CN115147746B - Saline-alkali geological identification method based on unmanned aerial vehicle remote sensing image - Google Patents
Saline-alkali geological identification method based on unmanned aerial vehicle remote sensing image Download PDFInfo
- Publication number
- CN115147746B CN115147746B CN202211068025.6A CN202211068025A CN115147746B CN 115147746 B CN115147746 B CN 115147746B CN 202211068025 A CN202211068025 A CN 202211068025A CN 115147746 B CN115147746 B CN 115147746B
- Authority
- CN
- China
- Prior art keywords
- target
- saline
- alkali
- preliminary
- region
- 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.)
- Active
Links
- 239000003513 alkali Substances 0.000 title claims abstract description 215
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000012545 processing Methods 0.000 claims abstract description 37
- 238000005457 optimization Methods 0.000 claims abstract description 21
- 238000000605 extraction Methods 0.000 claims abstract description 19
- 238000006243 chemical reaction Methods 0.000 claims abstract description 12
- 238000005520 cutting process Methods 0.000 claims abstract description 8
- 239000002585 base Substances 0.000 claims description 11
- 238000009826 distribution Methods 0.000 claims description 10
- 238000012549 training Methods 0.000 claims description 5
- 238000013507 mapping Methods 0.000 claims description 3
- 230000007704 transition Effects 0.000 claims 1
- 239000002689 soil Substances 0.000 abstract description 21
- 239000011780 sodium chloride Substances 0.000 description 18
- FAPWRFPIFSIZLT-UHFFFAOYSA-M Sodium chloride Chemical group [Na+].[Cl-] FAPWRFPIFSIZLT-UHFFFAOYSA-M 0.000 description 15
- 230000011218 segmentation Effects 0.000 description 9
- 230000010354 integration Effects 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000007596 consolidation process Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 150000003839 salts Chemical class 0.000 description 3
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000000877 morphologic effect Effects 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 239000003673 groundwater Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000003911 water pollution Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/17—Terrestrial scenes taken from planes or by drones
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/90—Dynamic range modification of images or parts thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Remote Sensing (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
Description
技术领域technical field
本发明涉及图像数据处理技术领域，具体涉及基于无人机遥感图像的盐碱地质识别方法。The invention relates to the technical field of image data processing, in particular to a saline-alkali geological identification method based on remote sensing images of drones.
背景技术Background technique
在自然环境影响或人工活动干扰下，土壤以及地下水中的盐分，通过毛细管随底层的水分上升到地表，而待到水分蒸发后，盐分往往聚集在地表土壤中，形成盐碱地。大范围的地质盐碱化，不仅会影响该范围内的土地质量，往往还会造成耕地减产和水体污染，甚至造成生物多样性减少。此外，地质盐碱化可能还会造成很多其他的严重的土地问题，对人类社会造成非常严重的直接或间接的经济损失。因此，对土壤进行盐碱识别至关重要。目前，在对土壤进行盐碱识别时，通常采用的方式为：首先，获取土壤地面图像，接着，将土壤地面图像输入到语义分割网络，然后，通过语义分割网络对土壤地面图像进行提取分割，最后将进行提取分割后得到的多个提取分割图像，分别输入到分类神经网络，通过分类神经网络，对多个提取分割图像进行盐碱程度检测，得到每个提取分割图像对应的盐碱程度。Under the influence of the natural environment or artificial activities, the salt in the soil and groundwater rises to the surface through the capillary with the water in the bottom layer, and after the water evaporates, the salt often accumulates in the surface soil, forming a saline-alkali land. Large-scale geological salinization will not only affect the quality of land within this range, but also often result in reduced cropland production, water pollution, and even a reduction in biodiversity. In addition, geological salinization may also cause many other serious land problems, causing very serious direct or indirect economic losses to human society. Therefore, it is very important to identify the salinity and alkalinity of the soil. At present, when salinity-alkali identification of soil is carried out, the usual method is as follows: firstly, the soil surface image is obtained, then, the soil surface image is input into the semantic segmentation network, and then the soil surface image is extracted and segmented through the semantic segmentation network, Finally, the multiple extracted and segmented images obtained after the extraction and segmentation are input to the classification neural network respectively, and the salinity degree detection is performed on the plurality of extracted and segmented images through the classification neural network, and the corresponding salinity degree of each extracted and segmented image is obtained.
然而，当采用上述方式时，经常会存在如下技术问题：However, when the above method is adopted, the following technical problems often exist:
由于语义分割网络的特性，往往会导致通过语义分割网络对土壤地面图像进行提取分割时，往往没有考虑到像素之间的关系，缺乏空间一致性，对细节不够敏感，因此，往往导致对土壤地面图像的提取分割不够精确，从而导致得到的多个提取分割图像中可能存在拍摄有多种盐碱程度的区域的提取分割图像，往往不能使后续分类神经网络精确的确定提取分割图像对应的盐碱程度，进而导致对土壤进行盐碱识别的准确度低下。Due to the characteristics of the semantic segmentation network, when extracting and segmenting the soil ground image through the semantic segmentation network, the relationship between pixels is often not considered, lacking spatial consistency, and not sensitive enough to details. The extraction and segmentation of the image is not accurate enough, resulting in multiple extraction and segmentation images that may contain extraction and segmentation images of areas with various levels of salinity, often unable to make the subsequent classification neural network accurately determine the corresponding salinity of the extraction and segmentation images Therefore, the accuracy of soil salinity identification is low.
发明内容Contents of the invention
本发明的内容部分用于以简要的形式介绍构思，这些构思将在后面的具体实施方式部分被详细描述。本发明的内容部分并不旨在标识要求保护的技术方案的关键特征或必要特征，也不旨在用于限制所要求的保护的技术方案的范围。The Summary of the Invention serves to introduce concepts in a simplified form that are described in detail in the Detailed Description that follows. The summary of the present invention is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to be used to limit the scope of the claimed technical solution.
为了解决对土壤进行盐碱识别的准确度低下的技术问题，本发明提出了基于无人机遥感图像的盐碱地质识别方法。In order to solve the technical problem of low accuracy in salinity-alkali identification of soil, the present invention proposes a saline-alkali geological identification method based on UAV remote sensing images.
本发明提供了基于无人机遥感图像的盐碱地质识别的方法，该方法包括：The invention provides a method for saline-alkali geological identification based on remote sensing images of drones, the method comprising:
通过安装在目标无人机上的目标相机，获取待检测盐碱情况的目标地面区域的待检测遥感图像；Through the target camera installed on the target UAV, obtain the remote sensing image of the target ground area to be detected for saline-alkali conditions;
对所述待检测遥感图像进行图像增强提取处理，得到目标遥感图像；performing image enhancement and extraction processing on the remote sensing image to be detected to obtain the target remote sensing image;
对所述目标遥感图像进行色彩空间转换处理，得到色彩遥感图像；performing color space conversion processing on the target remote sensing image to obtain a color remote sensing image;
对所述色彩遥感图像进行区域划分优化，得到初步最优盐碱区域集合；Carrying out region division optimization on the color remote sensing image to obtain a preliminary optimal saline-alkali region set;
对所述初步最优盐碱区域集合中的初步最优盐碱区域进行合并划分处理，得到目标盐碱区域集合；Merging and dividing the preliminary optimal saline-alkali regions in the preliminary optimal saline-alkali region set to obtain the target saline-alkali region set;
对所述目标盐碱区域集合中的每个目标盐碱区域进行切割处理，确定所述目标盐碱区域对应的目标盐碱图像，得到目标盐碱图像集合；Carrying out cutting processing on each target saline-alkali area in the target saline-alkali area set, determining a target saline-alkali image corresponding to the target saline-alkali area, and obtaining a target saline-alkali image set;
将所述目标盐碱图像集合中的每个目标盐碱图像输入到训练完成的盐碱程度分类网络，通过所述盐碱程度分类网络，确定所述目标盐碱图像对应的盐碱程度；Input each target saline-alkali image in the target saline-alkali image set to the trained salinity degree classification network, and determine the salinity degree corresponding to the target saline-alkali image through the salinity degree classification network;
根据所述目标盐碱图像集合中的各个目标盐碱图像对应的盐碱程度，生成表征所述目标地面区域的盐碱情况的目标盐碱信息。Target salinity information representing the salinity of the target ground area is generated according to the salinity degree corresponding to each target salinity image in the target salinity image set.
进一步的，所述对所述色彩遥感图像进行区域划分优化，得到初步最优盐碱区域集合，包括：Further, the region division and optimization of the color remote sensing image is carried out to obtain a preliminary optimal saline-alkali region set, including:
将所述色彩遥感图像，映射为目标无向图；Mapping the color remote sensing image into a target undirected graph;
根据所述目标无向图，通过最优化算法，确定所述初步最优盐碱区域集合。According to the target undirected graph, the preliminary optimal saline-alkaline region set is determined through an optimization algorithm.
进一步的，所述最优化算法的目标函数为：Further, the objective function of the optimization algorithm is:
其中，
进一步的，所述对所述初步最优盐碱区域集合中的初步最优盐碱区域进行合并划分处理，得到目标盐碱区域集合，包括：Further, the merging and dividing process is performed on the preliminary optimal saline-alkaline areas in the preliminary optimal saline-alkali area set to obtain the target saline-alkali area set, including:
对所述目标遥感图像进行归一化，得到目标归一化图像；normalizing the remote sensing image of the target to obtain a normalized image of the target;
根据所述初步最优盐碱区域集合中的每个初步最优盐碱区域在所述色彩遥感图像中的位置，确定所述初步最优盐碱区域对应的目标初步区域，得到目标初步区域集合，其中，所述目标初步区域集合中的目标初步区域是所述目标归一化图像中的区域；According to the position of each preliminary optimal saline-alkali area in the preliminary optimal saline-alkali area set in the color remote sensing image, determine the target preliminary area corresponding to the preliminary optimal saline-alkali area, and obtain the target preliminary area set , wherein the target preliminary regions in the set of target preliminary regions are regions in the target normalized image;
将所述目标归一化图像的目标数量个通道中的每个通道的通道值，划分为预设数目个通道等级，得到通道等级集合，其中，所述目标归一化图像的每个通道对应预设数目个通道等级，所述通道等级集合中的通道等级的数量为预设数目的目标数量次幂；Divide the channel value of each channel in the target number of channels of the target normalized image into a preset number of channel levels to obtain a channel level set, wherein each channel of the target normalized image corresponds to A preset number of channel levels, the number of channel levels in the channel level set is the power of the preset number of target numbers;
确定所述目标初步区域集合中的每个目标初步区域对应的颜色通道直方图；determining a color channel histogram corresponding to each target preliminary region in the target preliminary region set;
根据所述目标初步区域集合中的每两个目标初步区域对应的颜色通道直方图和所述通道等级集合，确定所述两个目标初步区域之间的第一区域合并指标；determining a first region merging index between the two target preliminary regions according to the color channel histogram corresponding to each two target preliminary regions in the target preliminary region set and the channel level set;
根据所述目标初步区域集合中的每两个目标初步区域，确定所述两个目标初步区域之间的第二区域合并指标；According to every two target preliminary areas in the set of target preliminary areas, determining a second area merging index between the two target preliminary areas;
根据所述目标初步区域集合中的每两个目标初步区域之间的第一区域合并指标和第二区域合并指标，确定所述两个目标初步区域之间的整体区域合并指标；determining an overall area integration index between the two target preliminary areas according to the first area integration index and the second area integration index between each two target preliminary areas in the target preliminary area set;
当所述目标初步区域集合中的两个目标初步区域之间的整体区域合并指标大于预先设置的判定阈值时，将所述两个目标初步区域划分为同一种目标初步区域；When the overall region merging index between two target preliminary regions in the target preliminary region set is greater than a preset judgment threshold, classify the two target preliminary regions as the same type of target preliminary region;
将同一种目标初步区域中的目标初步区域，确定为所述目标盐碱区域集合中的目标盐碱区域。The target preliminary areas in the same type of target preliminary areas are determined as the target saline-alkali areas in the set of target saline-alkali areas.
进一步的，所述确定所述两个目标初步区域之间的第一区域合并指标对应的公式为：Further, the formula for determining the first area merging index between the two target preliminary areas is:
其中，
进一步的，所述根据所述目标初步区域集合中的每两个目标初步区域，确定所述两个目标初步区域之间的第二区域合并指标，包括：Further, according to each two target preliminary areas in the set of target preliminary areas, determining the second area merging index between the two target preliminary areas includes:
将所述两个目标初步区域进行平移，使所述两个目标初步区域之间的距离为零，得到所述两个目标初步区域对应的拟合并区域；Translating the two preliminary target areas so that the distance between the two preliminary target areas is zero, and obtaining a fitting combined area corresponding to the two preliminary target areas;
根据所述两个目标初步区域和两个目标初步区域对应的拟合并区域，确定所述两个目标初步区域之间的第二区域合并指标。According to the two target preliminary areas and the fitting merged areas corresponding to the two target preliminary areas, a second area combining index between the two target preliminary areas is determined.
进一步的，所述确定所述两个目标初步区域之间的第二区域合并指标对应的公式为：Further, the formula for determining the second area merging index between the two target preliminary areas is:
其中，
进一步的，所述确定所述两个目标初步区域之间的整体区域合并指标对应的公式为：Further, the formula for determining the overall area merging index between the two target preliminary areas is:
其中，
进一步的，所述对所述待检测遥感图像进行图像增强提取处理，得到目标遥感图像，包括：Further, the performing image enhancement and extraction processing on the remote sensing image to be detected to obtain the target remote sensing image includes:
通过图像增强算法，对所述待检测遥感图像进行图像增强处理，得到待检测增强图像；Performing image enhancement processing on the remote sensing image to be detected through an image enhancement algorithm to obtain an enhanced image to be detected;
对所述待检测增强图像进行目标地面区域提取处理，得到所述目标遥感图像。The target ground area extraction process is performed on the enhanced image to be detected to obtain the target remote sensing image.
进一步的，所述盐碱程度分类网络的训练过程，包括：Further, the training process of the salinity degree classification network includes:
构建盐碱程度分类网络；Build a salinity classification network;
获取样本盐碱图像集合，其中，所述样本盐碱图像集合中的样本盐碱图像对应的标签为盐碱程度；Obtain a sample saline-alkali image set, wherein the label corresponding to the sample saline-alkali image in the sample saline-alkali image set is the degree of salinity;
利用所述样本盐碱图像集合和所述样本盐碱图像集合中的各个样本盐碱图像对应的标签，对盐碱程度分类网络进行训练，得到训练完成的盐碱程度分类网络。Using the sample saline-alkali image set and the labels corresponding to each sample saline-alkali image in the sample saline-alkali image set, train the salinity degree classification network to obtain a trained salinity degree classification network.
本发明具有如下有益效果：The present invention has following beneficial effect:
本发明的基于无人机遥感图像的盐碱地质识别方法，通过对待检测遥感图像进行图像处理，解决了对土壤进行盐碱识别的准确度低下的技术问题，提高了对土壤进行盐碱识别的准确度。首先，通过安装在目标无人机上的目标相机，获取待检测盐碱情况的目标地面区域的待检测遥感图像。由于待检测遥感图像上包含目标地面区域的信息，可以便于后续通过对待检测遥感图像进行图像处理，确定待检测遥感图像对应的盐碱情况，从而确定目标地面区域的盐碱情况，可以提高目标地面区域的盐碱情况确定的准确度。其次，对上述待检测遥感图像进行图像增强提取处理，得到目标遥感图像。对待检测遥感图像进行图像增强处理，往往可以提高待检测遥感图像的图像对比度，使得待检测遥感图像的可视化程度更高，可以便于后续对待检测遥感图像进行图像处理。其次，实际情况中，通过目标相机，获取的待检测遥感图像上往往不仅拍摄有目标地面区域，往往还会拍摄到除了目标地面区域之外的区域。然而，除了目标地面区域之外的区域往往不需要进行盐碱情况判断，因此，对上述待检测增强图像进行提取处理，得到只拍摄到目标地面区域的目标遥感图像，可以使目标地面区域之外的区域不再执行后续的步骤，可以减少计算量，可以减少计算资源的占用，可以提高对目标地面区域进行盐碱识别的效率。接着，对上述目标遥感图像进行色彩空间转换处理，得到色彩遥感图像。由于色彩遥感图像可以是HSV图像。HSV图像包括的H（Hue，色调）、S（Saturation，饱和度）和V（Value，明度）三个通道之间的相关性往往较小，往往更符合视觉特性。可以便于后续分析H、S和V三个通道。然后，对上述色彩遥感图像进行区域划分优化，得到初步最优盐碱区域集合。继续，对上述初步最优盐碱区域集合中的初步最优盐碱区域进行合并划分处理，得到目标盐碱区域集合。实际情况中，目标地面区域内的各个位置对应的盐碱程度往往不止一种，即目标地面区域往往可以包括多个盐碱程度不一的区域。因此，对上述色彩遥感图像进行区域划分优化合并，可以得到对应的盐碱程度不同的多个目标盐碱区域，可以便于后续更精确的判断目标地面区域的盐碱情况。之后，对上述目标盐碱区域集合中的每个目标盐碱区域进行切割处理，确定上述目标盐碱区域对应的目标盐碱图像，得到目标盐碱图像集合。将对应的盐碱情况不同的目标盐碱区域切割下来，得到与盐碱程度一一对应的目标盐碱图像，可以便于后续分析目标地面区域对应的多种盐碱程度以及每种盐碱程度在目标地面区域内的位置。而后，将上述目标盐碱图像集合中的每个目标盐碱图像输入到训练完成的盐碱程度分类网络，通过上述盐碱程度分类网络，确定上述目标盐碱图像对应的盐碱程度。最后，根据上述目标盐碱图像集合中的各个目标盐碱图像对应的盐碱程度，生成表征上述目标地面区域的盐碱情况的目标盐碱信息。因此，本发明通过对待检测遥感图像进行图像处理，解决了对土壤进行盐碱识别的准确度低下的技术问题，提高了对土壤进行盐碱识别的准确度。The saline-alkali geological identification method based on the remote sensing image of the unmanned aerial vehicle of the present invention solves the technical problem of low accuracy of saline-alkali identification of soil by performing image processing on the remote sensing image to be detected, and improves the accuracy of saline-alkali identification of soil Spend. First, through the target camera installed on the target UAV, the remote sensing image of the target ground area to be detected in the saline-alkali situation is obtained. Since the remote sensing image to be detected contains the information of the target ground area, it is convenient to perform image processing on the remote sensing image to be detected to determine the saline-alkali condition corresponding to the remote sensing image to be detected, thereby determining the salinity-alkali condition of the target ground area, which can improve the target ground area. The accuracy with which the salinity of the area is determined. Secondly, image enhancement extraction is performed on the remote sensing image to be detected to obtain the target remote sensing image. The image enhancement processing of the remote sensing image to be detected can often improve the image contrast of the remote sensing image to be detected, make the remote sensing image to be detected more visible, and facilitate the subsequent image processing of the remote sensing image to be detected. Secondly, in actual situations, through the target camera, the remote sensing image to be detected often not only captures the target ground area, but often also captures areas other than the target ground area. However, areas other than the target ground area often do not need to be judged on salinity. Therefore, the above-mentioned enhanced images to be detected are extracted and processed to obtain target remote sensing images that only capture the target ground area, which can make the area outside the target ground area Subsequent steps are not performed in the region, which can reduce the amount of calculation, reduce the occupation of computing resources, and improve the efficiency of salinity-alkali identification of the target ground area. Next, color space conversion processing is performed on the target remote sensing image to obtain a color remote sensing image. Since color remote sensing images can be HSV images. The correlation between the three channels of H (Hue, hue), S (Saturation, saturation) and V (Value, lightness) included in the HSV image is often small and often more in line with visual characteristics. It can facilitate the subsequent analysis of the three channels of H, S and V. Then, the above-mentioned color remote sensing images are divided and optimized to obtain a preliminary optimal set of saline-alkali regions. Continue to merge and divide the preliminary optimal saline-alkali regions in the above-mentioned preliminary optimal saline-alkali region set to obtain the target saline-alkali region set. In actual situations, there are usually more than one salinity degree corresponding to each position in the target ground area, that is, the target ground area may often include multiple areas with different salinity levels. Therefore, by performing region division optimization and merging on the above-mentioned color remote sensing images, multiple target saline-alkaline regions with different corresponding salinity degrees can be obtained, which can facilitate subsequent and more accurate judgment of the salinity-alkali condition of the target ground region. Afterwards, a cutting process is performed on each target saline-alkali region in the target saline-alkali region set, and a target saline-alkali image corresponding to the above-mentioned target saline-alkali region is determined to obtain a target saline-alkali image set. The target saline-alkali area with different salinity conditions is cut out to obtain the target saline-alkali image corresponding to the degree of salinity, which can facilitate subsequent analysis of various salinity degrees corresponding to the target ground area and each degree of salinity. The location within the target ground area. Then, each target saline-alkali image in the target saline-alkali image set is input to the trained salinity degree classification network, and the salinity degree corresponding to the above-mentioned target saline-alkali image is determined through the above-mentioned salinity degree classification network. Finally, target salinity information representing the salinity of the target ground area is generated according to the salinity degree corresponding to each target salinity image in the target salinity image set. Therefore, the present invention solves the technical problem of low accuracy of salinity-alkali identification of soil by performing image processing on the remote sensing image to be detected, and improves the accuracy of salinity-alkali identification of soil.
附图说明Description of drawings
为了更清楚地说明本发明实施例或现有技术中的技术方案和优点，下面将对实施例或现有技术描述中所需要使用的附图作简单的介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其它附图。In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or in the prior art, the following will briefly introduce the accompanying drawings that need to be used in the description of the embodiments or the prior art. Apparently, the appended The drawings are only some embodiments of the present invention, and those skilled in the art can also obtain other drawings based on these drawings without creative work.
图1为根据本发明的基于无人机遥感图像的盐碱地质识别方法的一些实施例的流程图；Fig. 1 is the flow chart of some embodiments of the saline-alkali geology identification method based on the remote sensing image of the drone according to the present invention;
图2为根据本发明的待检测增强图像和目标遥感图像示意图；Fig. 2 is a schematic diagram of an enhanced image to be detected and a remote sensing image of a target according to the present invention;
图3为根据本发明的两个目标初步区域之间距离为零时的示意图；Fig. 3 is a schematic diagram according to the present invention when the distance between two target preliminary regions is zero;
图4为根据本发明的目标归一化图像、目标盐碱区域和目标盐碱图像示意图。Fig. 4 is a schematic diagram of a target normalized image, a target saline region and a target saline image according to the present invention.
其中，图2中的附图标记包括：待检测增强图像201、图像区域202和目标遥感图像203。Wherein, reference numerals in FIG. 2 include: an enhanced image to be detected 201 , an
图3中的附图标记包括：第一目标初步区域301和第二目标初步区域302。Reference numerals in FIG. 3 include: a first target
图4中的附图标记包括：目标归一化图像401、第一目标盐碱区域402、第二目标盐碱区域403、第一目标盐碱图像404和第二目标盐碱图像405。Reference numerals in FIG. 4 include: a target normalized
具体实施方式Detailed ways
为了更进一步阐述本发明为达成预定发明目的所采取的技术手段及功效，以下结合附图及较佳实施例，对依据本发明提出的技术方案的具体实施方式、结构、特征及其功效，详细说明如下。在下述说明中，不同的“一个实施例”或“另一个实施例”指的不一定是同一个实施例。此外，一个或多个实施例中的特定特征、结构或特点可由任何合适形式组合。In order to further explain the technical means and effects of the present invention to achieve the intended purpose of the invention, the specific implementation, structure, features and effects of the technical solution proposed according to the present invention will be described in detail below in conjunction with the accompanying drawings and preferred embodiments. described as follows. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, the particular features, structures or characteristics of one or more embodiments may be combined in any suitable manner.
除非另有定义，本文所使用的所有的技术和科学术语与属于本发明的技术领域的技术人员通常理解的含义相同。Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field of the invention.
本发明提供了一种基于无人机遥感图像的盐碱地质识别方法，该方法包括以下步骤：The invention provides a saline-alkali geological identification method based on remote sensing images of drones, the method comprising the following steps:
通过安装在目标无人机上的目标相机，获取待检测盐碱情况的目标地面区域的待检测遥感图像；Through the target camera installed on the target UAV, obtain the remote sensing image of the target ground area to be detected for saline-alkali conditions;
对待检测遥感图像进行图像增强提取处理，得到目标遥感图像；Perform image enhancement and extraction processing on the remote sensing image to be detected to obtain the target remote sensing image;
对目标遥感图像进行色彩空间转换处理，得到色彩遥感图像；Perform color space conversion processing on the target remote sensing image to obtain the color remote sensing image;
对色彩遥感图像进行区域划分优化，得到初步最优盐碱区域集合；The color remote sensing image is divided and optimized to obtain a preliminary optimal set of saline-alkali regions;
对初步最优盐碱区域集合中的初步最优盐碱区域进行合并划分处理，得到目标盐碱区域集合；Merging and dividing the preliminary optimal saline-alkali regions in the preliminary optimal saline-alkali region set to obtain the target saline-alkali region set;
对目标盐碱区域集合中的每个目标盐碱区域进行切割处理，确定目标盐碱区域对应的目标盐碱图像，得到目标盐碱图像集合；Carry out cutting processing on each target saline-alkali area in the target saline-alkali area set, determine the target saline-alkali image corresponding to the target saline-alkali area, and obtain the target saline-alkali image set;
将目标盐碱图像集合中的每个目标盐碱图像输入到训练完成的盐碱程度分类网络，通过盐碱程度分类网络，确定目标盐碱图像对应的盐碱程度；Input each target saline-alkali image in the target saline-alkali image collection to the trained salinity degree classification network, and determine the corresponding salinity degree of the target saline-alkali image through the salinity degree classification network;
根据目标盐碱图像集合中的各个目标盐碱图像对应的盐碱程度，生成表征目标地面区域的盐碱情况的目标盐碱信息。According to the salinity degree corresponding to each target salinity image in the target salinity image set, target salinity information representing the salinity condition of the target ground area is generated.
下面对上述各个步骤进行详细展开：The above steps are described in detail below:
参考图1，示出了根据本发明的基于无人机遥感图像的盐碱地质识别方法的一些实施例的流程。该基于无人机遥感图像的盐碱地质识别方法，包括以下步骤：Referring to FIG. 1 , it shows the flow of some embodiments of the saline-alkali geology identification method based on UAV remote sensing images according to the present invention. The saline-alkali geological identification method based on UAV remote sensing images includes the following steps:
步骤S1，通过安装在目标无人机上的目标相机，获取待检测盐碱情况的目标地面区域的待检测遥感图像。Step S1, through the target camera installed on the target UAV, obtain the remote sensing image of the target ground area where the saline-alkali situation is to be detected to be detected.
在一些实施例中，可以通过安装在目标无人机上的目标相机，获取待检测盐碱情况的目标地面区域的待检测遥感图像。In some embodiments, the remote sensing image of the target ground area to be detected for saline-alkali conditions may be acquired by a target camera installed on the target drone.
其中，上述目标无人机是可以正常飞行的无人机。上述目标相机可以是安装在目标无人机上的遥感相机。上述目标地面区域可以是待检测盐碱情况的地面区域。上述待检测遥感图像可以是待检测盐碱情况的目标地面区域的遥感图像。Wherein, the above-mentioned target drone is a drone that can fly normally. The above-mentioned target camera may be a remote sensing camera installed on the target drone. The aforementioned target ground area may be a ground area where saline-alkali conditions are to be detected. The aforementioned remote sensing image to be detected may be a remote sensing image of a target ground area where saline-alkali conditions are to be detected.
作为示例，可以控制目标无人机移动至目标位置，通过目标相机，获取待检测盐碱情况的目标地面区域的待检测遥感图像。其中，目标位置可以是在目标地面区域的正上方的位置。目标位置与目标地面区域之间的距离可以是预先设置的高度，可以是根据实际情况设置的。目标相机的拍摄角度可以是预先设置的角度，可以是根据实际情况设置的。这些根据实际情况设置的数据，需要保证可以通过目标相机，获取待检测遥感图像。As an example, the target UAV can be controlled to move to the target position, and the target remote sensing image of the target ground area to be detected for saline-alkali conditions can be obtained through the target camera. Wherein, the target position may be a position directly above the target ground area. The distance between the target position and the target ground area may be a preset height, and may be set according to actual conditions. The shooting angle of the target camera may be a preset angle, or may be set according to actual conditions. These data set according to the actual situation need to ensure that the remote sensing image to be detected can be obtained through the target camera.
步骤S2，对待检测遥感图像进行图像增强提取处理，得到目标遥感图像。Step S2, performing image enhancement and extraction processing on the remote sensing image to be detected to obtain the remote sensing image of the target.
在一些实施例中，可以对上述待检测遥感图像进行图像增强提取处理，得到目标遥感图像。In some embodiments, image enhancement and extraction processing may be performed on the remote sensing image to be detected to obtain the remote sensing image of the target.
其中，上述目标遥感图像可以是进行图像增强提取处理后的待检测遥感图像。Wherein, the above target remote sensing image may be a remote sensing image to be detected after image enhancement and extraction processing.
作为示例，本步骤可以包括以下步骤：As an example, this step may include the following steps:
第一步，通过图像增强算法，对上述待检测遥感图像进行图像增强处理，得到待检测增强图像。The first step is to perform image enhancement processing on the remote sensing image to be detected through an image enhancement algorithm to obtain an enhanced image to be detected.
其中，上述待检测增强图像可以是进行图像增强处理后的待检测遥感图像。图像增强算法可以包括但不限于：直方图均衡化算法、拉普拉斯算法和伽马变换算法。Wherein, the above-mentioned enhanced image to be detected may be a remote sensing image to be detected after image enhancement processing. Image enhancement algorithms may include but not limited to: histogram equalization algorithm, Laplacian algorithm and gamma transform algorithm.
第二步，对上述待检测增强图像进行目标地面区域提取处理，得到上述目标遥感图像。The second step is to extract the target ground area from the enhanced image to be detected to obtain the remote sensing image of the target.
其中，上述目标遥感图像可以是目标地面区域对应的区域图像。Wherein, the above remote sensing image of the target may be an area image corresponding to the target ground area.
作为示例，如图2所示，可以对待检测增强图像201进行目标地面区域提取处理，得到目标遥感图像203。其中，图像区域202可以是目标地面区域对应在待检测增强图像201上的区域。将图像区域202，从待检测增强图像201上切割下来，可以得到目标遥感图像203。As an example, as shown in FIG. 2 , target ground region extraction processing may be performed on the enhanced image to be detected 201 to obtain a target
实际情况中，通过目标相机，获取的待检测遥感图像上往往不仅拍摄有目标地面区域，往往还会拍摄到除了目标地面区域之外的区域。然而，除了目标地面区域之外的区域往往不需要进行盐碱情况判断，因此，对上述待检测增强图像进行目标地面区域提取处理，得到只拍摄到目标地面区域的目标遥感图像，可以使目标地面区域之外的区域不再执行后续的步骤，可以减少计算量，可以减少计算资源的占用，可以提高对目标地面区域进行盐碱识别的效率。In actual situations, through the target camera, the remote sensing image to be detected often not only captures the target ground area, but often also captures areas other than the target ground area. However, areas other than the target ground area often do not need to be judged on salinity. Therefore, the target ground area extraction process is performed on the above-mentioned enhanced image to be detected to obtain a target remote sensing image that only captures the target ground area, which can make the target ground area Subsequent steps are not performed in areas outside the area, which can reduce the amount of calculation, reduce the occupation of computing resources, and improve the efficiency of salinity-alkali identification of the target ground area.
步骤S3，对目标遥感图像进行色彩空间转换处理，得到色彩遥感图像。Step S3, performing color space conversion processing on the target remote sensing image to obtain a color remote sensing image.
在一些实施例中，可以对上述目标遥感图像进行色彩空间转换处理，得到色彩遥感图像。In some embodiments, color space conversion processing may be performed on the target remote sensing image to obtain a color remote sensing image.
其中，上述色彩遥感图像可以是目标遥感图像转化成的HSV图像。Wherein, the above-mentioned color remote sensing image may be an HSV image converted from the target remote sensing image.
作为示例，可以通过现有技术，将目标遥感图像，转化为色彩遥感图像。As an example, the target remote sensing image can be converted into a color remote sensing image through the existing technology.
步骤S4，对色彩遥感图像进行区域划分优化，得到初步最优盐碱区域集合。In step S4, region division optimization is performed on the color remote sensing image to obtain a preliminary optimal set of saline-alkali regions.
在一些实施例中，可以对上述色彩遥感图像进行区域划分优化，得到初步最优盐碱区域集合。In some embodiments, region division and optimization may be performed on the above color remote sensing images to obtain a preliminary optimal set of saline-alkali regions.
作为示例，本步骤可以包括以下步骤：As an example, this step may include the following steps:
第一步，将上述色彩遥感图像，映射为目标无向图。The first step is to map the above color remote sensing image into a target undirected graph.
其中，上述目标无向图可以是色彩遥感图像映射的无向图。上述目标无向图的结点可以是色彩遥感图像中的像素点。上述目标无向图的边可以是色彩遥感图像中的像素点之间的连线。Wherein, the above target undirected graph may be an undirected graph of color remote sensing image mapping. The nodes of the above target undirected graph may be pixels in the color remote sensing image. The edges of the above target undirected graph may be the connection lines between the pixels in the color remote sensing image.
第二步，根据上述目标无向图，通过最优化算法，确定上述初步最优盐碱区域集合。In the second step, according to the above-mentioned target undirected graph, through an optimization algorithm, determine the above-mentioned preliminary optimal saline-alkali area set.
其中，上述初步最优盐碱区域集合中的初步最优盐碱区域对应的盐碱程度可以不同。盐碱程度可以是但不限于：正常区域、轻度盐碱区、中度盐碱区和重度盐碱区。Wherein, the salinity degrees corresponding to the preliminary optimal saline-alkaline regions in the above-mentioned preliminary optimal saline-alkali region set may be different. The salinity level can be, but is not limited to: normal area, mildly saline area, moderately saline area, and heavily saline area.
例如，上述最优化算法的目标函数可以为：For example, the objective function of the above optimization algorithm can be:
其中，
当目标函数值F取到最大值时，所划分得到的区域，即为初步最优盐碱区域。When the objective function value F reaches the maximum value, the divided area is the preliminary optimal saline-alkali area.
由于
步骤S5，对初步最优盐碱区域集合中的初步最优盐碱区域进行合并划分处理，得到目标盐碱区域集合。In step S5, the preliminary optimal saline-alkaline areas in the preliminary optimal saline-alkali area set are merged and divided to obtain a target saline-alkali area set.
在一些实施例中，可以对上述初步最优盐碱区域集合中的初步最优盐碱区域进行合并划分处理，得到目标盐碱区域集合。In some embodiments, the preliminary optimal saline-alkali regions in the above-mentioned preliminary optimal saline-alkali region set may be merged and divided to obtain the target saline-alkali region set.
其中，上述目标盐碱区域集合中的各个目标盐碱区域对应的盐碱程度可以不同。Wherein, the saline-alkali degree corresponding to each target saline-alkaline area in the target saline-alkali area set may be different.
作为示例，本步骤可以包括以下步骤：As an example, this step may include the following steps:
第一步，对上述目标遥感图像进行归一化，得到目标归一化图像。The first step is to normalize the remote sensing image of the above target to obtain the normalized image of the target.
其中，上述目标归一化图像可以进行归一化后的目标遥感图像。Wherein, the target normalized image above may be a normalized target remote sensing image.
第二步，根据上述初步最优盐碱区域集合中的每个初步最优盐碱区域在上述色彩遥感图像中的位置，确定上述初步最优盐碱区域对应的目标初步区域，得到目标初步区域集合。In the second step, according to the position of each preliminary optimal saline-alkali area in the above-mentioned preliminary optimal saline-alkali area set in the above-mentioned color remote sensing image, determine the target preliminary area corresponding to the above-mentioned preliminary optimal saline-alkali area, and obtain the target preliminary area gather.
其中，上述目标初步区域集合中的目标初步区域可以是上述目标归一化图像中的区域。Wherein, the target preliminary regions in the above target preliminary region set may be the regions in the above target normalized image.
例如，可以将初步最优盐碱区域在色彩遥感图像中的位置，确定为目标位置。将目标归一化图像中的目标位置所在的区域，确定为该初步最优盐碱区域对应的目标初步区域。其中，初步最优盐碱区域的形状大小可以与该初步最优盐碱区域对应的目标初步区域的形状大小一样。For example, the position of the preliminary optimal saline-alkali area in the color remote sensing image can be determined as the target position. The area where the target position in the target normalized image is located is determined as the preliminary target area corresponding to the preliminary optimal saline-alkali area. Wherein, the shape and size of the preliminary optimal saline-alkaline region may be the same as the shape and size of the target preliminary region corresponding to the preliminary optimal saline-alkali region.
由于色彩遥感图像是进行色彩空间转换后的目标遥感图像，所以，色彩遥感图像和目标遥感图像中相同位置的像素点对应的实际景物可以相同。由于初步最优盐碱区域集合中的初步最优盐碱区域可以是色彩遥感图像中的区域。目标初步区域集合中的目标初步区域可以是目标归一化图像中的区域。目标归一化图像可以是进行归一化后的目标遥感图像。所以，色彩遥感图像、目标遥感图像和目标归一化图像中相同位置的像素点对应的实际景物可以相同。因此，色彩遥感图像、目标遥感图像和目标归一化图像中相同位置的图像区域对应的实际景物可以相同。所以，第一位置可以与第二位置相同。第一位置可以是初步最优盐碱区域在色彩遥感图像中的位置。第二位置可以是初步最优盐碱区域对应的目标初步区域在目标归一化图像中的位置。Since the color remote sensing image is the target remote sensing image after color space conversion, the actual scene corresponding to the pixel at the same position in the color remote sensing image and the target remote sensing image may be the same. Because the preliminary optimal saline-alkali area in the set of preliminary optimal saline-alkali areas can be an area in the color remote sensing image. The target preliminary regions in the set of target preliminary regions may be regions in the target normalized image. The target normalized image may be a normalized remote sensing image of the target. Therefore, the actual scene corresponding to the pixels at the same position in the color remote sensing image, the target remote sensing image and the target normalized image can be the same. Therefore, the actual scene corresponding to the image area at the same position in the color remote sensing image, the target remote sensing image and the target normalized image may be the same. So, the first position can be the same as the second position. The first position may be the position of the preliminary optimal saline-alkaline area in the color remote sensing image. The second position may be the position of the target preliminary area corresponding to the preliminary optimal saline-alkaline area in the target normalized image.
第三步，将上述目标归一化图像的目标数量个通道中的每个通道的通道值，划分为预设数目个通道等级，得到通道等级集合。The third step is to divide the channel value of each channel in the target number of channels of the above-mentioned target normalized image into a preset number of channel levels to obtain a set of channel levels.
其中，上述目标数量可以是目标归一化图像的通道的数量。如，目标归一化图像的通道可以分别为RGB模式下的R（Red，红色）通道、G（Green，绿色）通道和B（Blue，蓝色）通道。上述目标数量可以是3。上述目标归一化图像的每个通道可以对应预设数目个通道等级。上述预设数目可以是预先设置的数目。如，上述预设数目可以是16。上述通道等级集合中的通道等级的数量可以为预设数目的目标数量次幂。如，上述通道等级集合中的通道等级的数量可以为16的3次幂。Wherein, the above target number may be the number of channels of the target normalized image. For example, the channels of the target normalized image may be R (Red, red) channel, G (Green, green) channel and B (Blue, blue) channel in RGB mode respectively. The above target number may be 3. Each channel of the above target normalized image may correspond to a preset number of channel levels. The aforementioned preset number may be a preset number. For example, the aforementioned preset number may be 16. The number of channel levels in the above channel level set may be a preset number raised to the power of the target number. For example, the number of channel levels in the above channel level set may be 16 to the third power.
例如，上述预设数目可以是2。目标归一化图像的R通道的通道值可以包括：0.1，0.2，0.6和0.7。可以将目标归一化图像的R通道的通道值，划分为2个通道等级。其中，0.1和0.2可以为一个通道等级。0.6和0.7可以为另一个通道等级。For example, the aforementioned preset number may be two. The channel values of the R channel of the target normalized image may include: 0.1, 0.2, 0.6 and 0.7. The channel value of the R channel of the target normalized image can be divided into 2 channel levels. Among them, 0.1 and 0.2 can be a channel level. 0.6 and 0.7 can be another channel grade.
第四步，确定上述目标初步区域集合中的每个目标初步区域对应的颜色通道直方图。The fourth step is to determine the color channel histogram corresponding to each target preliminary area in the target preliminary area set.
其中，颜色通道直方图可以是以通道等级为横轴值，以目标像素点的数量为纵轴值的直方图。目标像素点可以是目标初步区域内像素值属于通道等级的像素点。Wherein, the color channel histogram may be a histogram with the channel level as the horizontal axis value and the number of target pixel points as the vertical axis value. The target pixel point may be a pixel point whose pixel value belongs to the channel level in the target preliminary area.
第五步，根据上述目标初步区域集合中的每两个目标初步区域对应的颜色通道直方图和上述通道等级集合，确定上述两个目标初步区域之间的第一区域合并指标。In the fifth step, according to the color channel histogram corresponding to each two target preliminary regions in the above target preliminary region set and the above channel level set, determine the first region merging index between the above two target preliminary regions.
例如，上述确定上述两个目标初步区域之间的第一区域合并指标对应的公式可以为：For example, the above formula for determining the first area merging indicator between the above two target preliminary areas may be:
其中，
由于
第六步，根据上述目标初步区域集合中的每两个目标初步区域，确定上述两个目标初步区域之间的第二区域合并指标。In the sixth step, according to each two target preliminary areas in the target preliminary area set, the second area combination index between the above two target preliminary areas is determined.
例如，本步骤可以包括以下子步骤：For example, this step may include the following substeps:
第一子步骤，将上述两个目标初步区域进行平移，使上述两个目标初步区域之间的距离为零，得到上述两个目标初步区域对应的拟合并区域。In the first sub-step, the above-mentioned two target preliminary regions are translated, so that the distance between the above-mentioned two target preliminary regions is zero, and the fitting merged region corresponding to the above-mentioned two target preliminary regions is obtained.
其中，两个目标初步区域对应的拟合并区域可以是这两个目标初步区域合并在一起，得到的区域。Wherein, the fitting merged area corresponding to the two target preliminary areas may be an area obtained by merging these two target preliminary areas.
比如，当两个目标初步区域之间的距离大于零时，平移上述两个目标初步区域，使上述两个目标初步区域之间的距离为零，得到上述两个目标初步区域对应的拟合并区域。For example, when the distance between the two preliminary target regions is greater than zero, the above two preliminary target regions are shifted so that the distance between the two preliminary target regions is zero, and the fitting and sum corresponding to the above two preliminary target regions is obtained. area.
又如，当两个目标初步区域之间的距离为零时，不需要平移上述两个目标初步区域，可以直接将上述两个目标初步区域合并在一起，得到上述两个目标初步区域对应的拟合并区域。如图3所示，第一目标初步区域301和第二目标初步区域302之间的距离为零。第一目标初步区域301和第二目标初步区域302可以是两个目标初步区域。As another example, when the distance between the two preliminary target regions is zero, the above two preliminary target regions do not need to be translated, and the above two preliminary target regions can be directly merged together to obtain the pseudo Merge regions. As shown in FIG. 3 , the distance between the first
第二子步骤，根据上述两个目标初步区域和两个目标初步区域对应的拟合并区域，确定上述两个目标初步区域之间的第二区域合并指标。In the second sub-step, according to the above-mentioned two target preliminary areas and the fitting combined areas corresponding to the two target preliminary areas, determine the second area merging index between the above-mentioned two target preliminary areas.
比如，上述确定上述两个目标初步区域之间的第二区域合并指标对应的公式可以为：For example, the above-mentioned formula for determining the second area merger index between the above-mentioned two target preliminary areas can be:
其中，
第七步，根据上述目标初步区域集合中的每两个目标初步区域之间的第一区域合并指标和第二区域合并指标，确定上述两个目标初步区域之间的整体区域合并指标。The seventh step is to determine the overall area integration index between the above two target preliminary areas according to the first area integration index and the second area integration index between each two target preliminary areas in the above target preliminary area set.
例如，上述确定上述两个目标初步区域之间的整体区域合并指标对应的公式可以为：For example, the formula corresponding to the above-mentioned determination of the overall area merging index between the above-mentioned two target preliminary areas may be:
其中，
由于第p个目标初步区域和第q个目标初步区域之间的第一区域合并指标
第八步，当上述目标初步区域集合中的两个目标初步区域之间的整体区域合并指标大于预先设置的判定阈值时，将上述两个目标初步区域划分为同一种目标初步区域。In the eighth step, when the overall region merging index between the two target preliminary regions in the target preliminary region set is greater than a preset judgment threshold, the above two target preliminary regions are classified as the same type of target preliminary region.
其中，判定阈值可以是认为两个目标初步区域不是同一种目标初步区域的最小的整体区域合并指标。例如，判定阈值可以是0.75。同一种目标初步区域可以包括盐碱程度相同的目标初步区域。Wherein, the determination threshold may be the smallest overall region merging index that considers that two preliminary target regions are not of the same kind of preliminary target region. For example, the decision threshold may be 0.75. The same target preliminary area may include target preliminary areas with the same salinity degree.
第九步，将同一种目标初步区域中的目标初步区域，确定为上述目标盐碱区域集合中的目标盐碱区域。In the ninth step, the target preliminary areas in the same type of target preliminary areas are determined as the target saline-alkali areas in the set of target saline-alkali areas.
步骤S6，对目标盐碱区域集合中的每个目标盐碱区域进行切割处理，确定目标盐碱区域对应的目标盐碱图像，得到目标盐碱图像集合。Step S6 , cutting each target saline-alkali region in the target saline-alkali region set, determining the target saline-alkali image corresponding to the target saline-alkali region, and obtaining the target saline-alkali image set.
在一些实施例中，可以对上述目标盐碱区域集合中的每个目标盐碱区域进行切割处理，确定上述目标盐碱区域对应的目标盐碱图像，得到目标盐碱图像集合。In some embodiments, each target saline-alkali region in the target saline-alkali region set may be cut to determine a target saline-alkali image corresponding to the target saline-alkali region to obtain a target saline-alkali image set.
作为示例，如图4所示，可以将目标归一化图像401中的第一目标盐碱区域402和第二目标盐碱区域403进行切割处理，得到第一目标盐碱图像404和第二目标盐碱图像405。其中，第一目标盐碱区域402和第二目标盐碱区域403可以是两个目标盐碱区域。第一目标盐碱图像404和第二目标盐碱图像405可以是两个目标盐碱图像。As an example, as shown in FIG. 4 , the first target saline-
步骤S7，将目标盐碱图像集合中的每个目标盐碱图像输入到训练完成的盐碱程度分类网络，通过盐碱程度分类网络，确定目标盐碱图像对应的盐碱程度。Step S7, input each target salinity image in the target salinity image set to the trained salinity degree classification network, and determine the salinity degree corresponding to the target salinity degree classification network through the salinity degree classification network.
在一些实施例中，可以将上述目标盐碱图像集合中的每个目标盐碱图像输入到训练完成的盐碱程度分类网络，通过上述盐碱程度分类网络，确定上述目标盐碱图像对应的盐碱程度。In some embodiments, each target saline-alkaline image in the target saline-alkali image set can be input to the trained salinity degree classification network, and the salt corresponding to the above-mentioned target saline-alkali image is determined through the above-mentioned salinity degree classification network. alkalinity.
其中，盐碱程度分类网络可以用于判断目标盐碱图像对应的盐碱程度。盐碱程度分类网络可以是分类识别神经网络。Among them, the salinity degree classification network can be used to judge the salinity degree corresponding to the target salinity image. The salinity degree classification network may be a classification recognition neural network.
作为示例，上述盐碱程度分类网络的训练过程，可以包括以下步骤：As an example, the training process of the above-mentioned salinity degree classification network may include the following steps:
第一步，构建盐碱程度分类网络。The first step is to construct a salinity classification network.
本步骤可以通过现有技术实现，在此不再赘述。This step can be implemented through existing technologies, and will not be repeated here.
第二步，获取样本盐碱图像集合。The second step is to obtain a collection of sample saline-alkali images.
其中，样本盐碱图像集合中的样本盐碱图像可以是已知盐碱程度的地面图像。上述样本盐碱图像集合中的样本盐碱图像对应的标签可以为盐碱程度。Wherein, the sample saline-alkali image in the sample saline-alkali image set may be a ground image with a known degree of salinity. The label corresponding to the sample saline-alkaline image in the sample saline-alkali image set may be the degree of salinity.
第三步，利用上述样本盐碱图像集合和上述样本盐碱图像集合中的各个样本盐碱图像对应的标签，对盐碱程度分类网络进行训练，得到训练完成的盐碱程度分类网络。The third step is to use the sample saline-alkali image set and the labels corresponding to each sample saline-alkali image in the sample saline-alkali image set to train the salinity degree classification network to obtain the trained salinity degree classification network.
其中，训练盐碱程度分类网络的损失函数可以是交叉熵损失函数。Wherein, the loss function for training the salinity degree classification network may be a cross-entropy loss function.
本步骤可以通过现有技术实现，在此不再赘述。This step can be implemented through existing technologies, and will not be repeated here.
步骤S8，根据目标盐碱图像集合中的各个目标盐碱图像对应的盐碱程度，生成表征目标地面区域的盐碱情况的目标盐碱信息。Step S8, according to the salinity level corresponding to each target salinity image in the target salinity image set, generate target salinity information representing the salinity condition of the target ground area.
在一些实施例中，可以根据上述目标盐碱图像集合中的各个目标盐碱图像对应的盐碱程度，生成表征上述目标地面区域的盐碱情况的目标盐碱信息。In some embodiments, the target salinity information representing the salinity condition of the target ground area may be generated according to the salinity degree corresponding to each target salinity image in the target salinity image set.
作为示例，目标盐碱图像集合中的目标盐碱图像对应的盐碱程度可以分别为中度盐碱区和重度盐碱区。生成的表征目标地面区域的盐碱情况的目标盐碱信息可以是“该目标地面区域的中间部分是重度盐碱区，其余区域均为中度盐碱区，盐碱程度比较严重”。As an example, the salinity degrees corresponding to the target saline-alkaline images in the target saline-alkali image set may be moderately saline-alkali areas and severe saline-alkali areas. The generated target salinity information representing the salinity of the target ground area may be "the middle part of the target ground area is a severe saline area, and the rest of the areas are moderately saline areas, and the degree of salinity is relatively serious".
本发明的基于无人机遥感图像的盐碱地质识别方法，通过对待检测遥感图像进行图像处理，解决了对土壤进行盐碱识别的准确度低下的技术问题，提高了对土壤进行盐碱识别的准确度。首先，通过安装在目标无人机上的目标相机，获取待检测盐碱情况的目标地面区域的待检测遥感图像。由于待检测遥感图像上包含目标地面区域的信息，可以便于后续通过对待检测遥感图像进行图像处理，确定待检测遥感图像对应的盐碱情况，从而确定目标地面区域的盐碱情况，可以提高目标地面区域的盐碱情况确定的准确度。其次，对上述待检测遥感图像进行图像增强提取处理，得到目标遥感图像。对待检测遥感图像进行图像增强处理，往往可以提高待检测遥感图像的图像对比度，使得待检测遥感图像的可视化程度更高，可以便于后续对待检测遥感图像进行图像处理。其次，实际情况中，通过目标相机，获取的待检测遥感图像上往往不仅拍摄有目标地面区域，往往还会拍摄到除了目标地面区域之外的区域。然而，除了目标地面区域之外的区域往往不需要进行盐碱情况判断，因此，对上述待检测增强图像进行提取处理，得到只拍摄到目标地面区域的目标遥感图像，可以使目标地面区域之外的区域不再执行后续的步骤，可以减少计算量，可以减少计算资源的占用，可以提高对目标地面区域进行盐碱识别的效率。接着，对上述目标遥感图像进行色彩空间转换处理，得到色彩遥感图像。由于色彩遥感图像可以是HSV图像。HSV图像包括的H（Hue，色调）、S（Saturation，饱和度）和V（Value，明度）三个通道之间的相关性往往较小，往往更符合视觉特性。可以便于后续分析H、S和V三个通道。然后，对上述色彩遥感图像进行区域划分优化，得到初步最优盐碱区域集合。继续，对上述初步最优盐碱区域集合中的初步最优盐碱区域进行合并划分处理，得到目标盐碱区域集合。实际情况中，目标地面区域内的各个位置对应的盐碱程度往往不止一种，即目标地面区域往往可以包括多个盐碱程度不一的区域。因此，对上述色彩遥感图像进行区域划分优化合并，可以得到对应的盐碱程度不同的多个目标盐碱区域，可以便于后续更精确的判断目标地面区域的盐碱情况。之后，对上述目标盐碱区域集合中的每个目标盐碱区域进行切割处理，确定上述目标盐碱区域对应的目标盐碱图像，得到目标盐碱图像集合。将对应的盐碱情况不同的目标盐碱区域切割下来，得到与盐碱程度一一对应的目标盐碱图像，可以便于后续分析目标地面区域对应的多种盐碱程度以及每种盐碱程度在目标地面区域内的位置。而后，将上述目标盐碱图像集合中的每个目标盐碱图像输入到训练完成的盐碱程度分类网络，通过上述盐碱程度分类网络，确定上述目标盐碱图像对应的盐碱程度。最后，根据上述目标盐碱图像集合中的各个目标盐碱图像对应的盐碱程度，生成表征上述目标地面区域的盐碱情况的目标盐碱信息。因此，本发明通过对待检测遥感图像进行图像处理，解决了对土壤进行盐碱识别的准确度低下的技术问题，提高了对土壤进行盐碱识别的准确度。The saline-alkali geological identification method based on the remote sensing image of the unmanned aerial vehicle of the present invention solves the technical problem of low accuracy of saline-alkali identification of soil by performing image processing on the remote sensing image to be detected, and improves the accuracy of saline-alkali identification of soil Spend. First, through the target camera installed on the target UAV, the remote sensing image of the target ground area to be detected in the saline-alkali situation is obtained. Since the remote sensing image to be detected contains the information of the target ground area, it is convenient to perform image processing on the remote sensing image to be detected to determine the saline-alkali condition corresponding to the remote sensing image to be detected, thereby determining the salinity-alkali condition of the target ground area, which can improve the target ground area. The accuracy with which the salinity of the area is determined. Secondly, image enhancement extraction is performed on the remote sensing image to be detected to obtain the target remote sensing image. The image enhancement processing of the remote sensing image to be detected can often improve the image contrast of the remote sensing image to be detected, make the remote sensing image to be detected more visible, and facilitate the subsequent image processing of the remote sensing image to be detected. Secondly, in actual situations, through the target camera, the remote sensing image to be detected often not only captures the target ground area, but often also captures areas other than the target ground area. However, areas other than the target ground area often do not need to be judged on salinity. Therefore, the above-mentioned enhanced images to be detected are extracted and processed to obtain target remote sensing images that only capture the target ground area, which can make the area outside the target ground area Subsequent steps are not performed in the region, which can reduce the amount of calculation, reduce the occupation of computing resources, and improve the efficiency of salinity-alkali identification of the target ground area. Next, color space conversion processing is performed on the target remote sensing image to obtain a color remote sensing image. Since color remote sensing images can be HSV images. The correlation between the three channels of H (Hue, hue), S (Saturation, saturation) and V (Value, lightness) included in the HSV image is often small and often more in line with visual characteristics. It can facilitate the subsequent analysis of the three channels of H, S and V. Then, the above-mentioned color remote sensing images are divided and optimized to obtain a preliminary optimal set of saline-alkali regions. Continue to merge and divide the preliminary optimal saline-alkali regions in the above-mentioned preliminary optimal saline-alkali region set to obtain the target saline-alkali region set. In actual situations, there are usually more than one salinity degree corresponding to each position in the target ground area, that is, the target ground area may often include multiple areas with different salinity levels. Therefore, by performing region division optimization and merging on the above-mentioned color remote sensing images, multiple target saline-alkaline regions with different corresponding salinity degrees can be obtained, which can facilitate subsequent and more accurate judgment of the salinity-alkali condition of the target ground region. Afterwards, a cutting process is performed on each target saline-alkali region in the target saline-alkali region set, and a target saline-alkali image corresponding to the above-mentioned target saline-alkali region is determined to obtain a target saline-alkali image set. The target saline-alkali area with different salinity conditions is cut out to obtain the target saline-alkali image corresponding to the degree of salinity, which can facilitate subsequent analysis of various salinity degrees corresponding to the target ground area and each degree of salinity. The location within the target ground area. Then, each target saline-alkali image in the target saline-alkali image set is input to the trained salinity degree classification network, and the salinity degree corresponding to the above-mentioned target saline-alkali image is determined through the above-mentioned salinity degree classification network. Finally, target salinity information representing the salinity of the target ground area is generated according to the salinity degree corresponding to each target salinity image in the target salinity image set. Therefore, the present invention solves the technical problem of low accuracy of salinity-alkali identification of soil by performing image processing on the remote sensing image to be detected, and improves the accuracy of salinity-alkali identification of soil.
以上上述实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围，均应包含在本申请的保护范围之内。The above-mentioned embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still apply to the foregoing embodiments Modifications to the technical solutions described, or equivalent replacement of some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the application, and should be included in this application. within the scope of the application.
Claims (7)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211068025.6A CN115147746B (en) | 2022-09-02 | 2022-09-02 | Saline-alkali geological identification method based on unmanned aerial vehicle remote sensing image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211068025.6A CN115147746B (en) | 2022-09-02 | 2022-09-02 | Saline-alkali geological identification method based on unmanned aerial vehicle remote sensing image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115147746A CN115147746A (en) | 2022-10-04 |
| CN115147746B true CN115147746B (en) | 2022-11-29 |
Family
ID=83415144
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211068025.6A Active CN115147746B (en) | 2022-09-02 | 2022-09-02 | Saline-alkali geological identification method based on unmanned aerial vehicle remote sensing image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115147746B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116310842B (en) * | 2023-05-15 | 2023-08-04 | 菏泽市国土综合整治服务中心 | Soil saline-alkali area identification and division method based on remote sensing image |
| CN117630337B (en) * | 2024-01-04 | 2024-08-16 | 中国科学院华南植物园 | A method for monitoring coral sand salinity based on drones |
| CN117953400B (en) * | 2024-02-23 | 2024-08-23 | 广州市市政集团设计院有限公司 | Green micro-reconstruction system for old and old urban communities |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101614818A (en) * | 2009-07-09 | 2009-12-30 | 中国科学院遥感应用研究所 | A Radar Remote Sensing Monitoring Method for Soil Salinization |
| CN107784657A (en) * | 2017-09-29 | 2018-03-09 | 西安因诺航空科技有限公司 | A kind of unmanned aerial vehicle remote sensing image partition method based on color space classification |
| CN108256621A (en) * | 2016-12-28 | 2018-07-06 | 北京天诚同创电气有限公司 | Bee colony optimization method and device |
| CN108680509A (en) * | 2018-08-17 | 2018-10-19 | 山东农业大学 | A method for estimating soil salinity content in coastal saline areas |
| CN109447111A (en) * | 2018-09-20 | 2019-03-08 | 杭州师范大学 | A kind of remote sensing supervised classification method based on subclass training sample |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108647703B (en) * | 2018-04-19 | 2021-11-02 | 北京联合大学 | A Type Judgment Method of Saliency-Based Classified Image Library |
| CN114494081B (en) * | 2022-04-01 | 2022-07-05 | 武汉大学 | Unmanned aerial vehicle remote sensing mapping image enhancement method |
-
2022
- 2022-09-02 CN CN202211068025.6A patent/CN115147746B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101614818A (en) * | 2009-07-09 | 2009-12-30 | 中国科学院遥感应用研究所 | A Radar Remote Sensing Monitoring Method for Soil Salinization |
| CN108256621A (en) * | 2016-12-28 | 2018-07-06 | 北京天诚同创电气有限公司 | Bee colony optimization method and device |
| CN107784657A (en) * | 2017-09-29 | 2018-03-09 | 西安因诺航空科技有限公司 | A kind of unmanned aerial vehicle remote sensing image partition method based on color space classification |
| CN108680509A (en) * | 2018-08-17 | 2018-10-19 | 山东农业大学 | A method for estimating soil salinity content in coastal saline areas |
| CN109447111A (en) * | 2018-09-20 | 2019-03-08 | 杭州师范大学 | A kind of remote sensing supervised classification method based on subclass training sample |
Non-Patent Citations (3)
| Title |
|---|
| Land Cover Classification by using of Multi-source Remote Sensing Image based on ELM;Qing Ding et al.;《2019 SAR in Big Data Era (BIGSARDATA)》;20191007;1-4 * |
| 基于图像分割和密度聚类的遥感动目标分块提取;林翊钧等;《北京航空航天大学学报》;20180828;第44卷(第12期);2510-2520 * |
| 基于改进的遗传算法的多目标优化问题研究;孔德剑;《计算机仿真》;20120215;第29卷(第2期);213-215 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115147746A (en) | 2022-10-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115147746B (en) | Saline-alkali geological identification method based on unmanned aerial vehicle remote sensing image | |
| CN113160192B (en) | Visual sense-based snow pressing vehicle appearance defect detection method and device under complex background | |
| CN108416307B (en) | An aerial image pavement crack detection method, device and equipment | |
| CN111340824B (en) | An Image Feature Segmentation Method Based on Data Mining | |
| CN104966085B (en) | A kind of remote sensing images region of interest area detecting method based on the fusion of more notable features | |
| CN110309781B (en) | House damage remote sensing identification method based on multi-scale spectrum texture self-adaptive fusion | |
| CN111275696B (en) | Medical image processing method, image processing method and device | |
| CN112150493B (en) | Semantic guidance-based screen area detection method in natural scene | |
| CN107229917B (en) | A common salient target detection method for multiple remote sensing images based on iterative clustering | |
| CN113095263B (en) | Training method and device for pedestrian re-recognition model under shielding and pedestrian re-recognition method and device under shielding | |
| CN106934386B (en) | A kind of natural scene character detecting method and system based on from heuristic strategies | |
| CN107909081B (en) | A fast acquisition and fast calibration method for image datasets in deep learning | |
| CN107103317A (en) | Fuzzy license plate image recognition algorithm based on image co-registration and blind deconvolution | |
| CN112818775B (en) | Method and system for fast identification of forest roads based on region boundary pixel exchange | |
| CN110738676A (en) | A GrabCut Automatic Segmentation Algorithm Combining RGBD Data | |
| CN103020985B (en) | A kind of video image conspicuousness detection method based on field-quantity analysis | |
| CN109360179B (en) | Image fusion method and device and readable storage medium | |
| CN108629286B (en) | Remote sensing airport target detection method based on subjective perception significance model | |
| CN107392968B (en) | Image saliency detection method fused with color contrast map and color space distribution map | |
| CN107291855A (en) | A kind of image search method and system based on notable object | |
| CN109886146B (en) | Flood information remote sensing intelligent acquisition method and device based on machine vision detection | |
| CN106846322B (en) | The SAR image segmentation method learnt based on curve wave filter and convolutional coding structure | |
| CN104217440B (en) | A kind of method extracting built-up areas from remote sensing images | |
| CN109766823A (en) | A high-resolution remote sensing ship detection method based on deep convolutional neural network | |
| CN109447111A (en) | A kind of remote sensing supervised classification method based on subclass training sample |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20250116 Address after: No. 767 Huimin Street, Nanmingshan Street, Liandu District, Lishui City, Zhejiang Province, China 323000, Building 1-1, A202 Patentee after: Zhejiang Rongqi Technology Co.,Ltd. Country or region after: China Address before: 3A02, Block A, Duchuang Digital Innovation Center, Building 12, Jinhuafa Industrial Park, Helian Community, Longhua Street, Longhua District, Shenzhen, Guangdong 518000 Patentee before: GUANGDONG RONGQE INTELLIGENT TECHNOLOGY Co.,Ltd. Country or region before: China |