A method and system for monitoring oil-tea camellia planting areas taking into account forest age and growth suitability

Abstract

The present invention discloses a method and system for monitoring oil-tea camellia plantations that takes into account forest age and growth suitability. The system comprises the following steps: extracting multi-level features of newly planted and old oil-tea camellia forests from high-resolution and multispectral images using a deep learning model and a machine learning model, respectively, and outputting their distribution results; acquiring and integrating meteorological reanalysis data and terrain feature data, combining them with an oil-tea camellia growth suitability assessment, calculating and normalizing the suitability scores for each dimension, and generating an oil-tea camellia growth suitability distribution map; integrating the distribution results of new and old oil-tea camellia forests into preliminary oil-tea camellia planting areas, and then applying constraints using the suitability distribution map to ultimately obtain a revised oil-tea camellia planting area distribution map. This solution has high detection accuracy and is suitable for monitoring and managing large-scale oil-tea camellia resources.

Description

Technical Field

[0001] The present invention belongs to the fields of remote sensing geographic information systems and computers, and more specifically, relates to a method and system for monitoring oil-tea camellia planting areas taking into account forest age and growth suitability. Background Art

[0002] With the development of the national economy, the purchasing power of Chinese people has continued to rise, and the market for edible vegetable oils has continued to expand. The oil-tea camellia industry has become a key pillar of the agricultural economy in many regions, and the area of ​​oil-tea camellia cultivation in my country has shown a growing trend year by year. Therefore, conducting large-scale, precise monitoring of oil-tea camellia planting areas is a key measure to effectively improve land resource utilization efficiency and is of great significance for promoting the sustainable development of the oil-tea camellia industry.

[0003] While traditional manual field surveys can obtain accurate information on oil-tea camellia plantation attributes, they are time-consuming and costly. In contrast, remote sensing data, with its wide coverage and short revisit cycles, can quickly and accurately capture large-scale agricultural and forestland information, providing strong support for resource surveys and production assessments. In recent years, research has applied drone and satellite remote sensing data to tasks such as growth prediction, pest and disease monitoring, and rapid yield estimation of oil-tea camellia forests. Some studies have utilized single-modality remote sensing imagery, such as high-resolution optical or multispectral imagery, to extract oil-tea camellia forests, achieving some progress. However, oil-tea camellia forests present the challenge of large intra-class feature variability and low inter-class variability. Specifically, in high-resolution optical imagery, newly planted oil-tea camellia forests exhibit distinct texture features, while older oil-tea camellia forests are more similar in visual appearance to other forests. In medium-resolution multispectral or hyperspectral imagery, the spectral characteristics of newly planted and older oil-tea camellia forests also differ significantly. This results in poor extraction accuracy for oil-tea camellia forests using current single-modality data methods.

[0004] Prior art patent CN117994701A proposes a deep learning-based method for rapid tea oil yield prediction. The method involves collecting RGB images and videos of multiple fruiting tea oil trees; manually annotating the tea oil fruits in the collected images and videos to construct a target detection dataset; optimizing the tea oil fruit target detection model by adjusting the YOLOv8 model structure, and training the target detection network model using the processed dataset; performing target detection on each frame of the image using the trained network model; associating each detected tea oil fruit target with the corresponding target in the previous frame to form a target tracking system; managing and updating target trajectories to ensure accurate maintenance of the motion trajectory of each tea oil fruit target; and finally, constructing a rapid yield estimation model. This solution has a limited scope of application and is more suitable for yield assessment in small areas, local regions, or specific sample trees. For monitoring large-scale tea oil plantations, the workload and cost of image and video acquisition are high, making efficient coverage of large areas difficult to achieve. Summary of the Invention

[0005] In order to overcome the problems in the prior art of poor detection accuracy in oil-tea camellia planting areas and difficulty in achieving efficient coverage of large areas, the present invention provides a method and system for monitoring oil-tea camellia planting areas that takes into account forest age and growth suitability.

[0006] The primary purpose of the present invention is to solve the above technical problems, and the technical solutions of the present invention are as follows:

[0007] A first aspect of the present invention provides a method for monitoring an oil-tea camellia planting area taking into account forest age and growth suitability, comprising the following steps:

[0008] The deep learning model and machine learning model are used to extract the characteristics of newly planted oil-tea camellia forests and the characteristics of old oil-tea camellia forests from high-resolution images and multispectral images, respectively, and the distribution results of newly planted oil-tea camellia forests and old oil-tea camellia forests are output;

[0009] Obtain and use meteorological reanalysis data and terrain feature data, combined with the suitability characteristics of camellia oil tree growth, calculate the camellia oil tree growth suitability scores in each dimension, normalize the scores, and obtain a camellia oil tree growth suitability distribution map;

[0010] The distribution results of new oil-tea camellia forests were integrated with those of old oil-tea camellia forests to obtain a preliminary distribution map of oil-tea camellia planting areas.

[0011] Using the oil-tea camellia growth suitability distribution map as a priori knowledge reference, suitability constraints were imposed on the preliminary distribution map of oil-tea camellia planting areas to obtain the revised distribution map of oil-tea camellia planting areas.

[0012] Furthermore, the deep learning model adopts an encoder-decoder architecture, the encoder extracts multi-level features of the newly planted oil tea forest, and the multi-level features are passed to the decoder and converted into final segmentation results; the machine learning model is a random forest model.

[0013] Furthermore, the method for extracting the characteristics of the newly planted oil-tea camellia forest and the characteristics of the old oil-tea camellia forest includes the following steps:

[0014] Using high-resolution remote sensing images, a deep learning model is trained to extract the characteristics of newly planted oil-tea tree forests. The trained model is then applied to the predicted area to output the distribution results of newly planted oil-tea tree forests.

[0015] Using the distribution results of newly planted oil tea forests as a mask, a machine learning model is trained based on multispectral images to extract the features of old oil tea forests. The trained model is then applied to the image area to be predicted to output the distribution results of old oil tea forests.

[0016] Furthermore, high-resolution remote sensing images were used to train a deep learning model to extract the characteristics of the newly planted oil-tea tree forest, including the following steps:

[0017] Using high-resolution remote sensing images and expert judgment, combined with field research, we visually interpreted the newly planted oil-tea camellia forest area and carried out detailed annotation to obtain vector labels for the newly planted oil-tea camellia forest.

[0018] The vector labels of the newly planted oil-tea camellia forest are rasterized to obtain binary raster labels.

[0019] Pairing high-resolution images with raster labels, using non-overlapping sampling techniques, we obtained sample data of newly planted oil-tea camellia forests from the images, and divided the sample data into training and validation sets.

[0020] Using the training set and the validation set to train a deep learning model to extract multi-level features of the newly planted oil-tea camellia forest;

[0021] The trained deep learning model is applied to the area to be predicted, and the distribution results of newly planted oil tea forests are output.

[0022] Furthermore, using the distribution results of newly planted oil-tea tree forests as a mask, a machine learning model was trained based on multispectral imagery to extract the characteristics of old oil-tea tree forests, including the following steps:

[0023] Acquire satellite images and perform data preprocessing to obtain multispectral images;

[0024] Calculating multispectral image features, terrain features, polarization features, and time-series spectral features using multispectral images, and fusing and adding the features to the images;

[0025] In the study area, the distribution results of new oil-tea camellia forests were used as masks to select old oil-tea camellia forest sample points in the remaining area and randomly divided them into training set and validation set;

[0026] Using the training set and the validation set to train a machine learning model to extract multi-level features of the old oil-tea camellia forest;

[0027] The trained machine learning model is applied to the area to be predicted, and the distribution results of old tea oil forests are output.

[0028] Furthermore, the method for acquiring satellite images and performing data preprocessing is as follows: using the GEE platform, respectively calling Sentinel-2 or Landsat series satellite images, as well as DEM data and Sentinel-1 images within the same time range, filtering images of a specified time period, removing clouds and selecting the required bands, and performing radiation correction, atmospheric correction, and geometric correction on the images to obtain clear, cloud-free multispectral images.

[0029] Furthermore, the method for obtaining a distribution map of oil-tea camellia growth suitability comprises the following steps:

[0030] Collect literature and research data, organize the meteorological and environmental conditions required for the growth of oil-tea camellia, and build a suitability assessment system for oil-tea camellia growth, which is divided into n levels according to the degree of suitability;

[0031] Obtain meteorological reanalysis data and terrain feature data, reproject, resample, and convert the data types of the data, and combine them with the Camellia oleifera growth suitability assessment system to calculate a multi-dimensional Camellia oleifera growth suitability score;

[0032] A scoring grid is created in the area to be predicted, and the pixel score is calculated for each environmental factor. The scores of each factor are accumulated to calculate the comprehensive suitability score of each pixel. The comprehensive suitability score is normalized to obtain the oil-tea camellia growth suitability distribution map.

[0033] Furthermore, the meteorological reanalysis data and terrain feature data are ERA5 reanalysis data and Copernicus terrain data respectively.

[0034] Furthermore, the method for correcting the distribution map of oil-tea camellia planting areas comprises the following steps:

[0035] Initialize the suitability score threshold and create a Boolean value layer on the area to be predicted to indicate whether the suitability score of the corresponding area pixel is higher than the threshold;

[0036] Pixels originally identified as oil-tea camellia fields and with suitability scores higher than the threshold are retained;

[0037] The statistical data of oil-tea camellia planting areas and areas were introduced as auxiliary verification to adjust the suitability score threshold and obtain the distribution of oil-tea camellia planting areas with high confidence.

[0038] The second aspect of the present invention provides a tea oil planting area monitoring system that takes into account the age of the forest and the suitability of growth, including a memory and a processor, wherein the memory includes a tea oil planting area monitoring method program that takes into account the age of the forest and the suitability of growth, and when the tea oil planting area monitoring method program that takes into account the age of the forest and the suitability of growth is executed by the processor, it implements the steps of a tea oil planting area monitoring method that takes into account the age of the forest and the suitability of growth.

[0039] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0040] The present invention makes full use of different types of remote sensing data and combines deep learning models to accurately extract and correct oil tea planting areas. It has strong adaptability and flexibility and is suitable for large-scale oil tea resource monitoring and management. At the same time, this scheme combines the growth suitability assessment of oil tea, which not only takes into account the distribution of oil tea planting areas, but also further optimizes the monitoring results, excludes unsuitable areas, and improves accuracy. It extracts new and old oil tea forests in layers and corrects them in combination with growth suitability data to ensure that the monitoring results are not only accurate but also meet environmental requirements, and effectively deal with the problem of large inter-class differences and small intra-class differences in oil teas of different ages. It has strong versatility and stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to make the purpose and technical solution of the present invention clearer, the present invention provides the following drawings and descriptions:

[0042] Figure 1 A flow chart of a method provided by an embodiment of the present invention;

[0043] Figure 2 The overall technical roadmap provided for the embodiments of the present invention;

[0044] Figure 3 This is a sample example of a newly planted oil-tea tree forest dataset provided by an embodiment of the present invention;

[0045] Figure 4 Schematic diagram of the oil-tea camellia forest extraction result correction process provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that, in the absence of conflict, the embodiments of the present application and the features therein can be combined with each other.

[0047] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present invention is not limited to the specific embodiments disclosed below.

[0048] Example 1:

[0049] This paper provides a method for monitoring oil-tea camellia plantations that takes into account forest age and growth suitability. To obtain a dataset of newly planted oil-tea camellia forests with high-resolution remote sensing images, 50 ESRI Word Imagery Wayback images located in various regions of Guangdong Province were downloaded. The images have a resolution of approximately 0.5 meters and are clear and cloud-free, obtained through manual inspection. Figure 1 The figure shows a flow chart of a method for monitoring oil-tea camellia planting areas taking into account forest age and growth suitability provided by the present invention. Figure 2 The following is a general technical roadmap for monitoring camellia oil plantations that takes into account stand age and growth suitability.

[0050] S1: Use deep learning models and machine learning models to extract the characteristics of newly planted oil tea forests and old oil tea forests from high-resolution images and multispectral images respectively, and output the distribution results of newly planted oil tea forests and old oil tea forests.

[0051] More specifically, the deep learning model adopts an encoder-decoder architecture, the encoder extracts multi-level features of the newly planted oil tea forest, and the multi-level features are passed to the decoder and converted into the final segmentation results; the machine learning model is a random forest model.

[0052] More specifically, the method for extracting the characteristics of the newly planted oil-tea camellia forest and the characteristics of the old oil-tea camellia forest includes the following steps:

[0053] High-resolution remote sensing images are used to train a deep learning model to extract the characteristics of newly planted oil-tea tree forests. The trained model is then applied to the area to be predicted to output the distribution results of newly planted oil-tea tree forests.

[0054] The specific process is:

[0055] First, we demarcated the sample collection area based on the approximate distribution of oil-tea camellia plantations. We then downloaded the corresponding high-resolution imagery from the open-source ESRI Word Imagery website (https: / / livingatlas.arcgis.com / wayback). Based on the data description document, we performed preprocessing on the downloaded imagery, including radiometric correction and histogram matching. We then selected clear, cloud-free images to improve subsequent annotation efficiency.

[0056] Next, 50 selected high-resolution ArcGIS images were visually interpreted by experts to label newly planted oil-tea tree forests. Uncertain areas were verified through a combination of multi-source data comparison and field research, resulting in vector labels for newly planted oil-tea tree forests corresponding to each high-resolution image.

[0057] Finally, all label vector data was converted into binary class label images, where the oil-tea tree and other classes were represented by values ​​of 255 and 0, respectively. Using non-overlapping random sampling, a certain number of samples were collected from the sample area images and labels. The paired image and label data were cropped into 512×512 sample blocks to accommodate model training on a GPU. This yielded a dataset for monitoring newly planted oil-tea tree forests using high-resolution imagery, and the sample data was divided into a training set and a validation set.

[0058] In this embodiment, through the above steps, a total of 4265 pairs of newly planted oil tea forest samples were obtained, and some of the samples are shown as follows: Figure 3 These samples are randomly divided into training set and validation set in a ratio of 3:1.

[0059] Using the training set and the validation set to train a deep learning model to extract multi-level features of the newly planted oil-tea camellia forest;

[0060] To extract newly planted tea plantations from high-resolution images, this study constructed a deep learning-based network for extracting newly planted tea plantations. In this example, the model consists of three main modules: (1) a pixel-level feature extraction module based on SAM, which is used to extract pixel-level features of the image; (2) a Transformer-based instance-level feature extraction module, which is used to learn target features in the image; and (3) a classifier, which is used to finally extract the distribution of tea plants in the image.

[0061] Specifically, the pixel-level feature extraction module incorporates the SegmentAnything Model (SAM) as its backbone network, leveraging SAM's powerful edge feature extraction capabilities to extract pixel-level features of newly planted tea oil forests from high-resolution imagery. These features are passed to a pixel decoder, which, through progressive upsampling, generates a per-pixel embedding vector. These embeddings contain both local and global contextual information for each pixel in the image, providing the basis for subsequent mask prediction. Next, the instance-level feature extraction module receives the pixel features from the previous step and, using an attention mechanism, assigns importance weights to different parts of the features, generating several embedding vectors. These embedding vectors represent features of different regions in the image. Each embedding vector independently generates a class prediction and a corresponding mask embedding vector, representing the potential objects to be extracted in the image. Finally, the classifier takes these features and converts them into the final segmentation result. By performing a dot product operation on the embedding vectors obtained by the pixel-level feature extraction module and the mask embedding vectors from the instance-level feature extraction module, multiple, potentially overlapping, binary masks are predicted. These masks are processed using a sigmoid activation function to determine whether each pixel belongs to the newly planted tea oil forest category.

[0062] After building the network structure, set appropriate parameters for model training. In this example, the number of training cycles was set to 200, the optimizer was Adam, the initial learning rate was 0.0001, and the training batch size was 8. The deep learning model was trained using the training set constructed above, and the model performance was simultaneously tested in real time using the validation set, continuously saving the model parameters with the highest current accuracy.

[0063] This example uses Xingning City, Guangdong Province, as the research area. Finally, the proposed model is trained based on a dataset of newly planted oil-tea tree forests using high-resolution remote sensing images. This model is then applied to high-resolution images of Xingning City to output the distribution of newly planted oil-tea tree forests.

[0064] Using the distribution results of newly planted oil tea forests as a mask, a machine learning model is trained based on multispectral images to extract the features of old oil tea forests. The trained model is then applied to the image area to be predicted to output the distribution results of old oil tea forests.

[0065] The specific process is:

[0066] First, the required data was preprocessed on the GEE platform. Correlation functions were used to select Sentinel-2 imagery covering the specific timeframe of January 1, 2023, to December 31, 2023. Clouds were removed from the imagery using the built-in cloud removal function, and key bands such as B2-B8, B8A, B11, and B12 were selected. Multiple vegetation indices, such as the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and Building Structure Index (BSI), were then calculated using the imagery's inherent band information. Simultaneously, terrain features such as elevation, slope, and aspect were calculated using the terrain analysis tools provided by the GEE platform. These newly calculated indices and features were then added to the imagery as new bands. Sentinel-1 imagery covering the same timeframe was acquired, and statistical features such as mean, variance, and correlation, as well as their combined features, were calculated for both VV and VH polarization patterns.

[0067] In synthetic aperture radar (SAR) imaging, polarization mode refers to the polarization used when transmitting and receiving electromagnetic waves. Common dual-polarization modes, such as Sentinel-1 data, typically include the following two common polarization combinations: VV / VH or HH / HV. The "V" here stands for vertical polarization, and the "H" for horizontal polarization. The first letter represents the transmit polarization, and the second letter represents the receive polarization.

[0068] Next, the time series spectral features are combined. In terms of the combination of time series spectral features, first, for the time series features, the statistical features of the NDVI and BSI indices in different time periods (such as monthly and quarterly) are calculated, including the mean, median, maximum, minimum, phase and amplitude. For example, when calculating the monthly mean NDVI, by looping through the image data of each month, the reduceRegion function of the GEE platform is used to calculate the monthly NDVI mean in the form of regional averages; then the spectral features of the Sentinel-2 image, the newly calculated vegetation index and terrain features, the polarization pattern statistical features of the Sentinel-1 image, and the time series features are merged into a feature set. The feature set ultimately contains 50 different features, providing rich data information for subsequent classification.

[0069] Finally, the random forest algorithm was used for classification. In the study area, the distribution results of new oil tea forests were used as masks, and old oil tea forest sample points were screened in the remaining areas, and randomly divided into training sets and validation sets. Sample areas were extracted from the feature set, and a total of 2101 sample points were obtained, of which 1471 were used as training sets and 630 were used as validation sets. The random forest algorithm was used as the classifier, the number of decision trees was set to 100, and the number of features considered when each node was split was sqrt(50), that is, about 7 features. Through multiple experiments, it was determined that this parameter setting can achieve better classification results on this data set. The classifier was trained based on the training set data, and the classification function of the GEE platform was used to apply the trained classifier to the image of the study area to generate the classification results of the old oil tea forest.

[0070] In terms of model selection, the present invention adopts a deep learning model for newly planted oil tea forests, while a random forest model is selected for old oil tea forests, mainly based on the different characteristics presented by the two in remote sensing images. Since the newly planted oil tea forests have not yet formed a dense canopy and the row spacing is relatively obvious, clear texture differences are more likely to appear in sub-meter high-resolution images, and deep learning models are good at capturing such microscopic features. On the contrary, the visual differences between old oil tea forests and other woodlands in high-resolution images are not significant. The use of multispectral images can better utilize rich spectral information to distinguish old oil tea forests from other forest areas, and the random forest algorithm is widely used in multispectral classification and has stable effects. In order to cope with the significant differences in characteristics between newly planted and old oil tea forests, the solution adopts the idea of ​​layered extraction: deep learning is used for the texture features of high-resolution images, and random forests are used for the spectral differences of multispectral images.

[0071] S2: Obtain and use meteorological reanalysis data and terrain feature data, combine with the suitability characteristics of camellia oleifera growth, calculate the suitability scores of camellia oleifera growth in each dimension, normalize the scores, and obtain a camellia oleifera growth suitability distribution map.

[0072] More specifically, the meteorological reanalysis data and terrain feature data are ERA5 reanalysis data and Copernicus terrain data, respectively.

[0073] In this example, Xingning City, Guangdong Province was used as the research area, and an assessment of the suitability of camellia oleifera growth based on meteorological data was conducted to clarify the distribution of suitability for camellia oleifera growth in the area.

[0074] In terms of data collection and preprocessing, the meteorological data are precipitation data from August 1, 2023, to September 30, 2023, obtained from the NASA / GPM_L3 / IM ERG_V06 dataset on the GEE platform, and temperature data from October 1, 2023, to December 31, 2023, January 2023, and July 2023, obtained from the ECMWF / ERA5_LAND / MONTHLY_AGGR dataset. After screening, averaging, reprojection (EPSG:4326, 1km resolution), and bilinear resampling (500m resolution), single-band precipitation data (precipitation) and temperature data for different time periods (temperature, januaryTemp, and julyTemp) are obtained. The data types are checked and converted to Int types if they do not meet the requirements.

[0075] Soil data, including organic matter content (organic Matter), total nitrogen content (total Nitrogen), and pH value, were obtained from the Chinese Soil Organic Matter Dataset of the National Qinghai-Tibet Plateau Science Data Center. They were reprojected, resampled, and their data types checked to ensure that the data resolution was 500 m. The bands were then selected and renamed according to their actual band names.

[0076] The terrain data is DEM data obtained from the USGS / SRTMGL1_003 dataset. After reprojection, resampling, and data type check, the slope and aspect data are calculated.

[0077] In terms of constructing an assessment system for the suitability of camellia oil growth, through extensive collection of literature and field survey data, the requirements of camellia oil for meteorological, environmental and other conditions in different phenological periods were clarified, and the suitability of camellia oil growth was divided into three levels, namely suitable (score 2 points), relatively suitable (score 1 point) and unsuitable (score 0 point).

[0078] To calculate the suitability scores for each dimension of camellia oil-bearing tree growth, define the suitability score (suitabilityScore) function and create a single-band image (scoreImage) with all zero values ​​to store the final score. For each environmental factor, conditional statements are used to calculate a score raster based on the suitability level of camellia oil-bearing tree growth and accumulate it in scoreImage. The suitabilityScore function is then called to obtain a pixel-level suitability score raster.

[0079] S3: The distribution results of newly planted camellia oil trees are integrated with the distribution results of old camellia oil trees to obtain a preliminary distribution map of camellia oil tree planting areas. The camellia oil tree growth suitability distribution map is used as a priori knowledge reference to impose suitability constraints on the preliminary distribution map of camellia oil tree planting areas to obtain a revised camellia oil tree planting area distribution map.

[0080] Firstly, the extraction results of new and old camellia oil trees were integrated to obtain the preliminary camellia oil tree distribution results. Then, the target size was determined, the suitability score map was resampled to the size of the camellia oil tree planting distribution map, the resampled suitability score map was normalized, and the camellia oil tree growth suitability distribution map was used as a priori knowledge reference to revise the camellia oil tree planting area distribution map.

[0081] The specific process is:

[0082] Initialize the suitability score threshold and create a Boolean value layer on the area to be predicted to indicate whether the suitability score of the corresponding area pixel is higher than the threshold;

[0083] Pixels originally identified as oil-tea camellia fields and with suitability scores higher than the threshold are retained;

[0084] The oil-tea camellia planting area and area statistics data are introduced as auxiliary verification to adjust the suitability score threshold and obtain the distribution of oil-tea camellia planting areas with high confidence. In this embodiment, the process is as follows: Figure 4 As shown in the figure, taking Xingning City as an example, from left to right are: (a) initial results of oil tea forest extraction; (b) growth suitability distribution map; (c) corrected results.

[0085] The present invention makes full use of different types of remote sensing data and combines deep learning models to accurately extract and correct oil tea planting areas. It has strong adaptability and flexibility and is suitable for large-scale oil tea resource monitoring and management. At the same time, this scheme combines the growth suitability assessment of oil tea, which not only takes into account the distribution of oil tea planting areas, but also further optimizes the monitoring results, excludes unsuitable areas, and improves accuracy. It extracts new and old oil tea forests in layers and corrects them in combination with growth suitability data to ensure that the monitoring results are not only accurate but also meet environmental requirements, and effectively deal with the problem of large inter-class differences and small intra-class differences in oil teas of different ages. It has strong versatility and stability.

[0086] Example 2:

[0087] The present embodiment provides a system for monitoring tea oil planting areas taking into account forest age and growth suitability, comprising a memory and a processor, wherein the memory comprises a program for monitoring a tea oil planting area taking into account forest age and growth suitability, and when the program for monitoring a tea oil planting area taking into account forest age and growth suitability is executed by the processor, the steps of a method for monitoring a tea oil planting area taking into account forest age and growth suitability as described in Example 1 are implemented.

[0088] Obviously, the above embodiments of the present invention are merely examples for the purpose of clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Those skilled in the art will appreciate that other variations or modifications can be made based on the above description. It is not necessary and impossible to enumerate all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the claims of the present invention.

Claims

1. A method for monitoring oil-tea camellia planting areas taking into account forest age and growth suitability, characterized in that: The steps include: The deep learning model and machine learning model are used to extract the characteristics of newly planted oil-tea camellia forests and the characteristics of old oil-tea camellia forests from high-resolution images and multispectral images, respectively, and the distribution results of newly planted oil-tea camellia forests and old oil-tea camellia forests are output; Obtain and use meteorological reanalysis data and terrain feature data, combined with the suitability characteristics of camellia oil tree growth, calculate the camellia oil tree growth suitability scores in each dimension, normalize the scores, and obtain a camellia oil tree growth suitability distribution map; The distribution results of newly planted camellia oil trees were integrated with those of old camellia oil trees to obtain a preliminary distribution map of camellia oil tree planting areas. The camellia oil tree growth suitability distribution map was used as a priori knowledge reference, and suitability constraints were imposed on the preliminary distribution map of camellia oil tree planting areas to obtain a revised distribution map of camellia oil tree planting areas. The method for extracting the characteristics of the newly planted oil-tea camellia forest and the characteristics of the old oil-tea camellia forest comprises the following steps: Using high-resolution remote sensing images, a deep learning model is trained to extract the characteristics of newly planted oil-tea tree forests. The trained model is then applied to the predicted area to output the distribution results of newly planted oil-tea tree forests. Using the distribution results of newly planted oil tea forests as a mask, a machine learning model is trained based on multispectral images to extract the features of old oil tea forests. The trained model is then applied to the image area to be predicted to output the distribution results of old oil tea forests.

2. The method for monitoring camellia oil plantation areas taking into account forest age and growth suitability according to claim 1, characterized in that: The deep learning model adopts an encoder-decoder architecture. The encoder extracts multi-level features of the newly planted oil tea forest. The multi-level features are passed to the decoder and converted into the final segmentation results. The machine learning model is a random forest model.

3. The method for monitoring camellia oil plantation areas taking into account forest age and growth suitability according to claim 1, characterized in that: Using high-resolution remote sensing imagery to train a deep learning model to extract features of newly planted oil-tea tree forests involves the following steps: Using high-resolution remote sensing images and expert judgment, combined with field research, we visually interpreted the newly planted oil-tea camellia forest area and carried out detailed annotation to obtain vector labels for the newly planted oil-tea camellia forest. The vector labels of the newly planted oil-tea camellia forest are rasterized to obtain binary raster labels. Pairing high-resolution images with raster labels, using non-overlapping sampling techniques, we obtained sample data of newly planted oil-tea camellia forests from the images, and divided the sample data into training and validation sets. Using the training set and the validation set to train a deep learning model to extract multi-level features of the newly planted oil-tea camellia forest; The trained deep learning model is applied to the area to be predicted, and the distribution results of newly planted oil tea forests are output.

4. The method for monitoring camellia oil plantation areas taking into account forest age and growth suitability according to claim 1, characterized in that: Using the distribution of newly planted oil-tea tree forests as a mask, a machine learning model was trained based on multispectral imagery to extract the features of old oil-tea tree forests. The following steps were included: Acquire satellite images and perform data preprocessing to obtain multispectral images; Calculating multispectral image features, terrain features, polarization features, and time-series spectral features using multispectral images, and fusing and adding the features to the images; In the study area, the distribution results of new oil-tea camellia forests were used as masks to select old oil-tea camellia forest sample points in the remaining area and randomly divided them into training set and validation set; Using the training set and the validation set to train a machine learning model to extract multi-level features of the old oil-tea camellia forest; The trained machine learning model is applied to the area to be predicted, and the distribution results of old tea oil forests are output.

5. The method for monitoring oil-tea camellia planting areas taking into account forest age and growth suitability according to claim 4, characterized in that: The method for acquiring satellite images and performing data preprocessing is as follows: using the GEE platform, respectively calling Sentinel-2 or Landsat series satellite images, as well as DEM data and Sentinel-1 images within the same time range, filtering images of a specified time period, removing clouds and selecting the required bands, and performing radiation correction, atmospheric correction, and geometric correction on the images to obtain clear, cloud-free multispectral images.

6. The method for monitoring oil-tea camellia planting areas taking into account forest age and growth suitability according to claim 1, characterized in that: The method for obtaining a growth suitability distribution map of oil-tea camellia comprises the following steps: Collect literature and research data, organize the meteorological and environmental conditions required for the growth of oil-tea camellia, and build a suitability assessment system for oil-tea camellia growth, which is divided into n levels according to the degree of suitability; Obtain meteorological reanalysis data and terrain feature data, reproject, resample, and convert the data types of the data, and combine them with the Camellia oleifera growth suitability assessment system to calculate a multi-dimensional Camellia oleifera growth suitability score; A scoring grid is created in the area to be predicted, and the pixel score is calculated for each environmental factor. The scores of each factor are accumulated to calculate the comprehensive suitability score of each pixel. The comprehensive suitability score is normalized to obtain the oil-tea camellia growth suitability distribution map.

7. The method for monitoring oil-tea camellia planting areas taking into account forest age and growth suitability according to claim 6, characterized in that: The meteorological reanalysis data and topographic feature data are ERA5 reanalysis data and Copernicus topographic data respectively.

8. The method for monitoring oil-tea camellia planting areas taking into account forest age and growth suitability according to claim 1, characterized in that: The method for revising the distribution map of oil-tea camellia planting areas comprises the following steps: Initialize the suitability score threshold and create a Boolean value layer on the area to be predicted to indicate whether the suitability score of the corresponding area pixel is higher than the threshold; Pixels originally identified as oil-tea camellia fields and with suitability scores higher than the threshold are retained; The statistical data of oil-tea camellia planting areas and areas were introduced as auxiliary verification to adjust the suitability score threshold and obtain the distribution of oil-tea camellia planting areas with high confidence.

9. A monitoring system for oil-tea camellia planting areas taking into account forest age and growth suitability, characterized in that: The system includes: a memory and a processor, wherein the memory includes a program for monitoring a tea oil planting area taking into account forest age and growth suitability, and when the program for monitoring a tea oil planting area taking into account forest age and growth suitability is executed by the processor, the steps of a method for monitoring a tea oil planting area taking into account forest age and growth suitability as described in any one of claims 1 to 8 are implemented.

Images