：以样本协方差关联公式计算所述土壤温度时序输入向量和所述土壤湿度时序输入向量之间的样本协方差关联矩阵，以得到所述土壤温度-土壤湿度时序关联矩阵；其中，所述样本协方差关联公式为：In the above soil environment improvement method for cultivated land on hillside, the sample covariance correlation matrix between the soil temperature time series input vector and the soil moisture time series input vector is calculated to obtain the soil temperature-soil moisture time series correlation matrix, including: Calculate the sample covariance correlation matrix between the soil temperature time series input vector and the soil moisture time series input vector using the sample covariance correlation formula to obtain the soil temperature-soil moisture time series correlation matrix; wherein, the sample covariance correlation matrix The variance correlation formula is:
Mcov=WTXXTWM cov =W T XX T W
其中，W为所述土壤温度时序输入向量，X为所述土壤湿度时序输入向量，Mcov为所述土壤温度-土壤湿度时序关联矩阵，T表示向量的转置。Wherein, W is the soil temperature time series input vector, X is the soil moisture time series input vector, M cov is the soil temperature-soil moisture time series correlation matrix, and T represents the transpose of the vector.
在上述山坡耕地土壤环境改良方法中，基于所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量之间的语义交互融合特征，确定是否进行灌溉，包括：对所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量进行特征交互，以得到作物生长状态-作物生长条件语义交互融合特征向量；对所述作物生长状态-作物生长条件语义交互融合特征向量进行优化，以得到优化作物生长状态-作物生长条件语义交互融合特征向量；将所述优化作物生长状态-作物生长条件语义交互融合特征向量通过分类器以得到分类结果，所述分类结果用于表示是否进行灌溉。In the above soil environment improvement method for cultivated land on hillside, based on the local characteristics, the semantic interaction fusion feature between the crop growth state feature vector and the soil temperature-humidity time series feature vector is strengthened to determine whether to irrigate, including: The feature-enhanced crop growth status feature vector and the soil temperature-humidity time series feature vector perform feature interaction to obtain a crop growth status-crop growth condition semantic interactive fusion feature vector; the crop growth status-crop growth condition semantic interactive fusion feature The vector is optimized to obtain the optimized crop growth status-crop growth condition semantic interaction fusion feature vector; the optimized crop growth status-crop growth condition semantic interaction fusion feature vector is passed through the classifier to obtain the classification result, and the classification result is used for Indicates whether to irrigate.
在上述山坡耕地土壤环境改良方法中，对所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量进行特征交互，以得到作物生长状态-作物生长条件语义交互融合特征向量，包括：以投影特征交互公式来对所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量进行特征交互，以得到所述作物生长状态-作物生长条件语义交互融合特征向量；其中，所述投影特征交互公式为：In the above soil environment improvement method for cultivated land on hillside, feature interaction is performed on the local feature enhanced crop growth status feature vector and the soil temperature-humidity time series feature vector to obtain a crop growth status-crop growth condition semantic interaction fusion feature vector, The method includes: using a projection feature interaction formula to perform feature interaction on the local feature-enhanced crop growth state feature vector and the soil temperature-humidity time series feature vector to obtain the crop growth state-crop growth condition semantic interactive fusion feature vector; Wherein, the projection feature interaction formula is:
其中，V1是所述局部特征强化作物生长状态特征向量，V2所述土壤温度-湿度时序特征向量，Vf是所述作物生长状态-作物生长条件语义交互融合特征向量，/>表示投影交互处理。 Among them, V 1 is the local feature-enhanced crop growth state feature vector, V 2 is the soil temperature-humidity time series feature vector, V f is the crop growth state-crop growth condition semantic interaction fusion feature vector, /> Represents projection interaction processing.
在上述山坡耕地土壤环境改良方法中，将所述优化作物生长状态-作物生长条件语义交互融合特征向量通过分类器以得到分类结果，所述分类结果用于表示是否进行灌溉，包括：使用所述分类器的全连接层对所述优化作物生长状态-作物生长条件语义交互融合特征向量进行全连接编码，以得到全连接编码特征向量；将所述全连接编码特征向量输入所述分类器的Softmax分类函数，以得到所述优化作物生长状态-作物生长条件语义交互融合特征向量归属于各个分类标签的概率值，所述分类标签包括需要进行灌溉和不需要进行灌溉；将所述概率值中最大者对应的分类标签确定为所述分类结果。In the above-mentioned method for improving the soil environment of hillside cultivated land, the optimized crop growth state-crop growth condition semantic interaction fusion feature vector is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether irrigation is to be performed, including: using the fully connected layer of the classifier to fully connect the optimized crop growth state-crop growth condition semantic interaction fusion feature vector to obtain a fully connected encoded feature vector; inputting the fully connected encoded feature vector into the Softmax classification function of the classifier to obtain the probability value of the optimized crop growth state-crop growth condition semantic interaction fusion feature vector belonging to each classification label, and the classification labels include the need for irrigation and the need for irrigation; and determining the classification label corresponding to the largest of the probability values as the classification result.
与现有技术相比，本发明提供的山坡耕地土壤环境改良方法，通过获取并对作物生长状态图像进行生长状态特征提取和局部特征强化，获取多个预定时间点的土壤温度值和土壤湿度值并进行关联分析，以及基于局部特征强化作物生长状态特征向量和土壤温度-湿度时序特征向量之间的语义交互融合特征确定是否进行灌溉，提供了一种可以对山坡耕地土壤环境进行科学改良的方法，其提出的灌溉方法能够实现精准灌溉，准确性和效率都大幅提升，从而可以提高农作物的生长和产量，促进农业可持续发展。Compared with the existing technology, the soil environment improvement method for hillside farmland provided by the present invention obtains soil temperature values and soil moisture values at multiple predetermined time points by acquiring and performing growth state feature extraction and local feature enhancement on crop growth state images. Correlation analysis is performed, and the semantic interaction fusion feature between the local feature-strengthened crop growth state feature vector and the soil temperature-humidity time series feature vector is used to determine whether to irrigate, which provides a method that can scientifically improve the soil environment of hillside cultivated land. , the irrigation method proposed by it can achieve precise irrigation, and the accuracy and efficiency are greatly improved, which can improve the growth and yield of crops and promote the sustainable development of agriculture.
附图说明Description of drawings
通过结合附图对本发明实施例进行更详细的描述，本发明的上述以及其他目的、特征和优势将变得更加明显。附图用来提供对本发明实施例的进一步理解，并且构成说明书的一部分，与本发明实施例一起用于解释本发明，并不构成对本发明的限制。在附图中，相同的参考标号通常代表相同部件或步骤。The above and other objects, features and advantages of the present invention will become more apparent through a more detailed description of the embodiments of the present invention in conjunction with the accompanying drawings. The drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the description. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. In the drawings, like reference numbers generally represent like components or steps.
图1为本发明实施例提供的山坡耕地土壤环境改良方法的流程图。Figure 1 is a flow chart of a method for improving the soil environment of hillside farmland provided by an embodiment of the present invention.
图2为本发明实施例提供的山坡耕地土壤环境改良方法的架构示意图。Figure 2 is a schematic structural diagram of a soil environment improvement method for hillside farmland provided by an embodiment of the present invention.
图3为本发明实施例中对所述作物生长状态图像进行生长状态特征提取和局部特征强化，以得到局部特征强化作物生长状态特征向量的流程图。Figure 3 is a flow chart for performing growth state feature extraction and local feature enhancement on the crop growth state image to obtain a local feature enhanced crop growth state feature vector in an embodiment of the present invention.
图4为本发明实施例中对所述多个预定时间点的土壤温度值和土壤湿度值进行关联分析，以得到土壤温度-湿度时序特征向量的流程图。Figure 4 is a flow chart of performing correlation analysis on soil temperature values and soil moisture values at multiple predetermined time points to obtain soil temperature-humidity time series feature vectors in an embodiment of the present invention.
图5为本发明实施例中基于所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量之间的语义交互融合特征确定是否进行灌溉的流程图。Figure 5 is a flow chart for determining whether to perform irrigation based on the semantic interactive fusion feature between the local feature-enhanced crop growth state feature vector and the soil temperature-humidity time series feature vector in an embodiment of the present invention.
图6为本发明实施例中将所述优化作物生长状态-作物生长条件语义交互融合特征向量通过分类器，以得到分类结果的流程图。Figure 6 is a flow chart of passing the optimized crop growth status-crop growth condition semantic interactive fusion feature vector through a classifier to obtain a classification result in an embodiment of the present invention.
具体实施方式Detailed ways
下面，将参考附图详细地描述根据本发明的示例实施例。显然，所描述的实施例仅仅是本发明的一部分实施例，而不是本发明的全部实施例，应理解，本发明不受这里描述的示例实施例的限制。Hereinafter, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments described here.
图1为本发明实施例提供的山坡耕地土壤环境改良方法的流程图。图2为本发明实施例提供的山坡耕地土壤环境改良方法的架构示意图。如图1和图2所示，本发明实施例提供的山坡耕地土壤环境改良方法，包括步骤：S110，获取由摄像头采集的作物生长状态图像；S120，获取由温度传感器和湿度传感器采集的预定时间段内多个预定时间点的土壤温度值和土壤湿度值；S130，对所述作物生长状态图像进行生长状态特征提取和局部特征强化，以得到局部特征强化作物生长状态特征向量；S140，对所述多个预定时间点的土壤温度值和土壤湿度值进行关联分析，以得到土壤温度-湿度时序特征向量；S150，基于所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量之间的语义交互融合特征，确定是否进行灌溉。Figure 1 is a flow chart of a method for improving the soil environment of hillside farmland provided by an embodiment of the present invention. Figure 2 is a schematic structural diagram of a soil environment improvement method for hillside farmland provided by an embodiment of the present invention. As shown in Figures 1 and 2, the method for improving the soil environment of cultivated land on hillside provided by the embodiment of the present invention includes the steps: S110, obtain the crop growth status image collected by the camera; S120, obtain the predetermined time collected by the temperature sensor and humidity sensor. Soil temperature values and soil moisture values at multiple predetermined time points within the segment; S130, perform growth state feature extraction and local feature enhancement on the crop growth state image to obtain a local feature enhanced crop growth state feature vector; S140, perform growth state feature extraction on the crop growth state image. Perform correlation analysis on the soil temperature values and soil moisture values at the multiple predetermined time points to obtain the soil temperature-humidity time series feature vector; S150, strengthen the crop growth state feature vector and the soil temperature-humidity time series feature based on the local features Semantic interactions between vectors fuse features to determine whether to irrigate.
在上述山坡耕地土壤环境改良方法中，所述步骤S110，获取由摄像头采集的作物生长状态图像。通过摄像头采集作物生长状态图像能够获取作物在不同生长阶段的视觉信息。通过对作物生长状态图像进行分析和处理，可以提取出作物的生长状态特征，例如叶片颜色、形态、密度等。这些特征可以反映作物的生长状况和健康状况，对于判断作物的生长状态和土壤需求具有重要意义。In the above method for improving the soil environment of farmland on hillside, step S110 is to obtain images of crop growth status collected by the camera. Collecting images of crop growth status through cameras can obtain visual information of crops at different growth stages. By analyzing and processing crop growth status images, crop growth status characteristics can be extracted, such as leaf color, shape, density, etc. These characteristics can reflect the growth and health status of crops and are of great significance for judging the growth status and soil needs of crops.
在上述山坡耕地土壤环境改良方法中，所述步骤S120，获取由温度传感器和湿度传感器采集的预定时间段内多个预定时间点的土壤温度值和土壤湿度值。应可以理解，土壤温度和湿度是影响作物生长的重要环境因素之一，与作物的生长状态和需求密切相关。通过采集多个时间点的土壤温度和湿度数据，可以获得土壤温湿度的变化趋势和周期性规律，从而更好地了解土壤环境的动态变化，为土壤环境的分析和作物灌溉提供依据。In the above method for improving the soil environment of farmland on hillside, step S120 is to obtain soil temperature values and soil moisture values at multiple predetermined time points within a predetermined time period collected by the temperature sensor and the humidity sensor. It should be understood that soil temperature and humidity are one of the important environmental factors affecting crop growth and are closely related to the growth status and needs of the crop. By collecting soil temperature and humidity data at multiple time points, the changing trends and periodic patterns of soil temperature and humidity can be obtained, thereby better understanding the dynamic changes in the soil environment and providing a basis for soil environment analysis and crop irrigation.
在上述山坡耕地土壤环境改良方法中，所述步骤S130，对所述作物生长状态图像进行生长状态特征提取和局部特征强化，以得到局部特征强化作物生长状态特征向量。图3为本发明实施例中对所述作物生长状态图像进行生长状态特征提取和局部特征强化，以得到局部特征强化作物生长状态特征向量的流程图。如图3所示，所述步骤S130，包括：S131，将所述作物生长状态图像通过基于卷积神经网络模型的作物生长状态特征提取器，以得到作物生长状态特征图；S132，对所述作物生长状态特征图进行局部特征强化，以得到所述局部特征强化作物生长状态特征向量。In the above method for improving the soil environment of cultivated land on a hillside, in step S130, growth state feature extraction and local feature enhancement are performed on the crop growth state image to obtain a local feature enhanced crop growth state feature vector. Figure 3 is a flow chart for performing growth state feature extraction and local feature enhancement on the crop growth state image to obtain a local feature enhanced crop growth state feature vector in an embodiment of the present invention. As shown in Figure 3, the step S130 includes: S131, passing the crop growth status image through a crop growth status feature extractor based on the convolutional neural network model to obtain a crop growth status feature map; S132, applying the crop growth status feature map to the crop growth status feature extractor. The crop growth state feature map is subjected to local feature enhancement to obtain the local feature enhanced crop growth state feature vector.
具体地，卷积神经网络（Convolutional Neural Network，CNN）是一种专门用于处理图像数据的深度学习模型，其通过多层卷积和池化操作，能够自动学习图像中的特征表示。在作物生长状态特征提取过程中，卷积神经网络通过卷积层和池化层的组合，可以有效地获取图像中的局部信息，能够捕捉到作物生长状态图像中的细节特征，例如叶片的纹理、形状等重要特征，从而反映作物在不同生长阶段的生长状态。相比之下，传统的全连接神经网络在处理图像时需要将图像展平为一维向量，会丢失图像的空间结构信息。并且，卷积神经网络中的卷积层具有参数共享的特性，即同一卷积核在图像的不同位置共享权重。这样可以大大减少模型的参数量，提高模型的训练效率，并且使得模型对于平移、旋转等图像变换具有一定的不变性。对于所述作物生长状态图像，这种参数共享的特性可以更好地利用图像中的重复模式和局部特征，使得基于卷积神经网络模型的作物生长状态特征提取器能够从作物生长状态图像中提取出对作物生长状态判断有用的高级特征表示。Specifically, Convolutional Neural Network (CNN) is a deep learning model specially used to process image data. It can automatically learn feature representations in images through multi-layer convolution and pooling operations. In the process of crop growth state feature extraction, the convolutional neural network can effectively obtain local information in the image through the combination of convolution layer and pooling layer, and can capture detailed features in the crop growth state image, such as the texture of leaves. , shape and other important characteristics, thus reflecting the growth status of crops at different growth stages. In contrast, traditional fully connected neural networks need to flatten the image into a one-dimensional vector when processing the image, which will lose the spatial structure information of the image. Moreover, the convolutional layer in the convolutional neural network has the characteristic of parameter sharing, that is, the same convolution kernel shares weights at different positions in the image. This can greatly reduce the number of parameters of the model, improve the training efficiency of the model, and make the model have certain invariance to image transformations such as translation and rotation. For the crop growth status image, this parameter sharing feature can better utilize repeated patterns and local features in the image, allowing the crop growth status feature extractor based on the convolutional neural network model to extract from the crop growth status image. Develop high-level feature representations useful for judging crop growth status.
在本发明的技术方案中，为了进一步提升特征的表达能力，增强对作物生长状态的理解和描述，进而对所述作物生长状态特征图进行局部特征强化处理。在本发明的一个具体示例中，对所述作物生长状态特征图进行局部特征强化，以得到所述局部特征强化作物生长状态特征向量的编码方式是：将所述作物生长状态特征图通过局部信息高效建模模块，以得到所述局部特征强化作物生长状态特征向量。应可以理解，在所述作物生长状态特征图中，每个位置的特征都包含了与作物生长状态相关的信息。然而，不同位置的局部特征对作物生长状态的贡献程度可能是不同的。部分局部区域可能包含了更重要或更具代表性的特征，而其他区域则可能包含了噪声或不相关的信息。通过对作物生长状态特征图通过局部信息高效建模模块进行局部区域的分析和建模，能够强调和聚焦于重要的局部特征，进而提取出具有代表性的局部特征，更好地表示作物生长状态的细节和差异，从而提高作物生长状态特征的判别能力和表达能力，提供更全面的信息供后续的分析和决策使用。In the technical solution of the present invention, in order to further improve the expression ability of features and enhance the understanding and description of crop growth status, local feature enhancement processing is performed on the crop growth status feature map. In a specific example of the present invention, the encoding method of performing local feature enhancement on the crop growth state feature map to obtain the local feature enhanced crop growth state feature vector is: converting the crop growth state feature map through local information Efficient modeling module to obtain the local feature-enhanced crop growth state feature vector. It should be understood that in the crop growth status feature map, the characteristics of each position include information related to the crop growth status. However, the degree of contribution of local features at different locations to crop growth status may be different. Some local areas may contain more important or representative features, while other areas may contain noise or irrelevant information. By analyzing and modeling local areas of the crop growth status feature map through the local information efficient modeling module, it is possible to emphasize and focus on important local features, and then extract representative local features to better represent the crop growth status. details and differences, thereby improving the ability to distinguish and express the characteristics of crop growth status, and provide more comprehensive information for subsequent analysis and decision-making.
在一个具体示例中，所述步骤S132，包括：以局部特征强化公式对所述作物生长状态特征图进行局部信息高效建模，以得到所述局部特征强化作物生长状态特征向量；其中，所述局部特征强化公式为：In a specific example, step S132 includes: performing local information efficient modeling on the crop growth state feature map with a local feature enhancement formula to obtain the local feature enhanced crop growth state feature vector; wherein, The local feature enhancement formula is:
其中，Foutput为所述局部特征强化作物生长状态特征向量，Finput为所述作物生长状态特征图，GAP表示进行池化操作，ReLU表示进行ReLU激活处理，Conv1×1(·)表示进行基于1×1卷积核的卷积操作，Conv3×3(·)表示进行基于3×3卷积核的卷积操作。 Among them, F output is the local feature-enhanced crop growth state feature vector, F input is the crop growth state feature map, GAP means performing pooling operation, ReLU means performing ReLU activation processing, and Conv 1×1 (·) means performing Convolution operation based on 1×1 convolution kernel, Conv 3×3 (·) indicates convolution operation based on 3×3 convolution kernel.
在上述山坡耕地土壤环境改良方法中，所述步骤S140，对所述多个预定时间点的土壤温度值和土壤湿度值进行关联分析，以得到土壤温度-湿度时序特征向量。应可以理解，土壤温度和湿度是作物生长的重要环境因素，对作物的生长和发育具有直接影响。通过对多个时间点的土壤温度和湿度进行关联分析，可以挖掘出土壤温度和湿度之间的相关性，即它们在时间上的变化是否具有一定的同步性或相似性，进而发现土壤温湿度之间的潜在关联关系，例如温度升高时湿度是否会下降，或者湿度增加时温度是否会降低等，从而提供更全面的土壤状况评估信息，有助于更准确地判断是否需要进行灌溉操作。In the above method for improving the soil environment of cultivated land on a hillside, in step S140, correlation analysis is performed on the soil temperature values and soil moisture values at the plurality of predetermined time points to obtain a soil temperature-humidity time series feature vector. It should be understood that soil temperature and humidity are important environmental factors for crop growth and have a direct impact on crop growth and development. By performing correlation analysis on soil temperature and humidity at multiple time points, the correlation between soil temperature and humidity can be discovered, that is, whether their changes in time have certain synchronicity or similarity, and then the soil temperature and humidity can be discovered. Potential correlations between soil conditions, such as whether the humidity will decrease when the temperature increases, or whether the temperature will decrease when the humidity increases, etc., thereby providing more comprehensive soil condition assessment information and helping to more accurately determine whether irrigation operations are needed.
图4为本发明实施例中对所述多个预定时间点的土壤温度值和土壤湿度值进行关联分析，以得到土壤温度-湿度时序特征向量的流程图。如图4所示，所述步骤S140，包括：S141，将所述多个预定时间点的土壤温度值和土壤湿度值按照时间维度分别排列为土壤温度时序输入向量和土壤湿度时序输入向量；S142，对所述土壤温度时序输入向量和所述土壤湿度时序输入向量进行时序关联编码，以得到土壤温度-土壤湿度时序关联矩阵；S143，将所述土壤温度-土壤湿度时序关联矩阵通过基于卷积神经网络模型的土壤温度-湿度时序特征提取器，以得到所述土壤温度-湿度时序特征向量。Figure 4 is a flow chart of performing correlation analysis on soil temperature values and soil moisture values at multiple predetermined time points to obtain soil temperature-humidity time series feature vectors in an embodiment of the present invention. As shown in Figure 4, the step S140 includes: S141, arranging the soil temperature values and soil moisture values of the multiple predetermined time points into soil temperature time series input vectors and soil moisture time series input vectors according to the time dimension; S142 , perform time series correlation coding on the soil temperature time series input vector and the soil moisture time series input vector to obtain the soil temperature-soil moisture time series correlation matrix; S143, convert the soil temperature-soil moisture time series correlation matrix through convolution-based The soil temperature-humidity time series feature extractor of the neural network model is used to obtain the soil temperature-humidity time series feature vector.
应可以理解，土壤温度和湿度是随着时间推移而变化的。通过将多个预定时间点的土壤温度值和土壤湿度值按照时间维度排列，可以形成一系列的时间序列数据，用于提供土壤温湿度的历史记录，反映出土壤温湿度的变化趋势和周期性规律，进而捕捉到土壤温湿度随时间变化的动态特征。It should be understood that soil temperature and moisture change over time. By arranging the soil temperature values and soil moisture values at multiple predetermined time points according to the time dimension, a series of time series data can be formed to provide a historical record of soil temperature and humidity, reflecting the changing trend and periodicity of soil temperature and humidity. regularity, thereby capturing the dynamic characteristics of soil temperature and humidity changes over time.
在本发明的技术方案中，考虑到土壤温度和湿度是两个相互关联的变量，它们在一定程度上会相互影响。因此，对所述土壤温度时序输入向量和所述土壤湿度时序输入向量进行时序关联编码，以提取出土壤温湿度时序关联特征。在本发明的一个具体示例中，对所述土壤温度时序输入向量和所述土壤湿度时序输入向量进行时序关联编码，以得到土壤温度-土壤湿度时序关联矩阵的编码方式是：计算所述土壤温度时序输入向量和所述土壤湿度时序输入向量之间的样本协方差关联矩阵，以得到所述土壤温度-土壤湿度时序关联矩阵。也就是，通过计算所述土壤温度时序输入向量和所述土壤湿度时序输入向量之间的样本协方差关联矩阵来衡量土壤温度数据和湿度数据之间的关联程度。应可以理解，协方差是衡量两个变量之间关联性的统计量，它可以反映变量之间的线性关系和变动趋势。通过计算所述土壤温度时序输入向量和所述土壤湿度时序输入向量之间的样本协方差关联矩阵，可以获得土壤温度和湿度在不同时间点上的关联信息，进而了解土壤温度和湿度的变化趋势是否一致、是否存在滞后效应或周期性变化等信息，为进一步的分析和建模提供数据基础。In the technical solution of the present invention, considering that soil temperature and humidity are two interrelated variables, they will affect each other to a certain extent. Therefore, time series correlation coding is performed on the soil temperature time series input vector and the soil moisture time series input vector to extract soil temperature and humidity time series correlation features. In a specific example of the present invention, the encoding method of performing time series correlation coding on the soil temperature time series input vector and the soil moisture time series input vector to obtain the soil temperature-soil moisture time series correlation matrix is: calculating the soil temperature The sample covariance correlation matrix between the time series input vector and the soil moisture time series input vector is used to obtain the soil temperature-soil moisture time series correlation matrix. That is, the correlation degree between the soil temperature data and the humidity data is measured by calculating the sample covariance correlation matrix between the soil temperature time series input vector and the soil moisture time series input vector. It should be understood that covariance is a statistic that measures the correlation between two variables, and it can reflect the linear relationship and change trend between variables. By calculating the sample covariance correlation matrix between the soil temperature time series input vector and the soil moisture time series input vector, the correlation information of soil temperature and humidity at different time points can be obtained, thereby understanding the changing trends of soil temperature and humidity. Whether it is consistent, whether there is a lag effect or cyclical changes and other information provide a data basis for further analysis and modeling.
在一个具体示例中，所述步骤S142，包括：以样本协方差关联公式计算所述土壤温度时序输入向量和所述土壤湿度时序输入向量之间的样本协方差关联矩阵；其中，所述样本协方差关联公式为：In a specific example, step S142 includes: calculating a sample covariance correlation matrix between the soil temperature time series input vector and the soil moisture time series input vector using a sample covariance correlation formula; wherein, the sample covariance correlation matrix The variance correlation formula is:
Mcov=WTXXTWM cov =W T XX T W
其中，W为所述土壤温度时序输入向量，X为所述土壤湿度时序输入向量，Mcov为所述土壤温度-土壤湿度时序关联矩阵，T表示向量的转置。Wherein, W is the soil temperature time series input vector, X is the soil moisture time series input vector, M cov is the soil temperature-soil moisture time series correlation matrix, and T represents the transpose of the vector.
具体地，所述S143在确定土壤温度-湿度时序特征向量时，通过将所述土壤温度-土壤湿度时序关联矩阵通过基于卷积神经网络模型的土壤温度-湿度时序特征提取器获得。应可以理解，卷积神经网络通过卷积层和池化层的组合，能够对所述土壤温度-土壤湿度时序关联矩阵进行抽象表示，自动地学习到所述土壤温度-土壤湿度时序关联矩阵中的关联模式，有效地提取出所述土壤温度-土壤湿度时序关联矩阵中的局部关联特征，更好地捕捉土壤温度和湿度之间的动态变化、趋势和周期性等信息，以得到更具有判别性的土壤温度-湿度时序特征向量，从而提高土壤状况评估的准确性。Specifically, when determining the soil temperature-humidity time series feature vector in S143, the soil temperature-soil moisture time series correlation matrix is obtained by passing the soil temperature-soil moisture time series feature extractor based on the convolutional neural network model. It should be understood that the convolutional neural network can abstractly represent the soil temperature-soil moisture time series correlation matrix through the combination of the convolution layer and the pooling layer, and automatically learn the soil temperature-soil moisture time series correlation matrix. The correlation model effectively extracts the local correlation features in the soil temperature-soil moisture time series correlation matrix, and better captures the dynamic changes, trends and periodicity information between soil temperature and moisture to obtain more discriminative information. The characteristic vector of soil temperature-humidity time series can improve the accuracy of soil condition assessment.
在上述山坡耕地土壤环境改良方法中，所述步骤S150，基于所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量之间的语义交互融合特征，确定是否进行灌溉。应可以理解，所述局部特征强化作物生长状态特征向量通过作物的生长速率、叶片状态等特征反映了作物对水分的需求。而所述土壤温度-湿度时序特征向量反映了土壤的温湿度变化趋势和关联关系。通过将二者进行语义交互融合，可以综合考虑作物的需水情况和土壤的实际状况，更好地理解作物对土壤状况的响应，建立作物需水和土壤状况之间的联系，并更准确、全面地评估是否需要进行灌溉。In the above soil environment improvement method for cultivated land on a hillside, step S150 determines whether to perform irrigation based on the semantic interaction fusion feature between the local feature-enhanced crop growth state feature vector and the soil temperature-humidity time series feature vector. It should be understood that the local feature-enhanced crop growth state feature vector reflects the crop's demand for water through features such as the growth rate and leaf status of the crop. The soil temperature-humidity time series feature vector reflects the soil temperature and humidity change trend and correlation relationship. By integrating the two semantically interactively, we can comprehensively consider the water needs of crops and the actual conditions of the soil, better understand the response of crops to soil conditions, establish the connection between crop water needs and soil conditions, and make it more accurate and Thoroughly assess the need for irrigation.
图5为本发明实施例中基于所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量之间的语义交互融合特征，确定是否进行灌溉的流程图。如图5所示，所述步骤S150，包括：S151，对所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量进行特征交互，以得到作物生长状态-作物生长条件语义交互融合特征向量；S152，对所述作物生长状态-作物生长条件语义交互融合特征向量进行优化，以得到优化作物生长状态-作物生长条件语义交互融合特征向量；S153，将所述优化作物生长状态-作物生长条件语义交互融合特征向量通过分类器以得到分类结果，所述分类结果用于表示是否进行灌溉。Figure 5 is a flow chart for determining whether to perform irrigation based on the semantic interaction fusion feature between the local feature-enhanced crop growth status feature vector and the soil temperature-humidity time series feature vector in an embodiment of the present invention. As shown in Figure 5, the step S150 includes: S151, performing feature interaction on the local feature enhanced crop growth status feature vector and the soil temperature-humidity time series feature vector to obtain crop growth status-crop growth condition semantics Interactive fusion feature vector; S152, optimize the crop growth status-crop growth condition semantic interactive fusion feature vector to obtain the optimized crop growth status-crop growth condition semantic interactive fusion feature vector; S153, optimize the crop growth status -Crop growth condition semantic interactive fusion feature vector is passed through the classifier to obtain a classification result, which is used to indicate whether irrigation is performed.
具体地，所述S151中，特征交互是将不同特征之间的信息进行融合和交流的过程。在本发明的技术方案中，土壤温度和湿度时序特征向量反映了土壤的动态变化情况，而作物生长状态特征向量描述了作物的生长状态，如植株高度、叶片面积、生长速率等。通过使用投影层来将所述作物生长状态特征向量和所述土壤温度-湿度时序特征向量进行组合，能够捕捉到作物生长状态和作物生长条件之间的复杂关系和相互依赖，从而获取更全面的作物生长状态-作物生长条件的特征表示，进一步提高特征的表达能力。也就是说，所述作物生长状态-作物生长条件语义交互融合特征向量综合考虑了作物生长状态特征和土壤温度-湿度时序特征的信息，并通过投影层来进行特征交互以融合了二者之间的语义交互信息，能够更准确地描述作物的生长状态和作物生长条件之间的关系，为后续的分析和建模提供更有价值的特征表示。Specifically, in S151, feature interaction is a process of integrating and communicating information between different features. In the technical solution of the present invention, the soil temperature and humidity time series feature vectors reflect the dynamic changes of the soil, while the crop growth status feature vector describes the growth status of the crop, such as plant height, leaf area, growth rate, etc. By using a projection layer to combine the crop growth state feature vector and the soil temperature-humidity time series feature vector, the complex relationship and interdependence between the crop growth state and crop growth conditions can be captured, thereby obtaining a more comprehensive Crop growth status - characteristic representation of crop growth conditions to further improve the expression ability of characteristics. That is to say, the crop growth status-crop growth condition semantic interaction fusion feature vector comprehensively considers the information of crop growth status characteristics and soil temperature-humidity time series characteristics, and performs feature interaction through the projection layer to fuse the two. The semantic interactive information can more accurately describe the relationship between crop growth status and crop growth conditions, providing more valuable feature representation for subsequent analysis and modeling.
在一个具体示例中，所述步骤S151在具体实现时：以投影特征交互公式来对所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量进行特征交互，以得到所述作物生长状态-作物生长条件语义交互融合特征向量；其中，所述投影特征交互公式为：In a specific example, when step S151 is implemented, a projection feature interaction formula is used to perform feature interaction on the local feature enhanced crop growth state feature vector and the soil temperature-humidity time series feature vector to obtain the Crop growth status-crop growth condition semantic interaction fusion feature vector; wherein, the projection feature interaction formula is:
其中，V1是所述局部特征强化作物生长状态特征向量，V2所述土壤温度-湿度时序特征向量，Vf是所述作物生长状态-作物生长条件语义交互融合特征向量，/>表示投影交互处理。 Among them, V 1 is the local feature-enhanced crop growth state feature vector, V 2 is the soil temperature-humidity time series feature vector, V f is the crop growth state-crop growth condition semantic interaction fusion feature vector, /> Represents projection interaction processing.
具体地，所述S152在具体实现时可以包括：对所述局部特征强化作物生长状态特征向量与所述土壤温度-湿度时序特征向量进行融合优化，以得到校正特征向量；以及，将所述校正特征向量与所述作物生长状态-作物生长条件语义交互融合特征向量融合，以得到所述优化作物生长状态-作物生长条件语义交互融合特征向量。应可以理解，在上述技术方案中，所述作物生长状态特征图表示作物生长状态图像的基于卷积神经网络模型的图像语义特征。在通过所述局部信息高效建模模块后，所述局部信息高效建模模块能够通过不同感受野的卷积核来聚焦所述作物生长状态特征图中的局部信息，以使所述局部特征强化作物生长状态特征向量能够具有更高的特征表达分辨率。所述土壤温度-湿度时序特征向量表达的是土壤温度值和土壤湿度值的时域关联的基于卷积核的局部时域下温度-湿度的局部关联模式特征。考虑到所述土壤温度值、所述土壤湿度值和所述作物生长状态图像的数据源域分布差异以及各自受噪声的影响，在使用投影层来对所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量进行特征交互时，所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量由于源域分布差异而放大的高阶关联特征分布差异，所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量存在显著的交互对应稀疏性，从而影响所述作物生长状态-作物生长条件语义交互融合特征向量的表达效果，因此，为了提升融合时的分布信息表示一致性，本发明对所述局部特征强化作物生长状态特征向量与所述土壤温度-湿度时序特征向量进行融合优化。Specifically, the specific implementation of S152 may include: performing fusion optimization on the local feature-enhanced crop growth state feature vector and the soil temperature-humidity time series feature vector to obtain a correction feature vector; and, applying the correction to The feature vector is fused with the crop growth status-crop growth condition semantic interactive fusion feature vector to obtain the optimized crop growth status-crop growth condition semantic interactive fusion feature vector. It should be understood that in the above technical solution, the crop growth status feature map represents the image semantic features of the crop growth status image based on the convolutional neural network model. After passing the local information efficient modeling module, the local information efficient modeling module can focus on the local information in the crop growth state feature map through convolution kernels with different receptive fields, so as to strengthen the local features. The crop growth status feature vector can have higher feature expression resolution. The soil temperature-humidity time series feature vector expresses the local correlation pattern characteristics of temperature-humidity in the local time domain based on the convolution kernel of the time-domain correlation of soil temperature values and soil moisture values. Considering the difference in data source domain distribution of the soil temperature value, the soil moisture value and the crop growth status image and the influence of noise on each, a projection layer is used to enhance the local features of the crop growth status feature vector and When the soil temperature-humidity time series feature vector performs feature interaction, the local feature strengthens the crop growth state feature vector and the soil temperature-humidity time series feature vector amplified by the distribution difference of high-order correlation features due to the difference in source domain distribution, so The local feature-enhanced crop growth state feature vector and the soil temperature-humidity time series feature vector have significant interactive correspondence sparsity, thereby affecting the expression effect of the crop growth state-crop growth condition semantic interactive fusion feature vector. Therefore, in order to To improve the consistency of distribution information representation during fusion, the present invention performs fusion optimization on the local feature-enhanced crop growth state feature vector and the soil temperature-humidity time series feature vector.
基于此，在本发明的技术方案中，以如下融合优化公式对所述局部特征强化作物生长状态特征向量与所述土壤温度-湿度时序特征向量进行融合优化，以得到所述校正特征向量；其中，所述融合优化公式为：Based on this, in the technical solution of the present invention, the local feature-enhanced crop growth state feature vector and the soil temperature-humidity time series feature vector are fused and optimized with the following fusion optimization formula to obtain the correction feature vector; where , the fusion optimization formula is:
其中，V1是所述局部特征强化作物生长状态特征向量，V2是所述土壤温度-湿度时序特征向量，v1i、v2i和vci分别是所述局部特征强化作物生长状态特征向量V1、所述土壤温度-湿度时序特征向量V2和所述校正特征向量的特征值，和/>分别表示特征向量的1范数和2范数的平方，局部特征强化作物生长状态特征向量V1和土壤温度-湿度时序特征向量V2具有相同的特征向量长度L，且ε是权重超参数，log表示以2为底的对数函数值。Wherein, V 1 is the local feature enhanced crop growth state feature vector, V 2 is the soil temperature-humidity time series feature vector, v 1i , v 2i and v ci are the local feature enhanced crop growth state feature vector V respectively. 1. The soil temperature-humidity time series eigenvector V 2 and the eigenvalues of the correction eigenvector, and/> represent the square of the 1 norm and the 2 norm of the feature vector respectively. The local feature enhanced crop growth state feature vector V 1 and the soil temperature-humidity time series feature vector V 2 have the same feature vector length L, and ε is the weight hyperparameter, log represents the base 2 logarithmic function value.
这里，为了提升所述局部特征强化作物生长状态特征向量与所述土壤温度-湿度时序特征向量在特征融合场景下的分布信息表示的一致性，通过待融合特征向量的特征尺度和结构表示来预定义分布回归的绝对坐标，以作为特征值交叉几何配准的基准，这样，可以保持信息分布的刚性网格一致性，并利用概率倒角损失的思路来惩罚特征分布信息表示之间的基于距离的不对齐和不完全重叠，以实现所述局部特征强化作物生长状态特征向量与所述土壤温度-湿度时序特征向量的分布信息表示一致的特征融合。这样，再以由vci组成的校正特征向量与所述作物生长状态-作物生长条件语义交互融合特征向量融合，就可以提升所述作物生长状态-作物生长条件语义交互融合特征向量对所述局部特征强化作物生长状态特征向量和所述土壤温度-湿度时序特征向量的交互融合表达效果，从而提升其通过分类器得到的分类结果的准确性。Here, in order to improve the consistency of the distribution information representation of the local feature-enhanced crop growth state feature vector and the soil temperature-humidity time series feature vector in the feature fusion scenario, the feature scale and structure representation of the feature vector to be fused are used to predict Define the absolute coordinates of distribution regression as the basis for feature value cross-geometric registration. In this way, the rigid grid consistency of the information distribution can be maintained, and the idea of probabilistic chamfer loss is used to punish the distance-based distance between feature distribution information representations. misalignment and incomplete overlap to achieve consistent feature fusion with the distribution information representation of the local feature-enhanced crop growth state feature vector and the soil temperature-humidity time series feature vector. In this way, by fusing the correction feature vector composed of v ci with the crop growth status-crop growth condition semantic interactive fusion feature vector, the crop growth status-crop growth condition semantic interactive fusion feature vector can be improved for the local The feature enhances the interactive fusion expression effect of the crop growth state feature vector and the soil temperature-humidity time series feature vector, thereby improving the accuracy of the classification results obtained by the classifier.
具体地，所述S153中，利用分类器来学习作物生长状态和作物生长条件与灌溉需求之间的关系，并根据这种关系对新的样本进行分类。也就是，通过分类器来将输入的所述优化作物生长状态-作物生长条件语义交互融合特征向量映射到两个类别中的一个，即需要灌溉和不需要灌溉。这样，基于分类结果来决策是否进行灌溉，以实现灌溉的合理控制，进而改善耕地的土壤环境，提高农作物的生长和产量。Specifically, in S153, a classifier is used to learn the relationship between crop growth status and crop growth conditions and irrigation needs, and new samples are classified according to this relationship. That is, the input optimal crop growth status-crop growth condition semantic interaction fusion feature vector is mapped through a classifier to one of two categories, that is, irrigation is required and irrigation is not required. In this way, decisions are made on whether to irrigate based on the classification results to achieve reasonable control of irrigation, thereby improving the soil environment of cultivated land and increasing the growth and yield of crops.
图6为本发明实施例中将所述优化作物生长状态-作物生长条件语义交互融合特征向量通过分类器以得到分类结果的流程图。如图6所示，所述S153，包括：S1531，使用所述分类器的全连接层对所述优化作物生长状态-作物生长条件语义交互融合特征向量进行全连接编码，以得到全连接编码特征向量；S1532，将所述全连接编码特征向量输入所述分类器的Softmax分类函数，以得到所述优化作物生长状态-作物生长条件语义交互融合特征向量归属于各个分类标签的概率值，所述分类标签包括需要进行灌溉和不需要进行灌溉；S1533，将所述概率值中最大者对应的分类标签确定为所述分类结果。Figure 6 is a flow chart for passing the optimized crop growth status-crop growth condition semantic interactive fusion feature vector through a classifier to obtain a classification result in an embodiment of the present invention. As shown in Figure 6, the S153 includes: S1531, using the fully connected layer of the classifier to perform fully connected encoding on the optimized crop growth status-crop growth condition semantic interaction fusion feature vector to obtain fully connected encoding features Vector; S1532, input the fully connected encoding feature vector into the Softmax classification function of the classifier to obtain the probability value of the optimized crop growth status-crop growth condition semantic interactive fusion feature vector belonging to each classification label, the The classification labels include irrigation required and irrigation not required; S1533, determine the classification label corresponding to the largest probability value as the classification result.
综上，本发明实施例提供的山坡耕地土壤环境改良方法被阐明，其利用基于深度学习的人工智能技术来分析土壤温度和湿度之间的关联变化特征，并基于土壤温湿度关联变化特征与作物的生长状态之间的语义交互关联关系，进而判断土壤是否需要进行灌溉。这样，可以实现对山坡耕地土壤环境的科学改良，从而提高农作物的生长和产量，促进农业可持续发展。In summary, the method for improving the soil environment of cultivated land on hillside provided by the embodiment of the present invention has been clarified, which uses artificial intelligence technology based on deep learning to analyze the correlation change characteristics between soil temperature and humidity, and based on the correlation change characteristics of soil temperature and humidity with crops The semantic interaction relationship between the growth status of the soil, and then determine whether the soil needs irrigation. In this way, the soil environment of hillside farmland can be scientifically improved, thereby increasing the growth and yield of crops and promoting sustainable agricultural development.
最后应说明的是，以上实施例仅用以说明本发明的技术方案而非限制，尽管参照较佳实施例对本发明进行了详细说明，本领域的普通技术人员应当理解，可以对本发明的技术方案进行修改或等同替换，而不脱离本发明技术方案的精神和范围。Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified. Modifications or equivalent substitutions may be made without departing from the spirit and scope of the technical solution of the present invention.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410190945.8A CN117743975A (en) | 2024-02-21 | 2024-02-21 | Methods for improving soil environment on hillside cultivated land |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410190945.8A CN117743975A (en) | 2024-02-21 | 2024-02-21 | Methods for improving soil environment on hillside cultivated land |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117743975A true CN117743975A (en) | 2024-03-22 |
Family
ID=90261409
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410190945.8A Pending CN117743975A (en) | 2024-02-21 | 2024-02-21 | Methods for improving soil environment on hillside cultivated land |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117743975A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118135407A (en) * | 2024-05-07 | 2024-06-04 | 中邦生态环境有限公司 | Arbor nutrient solution preparation dynamic adjustment system and method based on big data |
| CN118212535A (en) * | 2024-05-17 | 2024-06-18 | 江西省林业科学院 | Wetland aquatic plant cultivation management system and method |
| CN118504793A (en) * | 2024-07-18 | 2024-08-16 | 中色地科矿产勘查股份有限公司 | Method for improving saline-alkali soil based on fly ash |
Citations (35)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110719336A (en) * | 2019-10-21 | 2020-01-21 | 江苏省农业科学院 | An IoT-based Irrigation Water Analysis and Monitoring System |
| CN110839519A (en) * | 2019-12-02 | 2020-02-28 | 成都信息工程大学 | Internet of things intelligent agricultural irrigation device and method based on deep learning |
| CN111480557A (en) * | 2020-04-23 | 2020-08-04 | 顾晓东 | Agricultural irrigation real-time monitoring regulation and control system based on big data Internet of things |
| CN111915332A (en) * | 2020-08-12 | 2020-11-10 | 马妮娜 | Agricultural product safety tracing, monitoring and managing system based on big data |
| CN113575388A (en) * | 2021-07-30 | 2021-11-02 | 沭阳翔玮生态农业开发有限公司 | Agricultural intelligent irrigation system based on artificial intelligence and big data |
| KR20220094945A (en) * | 2020-12-29 | 2022-07-06 | 안동대학교 산학협력단 | Intelligent autonomous irrigation system |
| WO2022253057A1 (en) * | 2021-06-01 | 2022-12-08 | 中国农业科学院蔬菜花卉研究所 | Intelligent moisture precise irrigation control system and method for fruit and vegetable cultivation in solar greenhouse |
| CN115482467A (en) * | 2022-09-29 | 2022-12-16 | 河南省景观规划设计研究院有限公司 | Automatic irrigation system in wisdom gardens |
| CN115633622A (en) * | 2022-06-06 | 2023-01-24 | 华南农业大学 | An orchard intelligent irrigation system and method thereof |
| CN116210573A (en) * | 2023-04-10 | 2023-06-06 | 上海华维可控农业科技集团股份有限公司 | Intelligent radiation monitoring and adjusting irrigation system and method based on Internet of things |
| CN116612435A (en) * | 2023-07-18 | 2023-08-18 | 吉林隆源农业服务有限公司 | Corn high-yield cultivation method |
| CN116649191A (en) * | 2023-06-14 | 2023-08-29 | 江西省科技基础条件平台中心(江西省计算中心) | Remote fertilization and irrigation control system and method based on PLC |
| CN116649159A (en) * | 2023-08-01 | 2023-08-29 | 江苏慧岸信息科技有限公司 | Edible fungus growth parameter optimizing system and method |
| CN116703638A (en) * | 2023-06-26 | 2023-09-05 | 宁波新蜂链农果业发展有限公司 | Greenhouse vegetable planting management system and method thereof |
| CN116686686A (en) * | 2023-04-12 | 2023-09-05 | 中国水利水电科学研究院 | Irrigation device and method |
| CN116862320A (en) * | 2023-09-04 | 2023-10-10 | 北京佳格天地科技有限公司 | Rural ecological environment assessment supervision method, system and storage medium |
| CN116918546A (en) * | 2023-07-11 | 2023-10-24 | 吉林隆源农业服务有限公司 | Corn high-yield planting and fertilizing method |
| CN117036089A (en) * | 2023-08-22 | 2023-11-10 | 哈尔滨辰昌农业科技有限公司 | Preparation method and system of biomass rice seedling raising tray |
| CN117036088A (en) * | 2023-08-21 | 2023-11-10 | 安阳市游园管理站 | Data acquisition and analysis method for identifying growth situation of greening plants by AI |
| CN117021402A (en) * | 2023-08-09 | 2023-11-10 | 江西新瑜新型建材有限公司 | Asphalt concrete preparation method and system thereof |
| CN117058622A (en) * | 2023-08-28 | 2023-11-14 | 海宁市立洁环保有限公司 | Intelligent monitoring system and method for sewage treatment equipment |
| CN117063820A (en) * | 2023-10-16 | 2023-11-17 | 武安市园林绿化管理局 | Intelligent garden monitoring platform based on big data and method thereof |
| CN117114627A (en) * | 2023-10-18 | 2023-11-24 | 日照市自然资源和规划局 | land resource management system |
| CN117114913A (en) * | 2023-07-03 | 2023-11-24 | 吉林隆源农业服务有限公司 | Intelligent agricultural data acquisition system based on big data |
| CN117136765A (en) * | 2023-09-04 | 2023-12-01 | 河南省农业农村科技教育中心(河南省农业广播电视学校) | Greenhouse control system and method based on intelligent agriculture |
| CN117237884A (en) * | 2023-11-06 | 2023-12-15 | 福建行无忧智能科技有限公司 | Interactive inspection robot based on berth positioning |
| CN117227005A (en) * | 2023-10-08 | 2023-12-15 | 安徽省凤阳县天瑞建材有限公司 | Production control system and method for concrete raw material processing |
| CN117270611A (en) * | 2023-11-22 | 2023-12-22 | 浙江威星电子系统软件股份有限公司 | Intelligent environmental control system and method for intelligent sports stadium |
| CN117288349A (en) * | 2023-09-21 | 2023-12-26 | 湖北鑫英泰系统技术股份有限公司 | Machine room cold channel micro-environment monitoring system |
| CN117313993A (en) * | 2023-09-15 | 2023-12-29 | 湖南省烟草公司湘西自治州公司 | Flue-cured tobacco cultivation management method based on temperature and altitude |
| CN117309786A (en) * | 2023-09-04 | 2023-12-29 | 湖南省吉祥天生物科技有限公司 | Detection and analysis method and system for organic fertilizer |
| CN117434082A (en) * | 2023-12-22 | 2024-01-23 | 广东岚瑞新材料科技集团有限公司 | Quality detection method and system for silica gel product production line |
| CN117540908A (en) * | 2023-11-07 | 2024-02-09 | 北京佳格天地科技有限公司 | Agricultural resource integration method and system based on big data |
| CN117535452A (en) * | 2024-01-09 | 2024-02-09 | 延边大学 | Online monitoring method and system for production of bacterial bran fermented feed |
| CN117558472A (en) * | 2024-01-11 | 2024-02-13 | 深圳大学 | Nuclear reactor cooling system and cooling control method thereof |
-
2024
- 2024-02-21 CN CN202410190945.8A patent/CN117743975A/en active Pending
Patent Citations (35)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110719336A (en) * | 2019-10-21 | 2020-01-21 | 江苏省农业科学院 | An IoT-based Irrigation Water Analysis and Monitoring System |
| CN110839519A (en) * | 2019-12-02 | 2020-02-28 | 成都信息工程大学 | Internet of things intelligent agricultural irrigation device and method based on deep learning |
| CN111480557A (en) * | 2020-04-23 | 2020-08-04 | 顾晓东 | Agricultural irrigation real-time monitoring regulation and control system based on big data Internet of things |
| CN111915332A (en) * | 2020-08-12 | 2020-11-10 | 马妮娜 | Agricultural product safety tracing, monitoring and managing system based on big data |
| KR20220094945A (en) * | 2020-12-29 | 2022-07-06 | 안동대학교 산학협력단 | Intelligent autonomous irrigation system |
| WO2022253057A1 (en) * | 2021-06-01 | 2022-12-08 | 中国农业科学院蔬菜花卉研究所 | Intelligent moisture precise irrigation control system and method for fruit and vegetable cultivation in solar greenhouse |
| CN113575388A (en) * | 2021-07-30 | 2021-11-02 | 沭阳翔玮生态农业开发有限公司 | Agricultural intelligent irrigation system based on artificial intelligence and big data |
| CN115633622A (en) * | 2022-06-06 | 2023-01-24 | 华南农业大学 | An orchard intelligent irrigation system and method thereof |
| CN115482467A (en) * | 2022-09-29 | 2022-12-16 | 河南省景观规划设计研究院有限公司 | Automatic irrigation system in wisdom gardens |
| CN116210573A (en) * | 2023-04-10 | 2023-06-06 | 上海华维可控农业科技集团股份有限公司 | Intelligent radiation monitoring and adjusting irrigation system and method based on Internet of things |
| CN116686686A (en) * | 2023-04-12 | 2023-09-05 | 中国水利水电科学研究院 | Irrigation device and method |
| CN116649191A (en) * | 2023-06-14 | 2023-08-29 | 江西省科技基础条件平台中心(江西省计算中心) | Remote fertilization and irrigation control system and method based on PLC |
| CN116703638A (en) * | 2023-06-26 | 2023-09-05 | 宁波新蜂链农果业发展有限公司 | Greenhouse vegetable planting management system and method thereof |
| CN117114913A (en) * | 2023-07-03 | 2023-11-24 | 吉林隆源农业服务有限公司 | Intelligent agricultural data acquisition system based on big data |
| CN116918546A (en) * | 2023-07-11 | 2023-10-24 | 吉林隆源农业服务有限公司 | Corn high-yield planting and fertilizing method |
| CN116612435A (en) * | 2023-07-18 | 2023-08-18 | 吉林隆源农业服务有限公司 | Corn high-yield cultivation method |
| CN116649159A (en) * | 2023-08-01 | 2023-08-29 | 江苏慧岸信息科技有限公司 | Edible fungus growth parameter optimizing system and method |
| CN117021402A (en) * | 2023-08-09 | 2023-11-10 | 江西新瑜新型建材有限公司 | Asphalt concrete preparation method and system thereof |
| CN117036088A (en) * | 2023-08-21 | 2023-11-10 | 安阳市游园管理站 | Data acquisition and analysis method for identifying growth situation of greening plants by AI |
| CN117036089A (en) * | 2023-08-22 | 2023-11-10 | 哈尔滨辰昌农业科技有限公司 | Preparation method and system of biomass rice seedling raising tray |
| CN117058622A (en) * | 2023-08-28 | 2023-11-14 | 海宁市立洁环保有限公司 | Intelligent monitoring system and method for sewage treatment equipment |
| CN117309786A (en) * | 2023-09-04 | 2023-12-29 | 湖南省吉祥天生物科技有限公司 | Detection and analysis method and system for organic fertilizer |
| CN117136765A (en) * | 2023-09-04 | 2023-12-01 | 河南省农业农村科技教育中心(河南省农业广播电视学校) | Greenhouse control system and method based on intelligent agriculture |
| CN116862320A (en) * | 2023-09-04 | 2023-10-10 | 北京佳格天地科技有限公司 | Rural ecological environment assessment supervision method, system and storage medium |
| CN117313993A (en) * | 2023-09-15 | 2023-12-29 | 湖南省烟草公司湘西自治州公司 | Flue-cured tobacco cultivation management method based on temperature and altitude |
| CN117288349A (en) * | 2023-09-21 | 2023-12-26 | 湖北鑫英泰系统技术股份有限公司 | Machine room cold channel micro-environment monitoring system |
| CN117227005A (en) * | 2023-10-08 | 2023-12-15 | 安徽省凤阳县天瑞建材有限公司 | Production control system and method for concrete raw material processing |
| CN117063820A (en) * | 2023-10-16 | 2023-11-17 | 武安市园林绿化管理局 | Intelligent garden monitoring platform based on big data and method thereof |
| CN117114627A (en) * | 2023-10-18 | 2023-11-24 | 日照市自然资源和规划局 | land resource management system |
| CN117237884A (en) * | 2023-11-06 | 2023-12-15 | 福建行无忧智能科技有限公司 | Interactive inspection robot based on berth positioning |
| CN117540908A (en) * | 2023-11-07 | 2024-02-09 | 北京佳格天地科技有限公司 | Agricultural resource integration method and system based on big data |
| CN117270611A (en) * | 2023-11-22 | 2023-12-22 | 浙江威星电子系统软件股份有限公司 | Intelligent environmental control system and method for intelligent sports stadium |
| CN117434082A (en) * | 2023-12-22 | 2024-01-23 | 广东岚瑞新材料科技集团有限公司 | Quality detection method and system for silica gel product production line |
| CN117535452A (en) * | 2024-01-09 | 2024-02-09 | 延边大学 | Online monitoring method and system for production of bacterial bran fermented feed |
| CN117558472A (en) * | 2024-01-11 | 2024-02-13 | 深圳大学 | Nuclear reactor cooling system and cooling control method thereof |
Non-Patent Citations (2)
| Title |
|---|
| JIANXIN LI 等: "USIP：Unsupervised Stable Interest Point Detection From 3D Point Clouds", 《2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION》, 27 February 2020 (2020-02-27), pages 361 - 370 * |
| 李德旺;许春雨;宋建成;田慕琴;邢希君;: "基于物联网的远程智能灌溉控制系统的开发", 《节水灌溉》, no. 10, 5 October 2017 (2017-10-05), pages 87 - 91 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118135407A (en) * | 2024-05-07 | 2024-06-04 | 中邦生态环境有限公司 | Arbor nutrient solution preparation dynamic adjustment system and method based on big data |
| CN118212535A (en) * | 2024-05-17 | 2024-06-18 | 江西省林业科学院 | Wetland aquatic plant cultivation management system and method |
| CN118504793A (en) * | 2024-07-18 | 2024-08-16 | 中色地科矿产勘查股份有限公司 | Method for improving saline-alkali soil based on fly ash |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111986099B (en) | Tillage monitoring method and system based on convolutional neural network with residual error correction fused | |
| Militante et al. | Plant leaf detection and disease recognition using deep learning | |
| CN117743975A (en) | Methods for improving soil environment on hillside cultivated land | |
| CN106971152B (en) | Method for detecting bird nest in power transmission line based on aerial images | |
| CN102013021A (en) | Tea tender shoot segmentation and identification method based on color and region growth | |
| CN117540908B (en) | Agricultural resource integration method and system based on big data | |
| Zhu et al. | Identifying carrot appearance quality by an improved dense CapNet | |
| CN109934256A (en) | A method for identifying weeds in paddy fields based on GA-ANN feature dimensionality reduction and SOM feature optimization | |
| CN111127423A (en) | Rice pest and disease identification method based on CNN-BP neural network algorithm | |
| CN117114913A (en) | Intelligent agricultural data acquisition system based on big data | |
| CN117765403A (en) | fertilizing method for improving lodging resistance and grain quality of crops | |
| CN116649191A (en) | Remote fertilization and irrigation control system and method based on PLC | |
| Paçal et al. | Data-efficient vision transformer models for robust classification of sugarcane | |
| CN114511787A (en) | Neural network-based remote sensing image ground feature information generation method and system | |
| CN118246720B (en) | Agricultural production management method and system applied to smart rural service system | |
| Paleti et al. | Sugar cane leaf disease classification and identification using deep machine learning algorithms | |
| Balaji et al. | Plant infirmity detection using vgg-16 convolutional neural network | |
| Kumawat et al. | Time-Variant Satellite Vegetation Classification Enabled by Hybrid Metaheuristic-Based Adaptive Time-Weighted Dynamic Time Warping | |
| He et al. | CMENet: A cross-modal enhancement network for tobacco leaf grading | |
| Suriya et al. | Development and Analysis of CNN based Disease Detection in Cotton Plants | |
| Nirmaladevi et al. | Prediction of leaf disease and pest detection using deep learning | |
| Tian et al. | Recognition method of corn and rice crop growth state based on computer image processing technology | |
| CN118072251B (en) | Tobacco pest identification method, medium and system | |
| Xu et al. | Identification of leek diseases based on deep learning algorithms | |
| Valarmathi et al. | Fruit Disease Prediction with Fertilizer Recommendation for Citrus Family using Deep Learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |