Target detection optimization method and system based on artificial intelligence

Abstract

The invention discloses a target detection optimization method and system based on artificial intelligence in the technical field of artificial intelligence. The system comprises a recognition device, a monitoring module and a segmentation module. The monitoring module is used for monitoring a tree shadow position and shadow change through an identification device; the segmentation module is used for segmenting the picture into a plurality of small grids; the system is characterized by comprising a learning prediction module, an analysis and judgment module and a control module, the learning prediction module is used for learning types of trees, growth states of the trees and sunrise and sunset time, and predicting changes of tree shadows in the road; the analysis and judgment module is used for analyzing the small checks and transmitting the small checks to the control module; and the control module is used for carrying out compensation processing on the areas of the shadow small checks according to the received position information of the shadow small checks. Through the identification device, the monitoring module, the segmentation module, the learning prediction module and the control module, interference generated by tree shadows is reduced when the identification device identifies the vehicle information.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an artificial intelligence-based target detection optimization method and system. Background Art

[0002] With the acceleration of urbanization and the continuous growth of traffic demand, road vehicle recognition has become an indispensable part of the intelligent transportation system. Traditional manual monitoring and recognition methods can no longer meet the needs of modern traffic management, and there are problems such as low efficiency and prone to errors. Therefore, the application of target detection optimization methods and systems based on artificial intelligence in the field of road vehicle recognition is particularly important.

[0003] Road vehicle recognition technology aims to achieve automatic detection, classification and tracking of vehicle images through advanced computer vision and deep learning algorithms. The core of this technology is to improve the accuracy and real-time performance of recognition to cope with complex and changing traffic environments and diverse vehicle types. However, in practical applications, road vehicle recognition faces many challenges, such as lighting changes, tree shadows, and vehicle occlusions.

[0004] The existing target detection optimization methods and systems for detecting vehicles in tree shadow areas currently rely mainly on image processing technology and license plate recognition algorithms, especially combined with deep learning methods. The shadows cast by trees on the road will significantly reduce the visibility and clarity of the license plate, making it difficult for the license plate recognition algorithm to accurately identify the license plate number. In particular, when the shadow covers most of the license plate area, the recognition rate will drop significantly, and interference factors such as shadows and tree textures may cause the license plate recognition algorithm to mistake non-license plate areas for license plates, or mistake characters on the license plate for other characters, thereby increasing the misrecognition rate. Therefore, it is necessary to propose an artificial intelligence-based target detection optimization method and system. Summary of the invention

[0005] In order to solve the above problems, the purpose of the present invention is to provide a target detection optimization method and system based on artificial intelligence, which reduces the interference caused by tree shadows when the recognition device recognizes vehicle information by learning a prediction module, a monitoring module, a segmentation module, an analysis and judgment module, a control module and an identification device for identifying road vehicle information.

[0006] In order to achieve the above-mentioned object, the technical solution of the present invention is as follows: an artificial intelligence-based target detection system, comprising an identification device for identifying road vehicle information, a monitoring module and a segmentation module;

[0007] The recognition device includes a camera, and the camera is used to recognize vehicle information, tree information, and tree shadow information on the road;

[0008] A monitoring module, used for monitoring the position and change of shadows cast on the road by trees on the actual road through an identification device;

[0009] A segmentation module is used to segment the entire picture monitored by the monitoring module through the recognition device into a number of small squares;

[0010] It is characterized by: comprising a learning prediction module, an analysis and judgment module and a control module;

[0011] A learning prediction module is used to learn the types of trees and the growth status of the trees as well as the sunrise and sunset times of different years, months and days, and the learning prediction module is also used to predict the changes of shadows produced by the trees on the road according to the learned sunrise and sunset times;

[0012] An analysis and judgment module is used to analyze the small squares, identify the small squares with shadows in the picture and transmit the position information of the small squares with shadows in the picture to the control module;

[0013] The control module is used to perform compensation processing on the areas of the shadow small squares according to the received position information of the shadow small squares.

[0014] The principle of the basic scheme is: predict the changes in shadows cast by trees on the road through a learning prediction module, divide the picture recognized by the recognition device into several small squares through a segmentation module, determine the location of the small squares where the shadows exist through an analysis and judgment module, and perform regional compensation for the small squares where the shadows exist through a control module.

[0015] The beneficial effects of the basic scheme are: 1. By learning and predicting the changing patterns of tree shadows through the learning prediction module and combining image processing technology in actual monitoring, the interference of tree shadows on vehicle recognition can be effectively reduced, and the accuracy of recognition can be significantly improved.

[0016] 2. The image is divided into small squares through the segmentation module, and only the small squares with shadows are compensated, which reduces unnecessary calculations and improves processing efficiency.

[0017] Furthermore, the learning prediction module includes a learning unit and a prediction unit. The learning unit is constructed based on the trained convolutional network model and is trained by inputting sunrise and sunset times. The convolutional neural network is used to output light intensity and angle changes in different time periods. The prediction unit is used to predict changes in shadows cast on the road by trees under the action of light based on the light intensity and angle changes in different time periods output by the learning unit and the tree information recognized by the camera.

[0018] The beneficial effect of the basic scheme is that by building a learning unit based on the trained convolutional network model, the system can more accurately learn and understand the impact of sunrise and sunset times on light intensity and angle changes. This deep learning method can capture complex light change patterns, thereby improving the prediction accuracy of tree shadow changes. The prediction unit can quickly predict the changes in tree shadows cast on the road under the action of light based on the output of the learning unit and the tree information recognized by the camera in real time.

[0019] Furthermore, the monitoring module is also used to monitor the appearance and speed of vehicles traveling on the road, including monitoring the appearance and speed of beyond-visual-range vehicles.

[0020] The beneficial effects of the basic scheme are: by monitoring the shape and speed of beyond-visual-range vehicles, it is helpful to predict and judge the type and speed of the vehicle in advance, and when the vehicle is identified by the identification device, it is helpful to reduce the consumption of computing power through early prediction and judgment.

[0021] Furthermore, the segmentation module adopts an adaptive grid division method to dynamically adjust the size and shape of the small squares according to the picture content.

[0022] The beneficial effects of the basic scheme are: the adaptive grid division method can automatically adjust the density of the grid according to the changing characteristics of the physical quantities in the calculation area, perform adaptive encryption in areas with more tree shadows, and better capture the image information of the shadow area. In non-shadow areas, the adaptive grid division method will use a sparser grid for calculation, thereby reducing unnecessary calculations and improving calculation efficiency.

[0023] Furthermore, the analysis and judgment module is also used to perform analysis and judgment based on the vehicle's appearance and speed data monitored by the monitoring module. If the vehicle's speed data is always maintained below 40km / h and the vehicle's appearance is small, it is auxiliary judged as a small car. If the vehicle's speed data exceeds 40km / h and the vehicle's appearance is medium, it is auxiliary judged as a medium-sized car. If the vehicle's appearance is large, it is auxiliary judged as a large car.

[0024] The beneficial effect of the basic solution is that the analysis and judgment module combines the vehicle's speed data and appearance characteristics for comprehensive analysis, and can preliminarily and accurately distinguish the basic categories of small cars, medium-sized cars and large cars based on real-time monitoring. This instant judgment mechanism significantly improves the intelligence level of the traffic monitoring system, making the judgment of vehicle types more accurate and efficient, and providing a reliable basis for subsequent management and decision-making.

[0025] Furthermore, small vehicles include bicycles, electric vehicles and motorcycles; medium-sized vehicles include sedans and SUVs; and large vehicles include buses and trucks.

[0026] The beneficial effect of the basic solution is that through the fine classification of vehicle types by cameras, the traffic management department can more accurately grasp the distribution and driving characteristics of various types of vehicles on the road, which helps to optimize the allocation of traffic resources.

[0027] Furthermore, it also includes a comparison module, which is used to compare the shadow change information predicted by the learning prediction module with the actual shadow change information monitored by the monitoring module. The comparison module is constructed based on a convolutional neural network model, and is trained by inputting the comparison value of the shadow change information predicted by the learning prediction module and the actual shadow change information monitored by the monitoring module, and continuously adjusts to improve the prediction accuracy.

[0028] The beneficial effect of the basic solution is that through the comparison module, the system can obtain the difference between the predicted value and the actual value in real time, and this difference information is used as the input of the convolutional neural network model for training. As the training progresses, the model can gradually learn more accurate prediction patterns, thereby continuously improving the accuracy of the prediction.

[0029] Furthermore, it also includes a data storage module for storing historical image data, shadow changes and license plate recognition results, and transmitting the image data, shadow changes and license plate recognition results to the learning unit for further learning.

[0030] The beneficial effect of the basic solution is that the data storage module provides a rich data source for the learning unit. These historical data include images, shadow changes and license plate recognition results in various scenarios, which are an important basis for the learning unit to optimize the algorithm and train the model. Through continuous learning and iteration, the system can gradually improve the recognition accuracy and adaptability, and better cope with complex and changing traffic environments.

[0031] Furthermore, it also includes an intelligent alarm module, which is used to automatically send early warning information to relevant personnel or departments when traffic accidents and road congestion are monitored.

[0032] The beneficial effect of the basic solution is that the intelligent alarm module can monitor road conditions in real time and immediately trigger the alarm mechanism when a traffic accident or congestion is detected. This immediacy ensures that relevant departments and personnel can quickly receive early warning information, thereby quickly launching emergency plans, shortening response time, and improving the efficiency of emergency handling.

[0033] An artificial intelligence-based target detection optimization method comprises: step 1, learning the tree types, tree growth status, light intensity and light angle changes through a learning unit, and predicting the changes of tree shadows on the road through a prediction unit;

[0034] Step 2: Monitor the types of trees on the road, the growth status of the trees, and the changes in shadows through the monitoring module, and monitor the speed of beyond-visual-range vehicles through the monitoring module to predict the type of vehicle in advance, thereby reducing the amount of system calculations;

[0035] Step 3, comparing the shadow changes monitored by the monitoring module with the shadow changes predicted by the prediction unit through a comparison module, and then making a prediction correction;

[0036] Step 4: dynamically adjust the size and shape of the small squares of the image content captured by the recognition device through the segmentation module;

[0037] Step 5: determine the location of the small square where the tree shadow exists through the analysis and judgment module;

[0038] Step six, the control module is used to compensate for the small squares where tree shadows exist, thereby improving the recognition accuracy of the recognition device.

[0039] The beneficial effects of the basic scheme are: 1. Through the learning unit's learning of tree species, growth status, light intensity and angle changes, the prediction unit can more accurately predict the changes in tree shadows on the road. This helps to improve the accuracy of shadow prediction in complex and changing natural environments, such as different lighting conditions and seasonal changes, thereby providing a more reliable reference for subsequent target monitoring.

[0040] 2. The monitoring module not only monitors the changes in trees and their shadows, but also monitors the speed of vehicles beyond visual range, so that the type of vehicle can be predicted in advance. This approach reduces the amount of calculation required for the system to conduct a comprehensive analysis of the entire picture, improving processing speed and efficiency. At the same time, accurate vehicle type prediction helps traffic management and planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 Schematic diagram of a system of an artificial intelligence-based target detection system in an embodiment of the present invention.

[0042] Figure 2 Schematic diagram of a comparison module of an artificial intelligence-based target detection system in an embodiment of the present invention.

[0043] Figure 3 This is a method step diagram of an artificial intelligence-based target detection optimization method in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The following is further described in detail through specific implementation methods:

[0045] Embodiment 1:

[0046] Basically as attached Figure 1 -Attached Figure 2Shown: An artificial intelligence-based target detection system includes an identification device for identifying road vehicle information, a monitoring module and a segmentation module.

[0047] The recognition device includes a camera, and the camera is used to recognize vehicle information, tree information and tree shadow information on the road.

[0048] The monitoring module is used to monitor the position and change of shadows cast by trees on the actual road through an identification device; the monitoring module is also used to monitor the shape and speed of vehicles traveling on the road, including the shape and speed monitoring of beyond-visual-range vehicles.

[0049] The segmentation module is used to segment the entire picture monitored by the monitoring module through the identification device into a number of small squares; the segmentation module adopts an adaptive grid division method to dynamically adjust the size and shape of the small squares according to the content of the picture.

[0050] The invention is characterized in that it comprises a learning prediction module, an analysis and judgment module and a control module.

[0051] The learning prediction module is used to learn the types of trees and the growth status of the trees as well as the sunrise and sunset times of different years, months and days, and the learning prediction module is also used to predict the changes in shadows produced by trees on the road according to the learned sunrise and sunset times.

[0052] The learning prediction module includes a learning unit and a prediction unit. The learning unit is constructed based on the trained convolutional network model and is trained by inputting sunrise and sunset times. The convolutional neural network is used to output the light intensity and angle changes in different time periods. The prediction unit is used to predict the changes in shadows cast on the road by trees under the action of light based on the light intensity and angle changes in different time periods output by the learning unit and the tree information identified by the camera.

[0053] The analysis and judgment module is used to analyze the small squares, identify the small squares with shadows in the picture and transmit the position information of the small squares with shadows in the picture to the control module; the analysis and judgment module is also used to analyze and judge according to the appearance of the vehicle monitored by the monitoring module and the speed data of the vehicle. If the speed data of the vehicle is always maintained below 40km / h and the appearance of the vehicle is small, it is auxiliary judged as a small car. If the speed data of the vehicle exceeds 40km / h and the appearance of the vehicle is medium, it is auxiliary judged as a medium-sized car. If the appearance of the vehicle is large, it is auxiliary judged as a large car. Small cars include bicycles, electric cars and motorcycles; medium-sized cars include sedans and SUVs; large cars include buses and freight trucks.

[0054] The control module is used to perform compensation processing on the areas of the shadow small squares according to the received position information of the shadow small squares.

[0055] It also includes a comparison module, which is used to compare the shadow change information predicted by the learning prediction module with the actual shadow change information monitored by the monitoring module. The comparison module is constructed based on a convolutional neural network model, and is trained by inputting the comparison value of the shadow change information predicted by the learning prediction module and the actual shadow change information monitored by the monitoring module, and continuously adjusts to improve the prediction accuracy.

[0056] It also includes a data storage module for storing historical image data, shadow changes and license plate recognition results, and transferring the image data, shadow changes and license plate recognition results to the learning unit for further learning.

[0057] It also includes an intelligent alarm module, which is used to automatically send early warning information to relevant personnel or departments when traffic accidents and road congestion are detected.

[0058] The specific implementation process is as follows: First, the camera is set up at the location where road monitoring is required, ensuring that the camera can clearly cover the road area, including trees, roads and moving vehicles. Then, the monitoring module can continuously collect road images, including road conditions at different time periods (such as sunrise, sunset and various time periods during the day), ensuring that rich lighting changes and shadow information are included.

[0059] The monitoring module will first conduct preliminary monitoring of the beyond-visual-range vehicles that appear in the picture for the first time, monitoring the type and speed of the beyond-visual-range vehicles. Then the analysis and judgment module will make further judgments based on the speed information and the appearance of the vehicle. Because there are fewer pixels of beyond-visual-range vehicles in the picture, by comparing and identifying the type and speed of the beyond-visual-range vehicles, it is judged that the vehicle is a small car (bicycles, electric vehicles and motorcycles), a medium-sized car (sedan and SUV) or a large car (bus and freight truck). After the analysis and judgment module makes a preliminary prediction and judgment on the vehicle type, when the vehicle is identified by the recognition device, it can be identified based on the preliminary judgment of the analysis and judgment module, which is conducive to reducing the consumption of computing power of the recognition device.

[0060] The monitoring module uses the road conditions collected by the camera to train the convolutional neural network in the learning unit, so that it can accurately output the light intensity and angle changes in different time periods. Then the prediction unit can predict the changes in tree shadows on the road through the light intensity and angle changes in different time periods output by the learning unit.

[0061] After the prediction unit makes a prediction, the shadow change information predicted by the prediction unit can be compared with the shadow change information actually monitored by the monitoring module through the comparison module to observe the accuracy of the prediction, and then input into the convolutional neural network of the comparison module for training, continuously adjusting the network parameters to improve the prediction accuracy.

[0062] When the camera shoots the road, the segmentation module will use an adaptive grid division method according to the content of the picture captured by the camera, dynamically adjust the size and shape of the small squares, and use finer small squares to divide the picture within the range covered by the largest tree shadow in the picture. The small squares are used to improve the picture fineness and fit the curve at the junction of the shadow area and the non-shadow area in the picture. At this time, the analysis and judgment module can analyze the small squares one by one, identify the curve at the junction of the shadow area and the non-shadow area, and pass its position information to the control module. The control module compensates and adjusts the shadow area according to the received position information of the intersection curve to reduce the impact of the shadow on vehicle recognition.

[0063] The data storage module will record historical image data, shadow changes and license plate recognition results in real time, and regularly pass these data to the learning unit for further learning to continuously optimize model performance.

[0064] The intelligent alarm module will continuously monitor road conditions and use the data from the analysis and judgment module and the preset traffic accident and congestion identification algorithm to determine whether there are any abnormal conditions. Once the intelligent alarm module detects a traffic accident or road congestion, it will immediately generate a warning message and send it to the relevant departments for alarm so that the relevant departments can take timely response measures.

[0065] Embodiment 2:

[0066] Basically as attached Figure 3 As shown: an artificial intelligence-based target detection optimization method, comprising: step 1, learning the types of trees, the growth status of the trees, the light intensity and the change of the light angle through a learning unit, and predicting the change of the shadow of the trees on the road through a prediction unit;

[0067] Step 2: Monitor the types of trees on the road, the growth status of the trees, and the changes in shadows through the monitoring module, and monitor the speed of beyond-visual-range vehicles through the monitoring module to predict the type of vehicle in advance, thereby reducing the amount of system calculations;

[0068] Step 3, comparing the shadow changes monitored by the monitoring module with the shadow changes predicted by the prediction unit through a comparison module, and then making a prediction correction;

[0069] Step 4: dynamically adjust the size and shape of the small squares of the image content captured by the recognition device through the segmentation module;

[0070] Step 5: determine the location of the small square where the tree shadow exists through the analysis and judgment module;

[0071] Step six, the control module is used to compensate for the small squares where tree shadows exist, thereby improving the recognition accuracy of the recognition device.

[0072] The specific implementation process is as follows: The learning unit uses historical data, including tree species, growth status, and changes in light intensity and angle at different time points, to perform deep learning through convolutional neural networks, extract key features and build a prediction model. The prediction unit, based on the output of the learning unit and combined with real-time camera data, predicts the precise location and change trend of tree shadows on the road in the future time period.

[0073] The monitoring module uses a camera to capture real-time road images and identify the type of trees, their growth status, and the real-time location of their shadows. At the same time, it uses image processing technology and deep learning algorithms to monitor the speed of vehicles beyond visual range and preliminarily predict the type of vehicle, such as small, medium, or large, based on the speed and vehicle profile, thereby optimizing the allocation of computing resources in advance and reducing unnecessary computing burdens.

[0074] The comparison module receives data from the monitoring module and the prediction unit in real time, uses a convolutional neural network to compare the shadow change differences between the two, and quickly iterates and corrects the prediction results to ensure the accuracy and real-time performance of shadow prediction.

[0075] The segmentation module uses adaptive grid division technology to dynamically adjust the size and shape of small squares according to the real-time image content provided by the monitoring module to more accurately divide the shadow area and vehicle area.

[0076] The analysis and judgment module analyzes the small squares one by one, identifies the small squares containing tree shadows, and marks their location information.

[0077] The control module performs local compensation adjustment on the small squares where tree shadows exist based on the position information provided by the analysis and judgment module to enhance the brightness contrast of the shadow area, thereby improving the recognition accuracy and stability of the recognition device for targets such as vehicles and pedestrians.

[0078] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0079] The above is only an embodiment of the present invention. The common sense such as the known specific structure and characteristics in the scheme is not described in detail here. The ordinary technicians in the relevant field know all the common technical knowledge in the technical field of the invention before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply the conventional experimental means before that date. The ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the protection scope of the present invention, which will not affect the effect of the implementation of the present invention and the practicality of the patent. The protection scope required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. An artificial intelligence-based target detection system: comprising an identification device for identifying road vehicle information, a monitoring module and a segmentation module; The recognition device includes a camera, and the camera is used to recognize vehicle information, tree information, and tree shadow information on the road; A monitoring module, used for monitoring the position and change of shadows cast on the road by trees on the actual road through an identification device; A segmentation module is used to segment the entire picture monitored by the monitoring module through the recognition device into a number of small squares; It is characterized by: comprising a learning prediction module, an analysis and judgment module and a control module; A learning prediction module is used to learn the types of trees and the growth status of the trees as well as the sunrise and sunset times of different years, months and days, and the learning prediction module is also used to predict the changes of shadows produced by the trees on the road according to the learned sunrise and sunset times; An analysis and judgment module is used to analyze the small squares, identify the small squares with shadows in the picture and transmit the position information of the small squares with shadows in the picture to the control module; The control module is used to perform compensation processing on the areas of the shadow small squares according to the received position information of the shadow small squares.

2. The artificial intelligence based target detection system according to claim 1, characterized in that: The learning prediction module includes a learning unit and a prediction unit. The learning unit is constructed based on the trained convolutional network model and is trained by inputting sunrise and sunset times. The convolutional neural network is used to output the light intensity and angle changes in different time periods. The prediction unit is used to predict the changes in shadows cast on the road by trees under the action of light based on the light intensity and angle changes in different time periods output by the learning unit and the tree information identified by the camera.

3. The artificial intelligence based target detection system according to claim 2, characterized in that: The monitoring module is also used to monitor the appearance and speed of vehicles traveling on the road, including the appearance and speed monitoring of beyond-visual-range vehicles.

4. The artificial intelligence-based target detection system according to claim 3, characterized in that: The segmentation module uses an adaptive grid division method to dynamically adjust the size and shape of the small squares according to the picture content.

5. The artificial intelligence-based target detection system according to claim 4, characterized in that: The analysis and judgment module is also used to analyze and judge based on the vehicle's appearance and speed data monitored by the monitoring module. If the vehicle's speed data is always maintained below 40km / h and the vehicle's appearance is small, it is auxiliary judged as a small car. If the vehicle's speed data exceeds 40km / h and the vehicle's appearance is medium, it is auxiliary judged as a medium-sized car. If the vehicle's appearance is large, it is auxiliary judged as a large car.

6. The artificial intelligence-based target detection system according to claim 5, characterized in that: Small cars include bicycles, electric cars and motorcycles; medium-sized cars include sedans and SUVs; large cars include buses and trucks.

7. The artificial intelligence-based target detection system according to claim 6, characterized in that: It also includes a comparison module, which is used to compare the shadow change information predicted by the learning prediction module with the actual shadow change information monitored by the monitoring module. The comparison module is constructed based on a convolutional neural network model, and is trained by inputting the comparison value of the shadow change information predicted by the learning prediction module and the actual shadow change information monitored by the monitoring module, and continuously adjusts to improve the prediction accuracy.

8. The artificial intelligence-based target detection system according to claim 7, characterized in that: It also includes a data storage module for storing historical image data, shadow changes and license plate recognition results, and transferring the image data, shadow changes and license plate recognition results to the learning unit for further learning.

9. The artificial intelligence-based target detection system according to claim 8, characterized in that: It also includes an intelligent alarm module, which is used to automatically send early warning information to relevant personnel or departments when traffic accidents and road congestion are detected.

10. An artificial intelligence-based target detection optimization method, characterized in that: include: Step 1: learn the tree types, tree growth status, light intensity and light angle changes through the learning unit, and predict the changes of tree shadows on the road through the prediction unit; Step 2: Monitor the types of trees on the road, the growth status of the trees, and the changes in shadows through the monitoring module, and monitor the speed of beyond-visual-range vehicles through the monitoring module to predict the type of vehicle in advance, thereby reducing the amount of system calculations; Step 3, comparing the shadow changes monitored by the monitoring module with the shadow changes predicted by the prediction unit through a comparison module, and then making a prediction correction; Step 4: dynamically adjust the size and shape of the small squares of the image content captured by the recognition device through the segmentation module; Step 5: determine the location of the small square where the tree shadow exists through the analysis and judgment module; Step six, the control module is used to compensate for the small squares where tree shadows exist, thereby improving the recognition accuracy of the recognition device.

Images