Color steel tile illegal building segmentation detection and geographic information recovery method and system

Abstract

The invention discloses a color steel tile illegal building segmentation detection and geographic information recovery method and system, and the method comprises the steps: obtaining image data of a training region and a test region, screening images containing color steel tile illegal buildings, and forming a color steel tile illegal building data set; marking the color steel tile building, and dividing marked data into training set data and verification set data; constructing an improved neural network model; the color steel tile illegal building data set is input into the improved neural network model for training, and an optimal model for color steel tile illegal building segmentation detection is obtained; the model is used to segment the color steel tile illegal building on the image data of the test area, and the geographic information recovery module is used to recover the geographic information. According to the method, the area and geographic information of the color steel tile illegal building can be obtained, and meanwhile, the advantage of processing a large amount of data is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of illegal building detection, and particularly to a method and system for segmenting and detecting color steel tile illegal buildings and restoring geographic information. Background Art

[0002] How to accurately, automatically and efficiently obtain the locations of such illegal buildings is crucial for accelerating urban development and the urbanization process.

[0003] In recent years, satellite remote sensing technology has developed rapidly. Currently, the resolution of domestic satellite remote sensing images can reach 1m - 0.5m, but it is still far from being able to detect tiny ground targets. The development of unmanned aerial photography technology has better solved this problem. Its advantages of being not restricted by regions and obtaining centimeter-level images have enabled it to be widely applied in the field of urban illegal building detection. At the same time, object detection based on deep learning has been intelligently implemented in various fields. Its efficient computing and autonomous learning of features and rules in images are the manifestations of its technical advantages. However, traditional deep learning models often have problems such as long training time and high performance requirements for the computing platform. Therefore, there is an urgent need for a deep learning model with fast training speed, low platform requirements and excellent performance to achieve rapid detection of color steel tile illegal buildings. Summary of the Invention

[0004] The purpose of this part is to outline some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Simplifications or omissions may be made in this part, as well as in the abstract and title of the present application, to avoid obscuring the purpose of this part, the abstract and the title. However, such simplifications or omissions shall not be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides a method and system for segmenting and detecting color steel tile illegal buildings and restoring geographic information to solve the problems that traditional deep learning models often have long training time and high performance requirements for the computing platform.

[0007] To solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides a method for segmenting and detecting color steel tile illegal buildings and restoring geographic information, including:

[0009] Obtaining image data of the training area and the test area through unmanned aerial photography, and screening images containing color steel tile illegal buildings to form a color steel tile illegal building data set;

[0010] Based on the selected dataset of illegal color steel tile buildings, label the color steel tile buildings, and divide the labeled data into training set data and validation set data;

[0011] Based on the deep learning model, construct an improved neural network model. The improved neural network model includes a five-stage encoder and a four-stage decoder, uses RSU modules with different numbers of layers, and optimizes the input, intermediate, and output channel numbers of all RSU modules through a genetic algorithm;

[0012] Based on the improved neural network model and the divided training set data, input the dataset of illegal color steel tile buildings into the improved neural network model for training to obtain the best model for the segmentation and detection of illegal color steel tile buildings;

[0013] Based on the best model for the segmentation and detection of illegal color steel tile buildings, use this model to segment the image data in the test area for illegal color steel tile buildings, and restore the geographic information through the geographic information restoration module.

[0014] As a preferred solution of the method for segmenting and detecting illegal color steel tile buildings and restoring geographic information according to the present invention, wherein:

[0015] The construction of the improved neural network model includes a five-stage encoder and a four-stage decoder;

[0016] The five-stage encoder is represented as En1, En2, En3, En4, En5;

[0017] The four-stage decoder is represented as De1, De2, De3, De4;

[0018] The RSU modules used in the En1 to En3 stages and the De1 to De3 stages are RSU modules with different numbers of layers, namely RSU-7 module, RSU-6 module, and RSU-5 module;

[0019] The RSU-4F module is used in the En4, En5, and De4 stages, and its number of layers is fixed at 4;

[0020] The input channel number I, intermediate channel number M, and output channel number O of the input data of all RSU modules are optimized and adjusted through a genetic algorithm.

[0021] As a preferred solution of the method for segmenting and detecting illegal color steel tile buildings and restoring geographic information according to the present invention, wherein:

[0022] The genetic algorithm includes the following steps:

[0023] Initialize the population: The population size is 200 individuals, and each individual represents each channel number configuration. The range of the channel number is between [0, 512], and the channel number must be divisible by 2;

[0024] Use GA slope function as the fitness function, and each channel number configuration needs to be trained for 10 epochs on the training dataset;

[0025] If a single epoch exceeds 10 minutes, discard the channel number scheme;

[0026] Perform selection operation according to the fitness evaluation result;

[0027] Randomly select two individuals from the current population as parents, and generate new individuals through crossover operation;

[0028] Perform mutation operation on the newly generated individuals, and set the mutation rate to 0.001;

[0029] Repeat the selection, crossover and mutation operations until the maximum number of iterations reaches 200 times;

[0030] After each iteration, update the population according to the evaluation result of the GA slope function;

[0031] Select the individual with the highest fitness in the final generation population as the best channel number configuration.

[0032] As a preferred solution of the color steel tile illegal building segmentation detection and geographic information restoration method described in the present invention, wherein:

[0033] The GA slope function is expressed as:

[0034]

[0035] where loss 10 is the loos value of the model when trained to 10 epochs, and loss1 is the loos value of the model when trained to the first epoch.

[0036] As a preferred solution of the color steel tile illegal building segmentation detection and geographic information restoration method described in the present invention, wherein:

[0037] Inputting the color steel tile illegal building dataset into the improved neural network model for training includes the following steps:

[0038] Set training parameters: Set the number of iterations to 100;

[0039] The Adam optimizer is adopted, with the initial learning rate of the optimizer set to 0.001 and the weight decay set to 0.0001. The binary cross-entropy with logits is used as the loss function;

[0040] The warm-up operation is completed at the 2nd epoch, and the cosine decay strategy is adopted to update the learning rate.

[0041] As a preferred solution of the color steel tile illegal building segmentation detection and geographic information restoration method described in the present invention, wherein:

[0042] The specific mathematical expression of the cosine decay strategy is:

[0043]

[0044] where i is the number of training rounds; and respectively represent the maximum and minimum values of the learning rate; T cur represents the number of completed rounds of training; T i represents the total number of epochs in the i-th training.

[0045] As a preferred solution of the color steel tile illegal building segmentation detection and geographic information restoration method described in the present invention, wherein:

[0046] The restoration of geographic information through the geographic information restoration module includes the following steps:

[0047] Use the best model for color steel tile illegal building segmentation detection to perform mask extraction on the color steel tile illegal buildings in the test area image patches;

[0048] Input the obtained classification results into the geographic information restoration module and splice them according to the original positions;

[0049] Input the test area UAV images into the geographic information restoration module, and use the GDAL package provided by Python to realize the restoration of geographic information for the spliced classification results.

[0050] In a second aspect, the present invention provides a color steel tile illegal building segmentation detection and geographic information restoration system, including:

[0051] An acquisition module, configured to acquire training area and test area image data based on UAV aerial photography, and screen images containing color steel tile illegal buildings to form a color steel tile illegal building data set;

[0052] A labeling module, configured to label the color steel tile buildings based on the already screened color steel tile illegal building data set, and divide the labeled data into training set data and validation set data;

[0053] The segmentation module constructs an improved neural network model based on a deep learning model. The improved neural network model includes a five-stage encoder and a four-stage decoder, uses RSU modules with different numbers of layers, and optimizes the input, intermediate, and output channel numbers of all RSU modules through a genetic algorithm.

[0054] The input module is used to input the color steel tile illegal building dataset into the improved neural network model for training based on the improved neural network model and the divided training set data, and obtain the best model for color steel tile illegal building segmentation detection.

[0055] The geographic information restoration module is used to segment the color steel tile illegal buildings in the image data of the test area based on the best model for color steel tile illegal building segmentation detection, and restore the geographic information through the geographic information restoration module.

[0056] In a third aspect, the present invention provides a computing device, including:

[0057] A memory for storing programs;

[0058] A processor for executing the computer-executable instructions, and when the computer-executable instructions are executed by the processor, the steps of the color steel tile illegal building segmentation detection and geographic information restoration method are implemented.

[0059] In a fourth aspect, the present invention provides a computer-readable storage medium, including: when the program is executed by a processor, the steps of the color steel tile illegal building segmentation detection and geographic information restoration method are implemented.

[0060] The beneficial effects of the present invention: The T-U2net proposed by the present invention realizes the automatic detection of color steel tile illegal buildings based on UAV images. Compared with the existing models, the T-U2net provided by the present invention can not only process small sample datasets in complex scenarios, but also is superior to the traditional U2net, Unet, and Unet++ in terms of training speed and prediction efficiency. The present invention avoids a large number of manual detection processes, has strong applicability, can complete detection tasks on UAV images in different regions, and is of great significance for quickly and efficiently obtaining urban color steel tile illegal building information. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. Among them:

[0062] Figure 1Schematic diagram of the basic process of a method for detecting and segmenting illegal color steel tile buildings and restoring geographic information provided by an embodiment of the present invention;

[0063] Figure 2 Structural diagram of an improved neural network model of a method for detecting and segmenting illegal color steel tile buildings and restoring geographic information provided by an embodiment of the present invention;

[0064] Figure 3 Schematic diagram of the final extracted mask result of illegal color steel tile buildings and the true distribution of a method for detecting and segmenting illegal color steel tile buildings and restoring geographic information provided by an embodiment of the present invention;

[0065] Figure 4 For a method for detecting and segmenting illegal color steel tile buildings and restoring geographic information provided by an embodiment of the present invention, a comparison chart of the present invention and a typical case in terms of GA slope And training time. Detailed implementation manners

[0066] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will describe the detailed implementation manners of the present invention with reference to the accompanying drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0067] In the following description, many specific details are set forth to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0068] Secondly, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that can be included in at least one implementation manner of the present invention. The "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that excludes other embodiments.

[0069] The present invention will be described in detail in conjunction with the schematic diagrams. When detailing the embodiments of the present invention, for ease of explanation, the cross-sectional views showing the device structure will be enlarged locally not in accordance with the general scale, and the schematic diagrams are only examples and should not limit the protection scope of the present invention herein. In addition, in actual production, three-dimensional spatial dimensions including length, width, and depth should be included.

[0070] Meanwhile, in the description of the present invention, it should be noted that the orientation or positional relationship indicated by terms such as "upper, lower, inner, and outer" is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present invention. In addition, the terms "first, second, or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0071] Unless otherwise clearly specified and defined in the present invention, the terms "installation, connection, and coupling" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can also be a mechanical connection, an electrical connection, or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0072] Embodiment 1

[0073] Referring to Figures 1-3 , an embodiment of the present invention provides a method for detecting and restoring geographical information of illegal color steel tile buildings, as shown in Figure 1 and includes the following steps:

[0074] S1: Obtain the image data of the training area and the test area based on the aerial photography method of the unmanned aerial vehicle (UAV), and screen the images containing illegal color steel tile buildings to form a dataset of illegal color steel tile buildings;

[0075] In the embodiment of the present application, since the image quality of the UAV is easily affected by weather conditions, it is necessary to preferably shoot on a sunny and cloudless day.

[0076] S2: Based on the dataset of illegal color steel tile buildings that has been screened, label the color steel tile buildings, and divide the labeled data into training set data and validation set data;

[0077] In the embodiment of the present application, the Agisoft meatshape software is used for the post-processing of the UAV aerial images to generate high-quality images of the training area and the test area, and the output image format is.tif. Based on the Opencv library provided by Python, the generated images of the training area and the test area are cropped to obtain image blocks with a size of 512*512; the numerous small image blocks generated in the training area are manually screened and the Labelme annotation software is used to label the color steel tile buildings. The labeled data is divided into training and validation datasets according to the ratio of 8:2, and the image blocks generated in the test area are used as test data.

[0078] S3: Based on the deep learning model, construct an improved neural network model. The improved neural network model includes a five-stage encoder and a four-stage decoder, uses RSU modules with different numbers of layers, and optimizes the input, intermediate, and output channel numbers of all RSU modules through a genetic algorithm;

[0079] In the embodiment of this application, based on the U 2 net, construct a neural network model, and on this basis, perform structural adjustment and channel number fine-tuning on it. The improved neural network model is named T-U 2 net, where the genetic algorithm code and the GA slope function are both written in the Python environment. The GA slope function will be used as the fitness function of the genetic algorithm to search for the optimal solutions of the channel numbers of each module of T-U 2 net. Considering that the increase in network complexity will lead to too long training time, the network configurations with too long training time will be excluded, that is, if a single epoch exceeds 10 minutes, the channel number scheme will be discarded.

[0080] In the embodiment of this application, the GA slope function is expressed as:

[0081]

[0082] where loss 10 is the loos value of the model when trained to 10 epochs, and loss1 is the loos value of the model when trained to the 1st epoch.

[0083] Preferably, on the basis of retaining the original large "U" - shaped structure, T-U 2 net performs deployment optimization on the global encoder (RSU (I,M,O) ) and the local encoder (RSU-4F (I,M,O) ). As shown in Figure 2 , the main network of T-U 2 net consists of a five-stage encoder (En1 to En5) and a four-stage decoder (De1 to De4)). In the stages of En1 to En3 and De1 to De3, the corresponding RSU modules are RSU-7 (I,M,O) , RSU-6 (I,M,O) and RSU-5 (I,M,O) respectively; in En4, En5 and De4, the corresponding RSU modules are all RSU-4F (I ,M,O) modules.

[0084] Preferably, RSU-L (I,M,O)(L = 5, 6, 7), where L is the number of layers of the RSU module, and the number of layers of the RSU-4F module is 4; RSU-L (I,M,O) (L = 5, 6, 7, 4F), where I is the number of channels of the input data, M is the number of intermediate channels, and O is the number of channels of the output data.

[0085] Preferably, T-U 2 For all channels inside the net, the optimal number of channels is searched using the genetic algorithm. Each sub-model obtained by the search is trained for 10 epochs on the training dataset, and the rate of decrease of the obtained loss function is used as an indicator to quantify the quality of each model. At the same time, there is a specific ratio relationship between the input and output channel numbers of each RSU module to accelerate convergence, that is, the ratio of M to O of the RSU-L and RSU-4F modules is not less than 1 and I is not greater than 3.

[0086] Preferably, for the En1 stage, the final obtained I is 3, M is 16, and O is 32; for the En2 to En5 stages, I, M, and O are respectively 32, 16, and 32; for the RSU-L of the De1 to De4 stages (I,M,O) (L = 5, 6, 7, 4F) module, I is 64, M is 16, and O is 32 for all.

[0087] In the embodiment of the present application, T-U 2 All channel numbers of the net are randomly initialized within the interval [0, 512], and the initialized channel numbers must be divisible by 2. The parameter settings of the genetic algorithm are: population size 200, number of iterations 200 times, and mutation rate 0.001.

[0088] S4: Based on the improved neural network model and the divided training set data, input the color steel tile illegal building dataset into the improved neural network model for training to obtain the best model for the segmentation detection of color steel tile illegal buildings;

[0089] In the embodiment of the present application, the epoch (number of iterations) is set to 100, the initial learning rate is 0.001. By the end of the 2nd epoch, the warmup operation is completed. During the training process, the Adam optimizer is adopted, and the loss function uses binary cross entropy with logits, and the calling method is torch.nn.functional.binary_cross_entropy_with_logits; the initial learning rate of the optimizer is set to 0.001, the weight decay is set to 0.0001, and other parameters remain default; the learning rate update method adopts both the warmup operation and the Cosine Decay strategy, and the number of cycles of the warmup operation is set to 2.

[0090] In the embodiment of the present application, the specific mathematical expression of the Cosine Decay strategy is:

[0091]

[0092] where i is the number of training rounds; and respectively represent the maximum and minimum values of the learning rate, which define the range of the learning rate; T cur represents how many rounds of training have been completed currently; T i represents the total number of epochs in the i-th training.

[0093] S5: Based on the best model for the segmentation and detection of color steel tile illegal buildings, use this model to segment the image data in the test area for color steel tile illegal buildings, and restore the geographical information through the geographical information restoration module.

[0094] In the embodiment of the present application, as Figure 3 shown, the geographical information restoration module (Geo IRM ) realizes its functions based on the GDAL package provided by Python, mainly including the splicing and geographical information restoration parts. The splicing algorithm splices the numerous color steel tile illegal building mask results output by the segmentation module according to the original positions, so that it covers the complete test area; the geographical information restoration algorithm restores the geographical information of the spliced color steel tile illegal building mask results. The operation of the geographical information restoration module is based on the UAV image data, that is, it can realize the restoration of the geographical position information of the color steel tile illegal building mask results in different regions.

[0095] Preferably, Geo IRMThe functions are implemented using the GDAL package provided by Python. The results output by this module are used for subsequent area calculation and special applications.

[0096] This embodiment also provides a color steel tile illegal building segmentation detection and geographic information restoration system, including:

[0097] An acquisition module, configured to acquire training area and test area image data based on the aerial photography method of an unmanned aerial vehicle, and screen the images containing color steel tile illegal buildings to form a color steel tile illegal building data set;

[0098] A labeling module, configured to label the color steel tile buildings based on the screened color steel tile illegal building data set, and divide the labeled data into training set data and validation set data;

[0099] A segmentation module, which constructs an improved neural network model based on a deep learning model. The improved neural network model includes a five-stage encoder and a four-stage decoder, uses RSU modules with different numbers of layers, and optimizes the input, intermediate, and output channel numbers of all RSU modules through a genetic algorithm;

[0100] An input module, configured to input the color steel tile illegal building data set into the improved neural network model for training based on the improved neural network model and the divided training set data, and obtain the best model for color steel tile illegal building segmentation detection;

[0101] A geographic information restoration module, configured to segment the color steel tile illegal buildings in the image data of the test area based on the best model for color steel tile illegal building segmentation detection, and perform geographic information restoration through the geographic information restoration module.

[0102] Furthermore, it further includes:

[0103] A memory, configured to store programs;

[0104] A processor, configured to load the program to execute the color steel tile illegal building segmentation detection and geographic information restoration method.

[0105] This embodiment also provides a computer-readable storage medium, which stores a program. When the program is executed by a processor, it implements the color steel tile illegal building segmentation detection and geographic information restoration method.

[0106] The storage medium proposed in this embodiment and the color steel tile illegal building segmentation detection and geographic information restoration method proposed in the above embodiment belong to the same inventive concept. The technical details not described in detail in this embodiment can be referred to in the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0107] From the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software and necessary general-purpose hardware. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present invention, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as a floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk or optical disc of a computer, etc., including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of the various embodiments of the present invention.

[0108] Embodiment 2

[0109] Refer to Figure 4 , which is an embodiment of the present invention, provides a method for detecting and dividing illegal color steel tile buildings and restoring geographic information. In order to verify its beneficial effects, the comparison results of two schemes are provided.

[0110] From Figure 3 , it can be seen that the finally searched T-U2net of the present invention can converge quickly in a short time, which indicates that the reasonable design of the channel number of the RSU module can balance the model between GAslope and training time, thus greatly improving the efficiency of the deep learning model. Compared with the artificially designed U2net, Unet and Unet++, the present invention shows obvious advantages, which will provide an important reference for the further exploration of small deep learning models.

[0111] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A method for segmentation detection and geographic information recovery of color steel tile illegal buildings, characterized in that: include: Obtain image data of the training area and the test area, and filter images containing illegal colored steel tile buildings to form a dataset of illegal colored steel tile buildings; Based on the screened dataset of illegal colored steel tile buildings, the colored steel tile buildings are labeled, and the labeled data is divided into training set data and validation set data; Based on the deep learning model, an improved neural network model is constructed, wherein the improved neural network model includes a five-stage encoder and a four-stage decoder, uses RSU modules with different numbers of layers, and optimizes the number of input, intermediate and output channels of all RSU modules through a genetic algorithm; Based on the improved neural network model and the divided training set data, the color steel tile illegal building data set is input into the improved neural network model for training, and the best model for the segmentation detection of color steel tile illegal buildings is obtained; Based on the optimal model for the segmentation and detection of illegal buildings with colored steel tiles, the model is used to segment the image data of the test area, and the geographic information is restored through the geographic information recovery module.

2. The method for segmentation detection and geographic information recovery of color steel tile illegal buildings as claimed in claim 1, characterized in that: The improved neural network model is constructed, including a five-stage encoder and a four-stage decoder; The five-stage encoder is represented as En1, En2, En3, En4, En5; The four-stage decoder is shown as De1, De2, De3, De4; The En1 to En3 phase and the De1 to De3 phase use RSU modules with different numbers of layers, namely RSU-7 module, RSU-6 module, and RSU-5 module; The En4, En5, and De4 stages use RSU-4F modules, with a fixed number of layers of 4; The number of input data channels I, the number of intermediate channels M, and the number of output data channels O of all RSU modules are optimized and adjusted using genetic algorithms.

3. The method for segmentation detection and geographic information recovery of color steel tile illegal buildings as described in claim 1 or 2, characterized in that: The genetic algorithm comprises the following steps: Initialize the population: The population size is 200 individuals, each individual represents each channel number configuration, the channel number range is between [0,512], and the channel number must be divisible by 2; Using GA slope The function is used as the fitness function, and each channel number configuration is trained on the training data set for 10 epochs; If a single epoch exceeds 10 minutes, the channel number scheme will be abandoned; According to the fitness evaluation results, the selection operation is performed; Randomly select two individuals from the current population as parents and generate new individuals through crossover operation; Perform mutation operation on the newly generated individuals, and the mutation rate is set to 0.001; Repeat the selection, crossover and mutation operations until the maximum number of iterations is 200; After each iteration, according to GA slope The evaluation results of the function update the population; In the final generation of the population, the individual with the highest fitness is selected as the optimal channel number configuration.

4. The method for segmentation detection and geographic information recovery of color steel tile illegal buildings as claimed in claim 3, characterized in that: The GA slope The function is expressed as: Among them, loss 10 is the loos value of the model when trained to 10 epochs, and loss1 is the loos value of the model when trained to the first epoch.

5. The method for segmentation detection and geographic information recovery of color steel tile illegal buildings as claimed in claim 4, characterized in that: The step of inputting the color steel tile illegal building data set into the improved neural network model for training includes the following steps: Set training parameters: the number of iterations is set to 100, and the initial learning rate is set to 0.001; The Adam optimizer is used, and the loss function uses binary cross entropy with logits; The warm-up operation is completed in the second epoch, and the cosine decay strategy is used to update the learning rate.

6. The method for segmentation detection and geographic information recovery of color steel tile illegal buildings as claimed in claim 5, characterized in that: The specific mathematical expression of the cosine attenuation strategy is: Among them, i is the round of training; and Respectively represent the maximum and minimum values ​​of the learning rate; T cur Indicates the number of rounds of training that have been completed; T i Represents the total number of epochs in the i-th training.

7. The method for segmentation detection and geographic information recovery of color steel tile illegal buildings as claimed in claim 6, characterized in that: The recovery of geographic information by the geographic information recovery module comprises the following steps: Use the best model for the segmentation and detection of illegal colored steel tile buildings to extract masks from illegal colored steel tile buildings in the image blocks of the test area; The classification results are passed to the geographic information recovery module and spliced ​​according to the original position; The drone images of the test area are transferred to the geographic information recovery module, and the GDAL package provided by Python is used to restore the geographic information of the spliced ​​classification results.

8. A system based on the method for segmentation detection and geographic information recovery of color steel tile illegal buildings according to claim 1, characterized in that: The acquisition module is used to acquire the image data of the training area and the test area based on the drone aerial photography method, and to filter the images containing the illegal buildings with color steel tiles to form the illegal buildings with color steel tiles data set; The labeling module is used to label the color steel tile buildings based on the screened color steel tile illegal building data set, and divide the labeled data into training set data and verification set data; The segmentation module constructs an improved neural network model based on the deep learning model. The improved neural network model includes a five-stage encoder and a four-stage decoder, uses RSU modules with different numbers of layers, and optimizes the number of input, intermediate and output channels of all RSU modules through a genetic algorithm; An input module is used to input the color steel tile illegal building data set into the improved neural network model for training based on the improved neural network model and the divided training set data, so as to obtain the best model for the segmentation detection of color steel tile illegal buildings; The geographic information recovery module is used to segment the illegal colored steel tile buildings based on the best model for segmentation detection of illegal colored steel tile buildings. The model is used to segment the illegal colored steel tile buildings in the image data of the test area, and the geographic information is recovered through the geographic information recovery module.

9. An electronic device, characterized in that: include: Memory, used to store programs; A processor is used to load the program to execute the steps of the method for segmentation detection and geographic information recovery of illegal colored steel tile buildings as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a program, characterized in that: When the program is executed by the processor, the steps of the method for segmentation detection and geographic information recovery of illegal buildings made of colored steel tiles as described in any one of claims 1 to 7 are implemented.

Images