CN117830857A - Atlas200 DK-based space target detection method and related equipment - Google Patents

Atlas200 DK-based space target detection method and related equipment Download PDF

Info

Publication number
CN117830857A
CN117830857A CN202410117710.6A CN202410117710A CN117830857A CN 117830857 A CN117830857 A CN 117830857A CN 202410117710 A CN202410117710 A CN 202410117710A CN 117830857 A CN117830857 A CN 117830857A
Authority
CN
China
Prior art keywords
target detection
model
atlas200
atlas200dk
yolo
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202410117710.6A
Other languages
Chinese (zh)
Inventor
刘曦
於志文
周美娟
贺鹏超
王宁
毛远宏
王天行
何海峰
王亮
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Northwestern Polytechnical University
Xian Microelectronics Technology Institute
Original Assignee
Northwestern Polytechnical University
Xian Microelectronics Technology Institute
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Northwestern Polytechnical University, Xian Microelectronics Technology Institute filed Critical Northwestern Polytechnical University
Priority to CN202410117710.6A priority Critical patent/CN117830857A/en
Publication of CN117830857A publication Critical patent/CN117830857A/en
Pending legal-status Critical Current

Links

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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • 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
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
    • G06V10/449—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters
    • G06V10/451—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters with interaction between the filter responses, e.g. cortical complex cells
    • G06V10/454—Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
    • 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
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/07—Target detection

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Multimedia (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computational Linguistics (AREA)
  • Databases & Information Systems (AREA)
  • Medical Informatics (AREA)
  • Biodiversity & Conservation Biology (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a space target detection method based on Atlas200DK and related equipment, the method is based on Atlas200DK, resnet18 is adopted as a backbone network to replace an original feature extraction network Darknet53 in a YOLO v3 network model, the combination greatly improves the detection rate of the model, meanwhile, the problem of suitability between an algorithm and a processor is considered, the target detection model is subjected to model file conversion processing to ensure that the optimized algorithm can be normally executed, finally, an obtained space target image is input into the target detection model supported by Atlas200DK, a target detection result is obtained through analysis processing, and the reasoning speed of the model can be ensured through the optimization and the suitability adjustment of the YOLO v3 network model even when facing a large amount of data, so that the processing effect of high speed and low power consumption is achieved; the method can improve the processing speed and the detection precision of the space target monitoring, and meets the requirement of the current spaceborne computer system on the high reliability of the space target monitoring.

Description

Atlas200 DK-based space target detection method and related equipment
Technical Field
The invention belongs to the field of spaceborne computers, and particularly relates to a space target detection method based on Atlas200DK and related equipment.
Background
The space target detection is an important part of the space target monitoring field, and different target detection algorithms are suitable for different application scenes, wherein the YOLO series target detection algorithm is widely applied due to the efficient detection performance.
With the development of the era, the space-borne computer system has raised higher requirements on the reliability of space target monitoring, and in general, the space image information is more, and the calculation amount is larger in the process of space target identification.
At present, by adopting computing devices such as a CPU and a GPU, the problems of low processing speed, high power consumption and the like in the aspect of data processing are caused by adopting a traditional YOLO series algorithm, and the requirement of a space target monitoring by a space-borne computer system on high reliability cannot be met obviously.
Disclosure of Invention
In order to overcome the defects of the technology, the invention provides a spatial target detection method based on Atlas200DK and related equipment, which can solve the technical problems of low processing speed and high power consumption of the existing CPU, GPU and other computing equipment under the condition of large data processing capacity.
In order to achieve the above purpose, the invention adopts the following technical scheme:
a space target detection method based on Atlas200DK is applied to Atlas200DK, and comprises the following steps:
acquiring a space target image on Atlas200 DK;
inputting a space target image into a target detection model supported by Atlas200DK, and obtaining a target detection result through analysis;
the target detection model supported by Atlas200DK is obtained by converting a model file from the target detection model; the target detection model is obtained by training and testing a model structure constructed by YOLO v3 and Resnet18.
Further, the constructing step of the target detection model includes:
and replacing the feature extraction network in the YOLO v3 network model with a Resnet18 network to obtain a YOLO-R network structure.
Further, the specific steps of performing model file conversion processing on the target detection model include:
and converting the target detection model into an om model file, and outputting the target detection model supported by Atlas200 DK.
Further, the target detection model adopts a prototxt model file under a Caffe framework.
Further, in the process of converting the model file, the customized network modification and AIPP configuration are carried out on the prototxt model file under the Caffe framework.
Further, an auxiliary development tool ATC in a software stack of the lifting AI processor is adopted to convert the target detection model into an om model file, wherein the model conversion is completed by executing an ATC command, and a Mind Studio is adopted to correct the conversion result.
Further, the obtained space target image is input into a target detection model supported by Atlas200DK to obtain a reasoning detection result, the reasoning detection result is analyzed, and bounding box screening processing is carried out by combining IoU with non-maximum suppression to obtain the target detection result.
A space target detection system based on Atlas200DK is used for realizing the steps of the space target detection method based on Atlas200DK, and comprises the following steps:
the image acquisition module is used for acquiring a space target image on Atlas200 DK;
the target detection module is used for inputting the acquired space target image into a target detection model supported by Atlas200DK, and obtaining a target detection result through analysis;
the target detection model supported by Atlas200DK is obtained by converting a model file from the target detection model; the target detection model is obtained by training and testing a model structure constructed by YOLO v3 and Resnet18.
An apparatus, comprising:
a memory for storing a computer program;
and the processor is used for realizing the steps of the spatial target detection method based on Atlas200DK when executing the computer program.
A computer readable storage medium storing a computer program for implementing the steps of the Atlas200 DK-based spatial target detection method described above when executed by a processor.
Compared with the prior art, the invention has the following beneficial effects:
the invention also provides a space target detection method based on Atlas200DK, which is based on Atlas200DK, adopts Resnet18 as backbone network to replace original feature extraction network Darknet53 in YOLO v3 network model, and the combination greatly improves the detection rate of the model, and simultaneously considers the problem of suitability between algorithm and processor, through carrying out model file conversion processing on target detection model, ensuring that the optimized algorithm can be normally executed, finally inputting the obtained space target image into target detection model supported by Atlas200DK, obtaining target detection result through analysis processing, and through optimizing YOLO v3 network model and adjusting the suitability, even if a large amount of data is faced, the reasoning speed of the model can be ensured, and the processing effect of high speed and low power consumption can be achieved; the method can improve the processing speed and the detection precision of the space target monitoring, and meets the requirement of the current spaceborne computer system on the high reliability of the space target monitoring.
Drawings
Fig. 1 is a frame diagram of a spatial target detection method flow based on Atlas200DK according to an embodiment of the present invention;
FIG. 2 is a diagram of a YOLO-R network structure according to an embodiment of the present invention;
fig. 3 is a flowchart of a spatial target detection method based on Atlas200DK provided by the invention;
fig. 4 is a schematic structural diagram of a spatial target detection system based on Atlas200 DK.
Detailed Description
The invention provides a space target detection method based on Atlas200DK, as shown in figure 3, comprising the following steps:
s1: a spatial target image on Atlas200DK was acquired.
S2: and inputting the space target image into a target detection model supported by Atlas200DK, and obtaining a target detection result through analysis processing.
The target detection model supported by Atlas200DK is obtained by converting a model file from the target detection model; the target detection model is obtained by training and testing a model structure constructed by YOLO v3 and Resnet18.
The construction step of the target detection model comprises the following steps:
and replacing the feature extraction network in the YOLO v3 network model with a Resnet18 network to obtain a YOLO-R network structure.
The specific steps of performing model file conversion processing on the target detection model include:
and converting the target detection model into an om model file, and outputting the target detection model supported by Atlas200 DK.
Here, the object detection model employs a prototxt model file under the Caffe framework.
In the process of converting the model file, the customized network modification and AIPP configuration are carried out on the prototxt model file under the Caffe framework.
Specifically, an auxiliary development tool ATC in a software stack of a lifting AI processor is adopted to convert the target detection model into an om model file, wherein the model conversion is completed by executing an ATC command, and a Mind Studio is adopted to correct the conversion result.
Specifically, the obtained space target image is input into a target detection model supported by Atlas200DK to obtain a reasoning detection result, the reasoning detection result is analyzed, and bounding box screening processing is performed by combining IoU with non-maximum suppression to obtain the target detection result.
As shown in fig. 4, the present invention further provides a spatial target detection system based on Atlas200DK, including: the image acquisition module is used for acquiring a space target image on Atlas200 DK; the target detection module is used for inputting the acquired space target image into a target detection model supported by Atlas200DK, and obtaining a target detection result through analysis; the target detection model supported by Atlas200DK is obtained by converting a model file from the target detection model; the target detection model is obtained by training and testing a model structure constructed by YOLO v3 and Resnet18.
The invention also provides an apparatus comprising: a memory for storing a computer program; and the processor is used for realizing the spatial target detection method based on Atlas200DK when executing the computer program.
The processor, when executing the computer program, implements the steps of Atlas200 DK-based spatial target detection described above, for example: acquiring a space target image on Atlas200 DK; inputting the acquired space target image into a target detection model supported by Atlas200DK, and obtaining a target detection result through analysis; the target detection model supported by Atlas200DK is obtained by converting a model file from the target detection model; the target detection model is obtained by training and testing a model structure constructed by YOLO v3 and Resnet18.
Alternatively, the processor may implement functions of each module in the above system when executing the computer program, for example: the image acquisition module is used for acquiring a space target image on Atlas200 DK; the target detection module is used for inputting the acquired space target image into a target detection model supported by Atlas200DK, and obtaining a target detection result through analysis; the target detection model supported by Atlas200DK is obtained by converting a model file from the target detection model; the target detection model is obtained by training and testing a model structure constructed by YOLO v3 and Resnet18.
The computer program may be divided into one or more modules/units, which are stored in the memory and executed by the processor to accomplish the present invention, for example. The one or more modules/units may be a series of computer program instruction segments capable of performing a predetermined function, the instruction segments describing the execution of the computer program in the Atlas200 DK-based spatial target detection apparatus. For example, the computer program may be split into an image acquisition module and a target detection module; the specific functions of each module are as follows:
the image acquisition module is used for acquiring a space target image on Atlas200 DK; the target detection module is used for inputting the acquired space target image into a target detection model supported by Atlas200DK, and obtaining a target detection result through analysis; the target detection model supported by Atlas200DK is obtained by converting a model file from the target detection model; the target detection model is obtained by training and testing a model structure constructed by YOLO v3 and Resnet18.
The Atlas200 DK-based space target detection device can be a desktop computer, a notebook computer, a palm computer, a cloud server and other computing devices. The Atlas200 DK-based spatial target detection device may include, but is not limited to, a processor, memory. It will be appreciated by those skilled in the art that the above is an example of an Atlas200 DK-based spatial target detection apparatus, and is not limited to an Atlas200 DK-based spatial target detection apparatus, and may include more components than those described above, or may combine some components, or different components, e.g., the Atlas200 DK-based spatial target detection apparatus may further include an input-output apparatus, a network access apparatus, a bus, etc.
The processor may be a central processing unit (Central Processing Unit, CPU), other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. The general purpose processor may be a microprocessor or the processor may also be any conventional processor, etc., which is a control center for the Atlas200 DK-based spatial target detection, and connects various parts of the entire Atlas200 DK-based spatial target detection apparatus using various interfaces and lines.
The memory may be used to store the computer program and/or module, and the processor may implement various functions of the Atlas200 DK-based spatial target detection apparatus by running or executing the computer program and/or module stored in the memory, and invoking data stored in the memory.
The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program (such as a sound playing function, an image playing function, etc.) required for at least one function, and the like; the storage data area may store data (such as audio data, phonebook, etc.) created according to the use of the handset, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, memory, plug-in hard disk, smart Media Card (SMC), secure Digital (SD) Card, flash Card (Flash Card), at least one disk storage device, flash memory device, or other volatile solid-state storage device.
The invention also provides a computer readable storage medium storing a computer program which when executed by a processor implements the steps of the Atlas200 DK-based spatial target detection method.
The modules/units integrated with the Atlas200 DK-based spatial target detection system may be stored in a computer readable storage medium if implemented in the form of software functional units and sold or used as a stand alone product.
Based on such understanding, the implementation of all or part of the above-mentioned Atlas200 DK-based spatial target detection method according to the present invention may also be accomplished by instructing related hardware by a computer program, where the computer program may be stored in a computer-readable storage medium, and the computer program may implement the steps of the above-mentioned Atlas200 DK-based spatial target detection method when executed by a processor. The computer program comprises computer program code, and the computer program code can be in a source code form, an object code form, an executable file or a preset intermediate form and the like.
The computer readable storage medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), an electrical carrier signal, a telecommunications signal, a software distribution medium, and so forth.
It should be noted that the computer readable storage medium may include content that is subject to appropriate increases and decreases as required by jurisdictions and by jurisdictions in which such computer readable storage medium does not include electrical carrier signals and telecommunications signals.
The invention is further illustrated by the following examples in conjunction with the accompanying drawings:
examples
As mentioned in the background: in the existing spaceborne computer system, computing devices such as a CPU and a GPU are adopted, and the problems of low processing speed, high power consumption and the like in the aspect of data processing are caused by adopting a traditional YOLO series algorithm, so that the requirement of the spaceborne computer system on high reliability of space target monitoring can not be met obviously.
In order to solve the above problems, the present embodiment provides a spatial target detection method based on Atlas200DK, which mainly uses an improved YOLO v3 algorithm to perform target detection, where the improved YOLO v3 algorithm has efficient detection performance, and the Atlas200DK can further improve the reasoning speed of the algorithm, so as to complete efficient detection tasks.
The main steps of the detection method are as follows: the method comprises a model training stage and a model reasoning stage; as shown in FIG. 1, the model training stage is mainly implemented at a server side to obtain a target detection model, the invention is mainly aimed at optimizing a lifting AI processor, the feature extraction network of YOLO v3 is replaced by Resnet18 to finish optimization, and the model reasoning stage is implemented on Atlas200 DK. Based on this, the main idea of the invention is:
(1) Algorithm optimization
Atlas200DK is used as a hardware product in the form of a developer suite, the rising 310 chip is used as the core of the acceleration module, and the rising 310 chip improves the algorithm performance through strong computing power. Therefore, how to fully utilize the utilization rate of the new chip computing unit of the rising AI series is considered when the algorithm is designed, and the algorithm optimization is realized mainly by modifying the network structure in the algorithm optimization process.
The YOLO v 3-based target detection algorithm used in the present embodiment is characterized in that the extraction network is a dark net53, and the traditional YOLO v3 has a relatively slow computing speed in the face of a large amount of computing data, and cannot meet the requirement of the current on-board computer system on high reliability of space target monitoring, so in order to improve the detection rate of the model, in the present embodiment, the dark net53 of the YOLO v3 is replaced with a backbone network with a shallower depth to complete feature extraction. After the research on each case of application development using Atlas200DK in the rising community is completed, a more typical classical case is to use Resnet18 to complete target recognition and detection tasks, and good performance of Resnet18 can be seen through each case, which is one of the invention points of the present invention. In the practical implementation process of Atlas200DK, the suitability of the algorithm to the rising AI processor needs to be considered first, and if the situation that the algorithm cannot be adapted occurs, the algorithm will bring great resistance to the subsequent implementation of the application. Meanwhile, the Darknet53 network is used as an important variant network of Resnet, so that the YOLO v3 target detection algorithm can be guaranteed to be normally executed after being optimized, and the Darknet53 is considered to be gradually replaced by a Resnet series network structure similar to the Resnet series network structure, namely Resnet18.
As shown in fig. 2, YOLO v3 was optimized by the above-described optimization method to obtain a YOLO-R structure, and the flow was similar to that of original YOLO v3 when the structure was used for target detection of the algorithm. Firstly, feature extraction is completed by using a Resnet18 network, three features with different scales are respectively extracted, and then, the output of target results with different scales is completed through convolution network, up-sampling and other processes. Next, a target frame prediction is performed, and a useful frame is extracted from the feature, where the target frame prediction is the same as the original YOLO v3 network. And finally, screening out a final result frame according to a non-maximum suppression algorithm.
(2) Integral configuration of experiment platform
In the model reasoning stage in Atlas200DK, the original model file needs to be converted into an offline network model om model file supported by the lifting AI processor. In the lifting AI processor, model files under the deep learning framework of the current mainstream are supported, such as Caffe, mindSpore, tensorFlow. The embodiment uses a model file prototxt file and a weight file caffeodel under the Caffe framework to successfully transform the model through a model transformation tool.
After model conversion is completed in the model reasoning stage, application development is carried out according to the development flow of Atlas200DK and by combining with ACL development API. The general application development flow is as follows: preprocessing media data, loading a model for model reasoning, analyzing a model reasoning result, and finally processing the obtained reasoning result. After the application development is completed, model deployment and model operation can be performed according to development languages of different applications. In the Atlas200DK developer suite of the current version, APIs are provided that are divided into a c++ version and a python version, so the application development language is mainly c++ language or python language. In the embodiment, the C++ language is used for model deployment and model operation, and the compiling work of engineering files is finished in a development environment, and then the deployment and operation on a developer board are carried out.
(3) Algorithm and application development implementation
In Atlas200DK, the existing model must be converted into an offline model om file supported by the lifting AI processor, so that it can be used for subsequent application development, and for implementing and completing the industry. When the model conversion is carried out, an auxiliary development tool ATC in a software stack of a lifting AI processor is mainly used, and two different model conversion methods are provided for a user, namely a model conversion tool of Mindstudio and ATC command line model conversion. When the model is converted, the customized network modification and AIPP configuration are required according to the requirement. The customized network modification is mainly aimed at modifying a model file (prototxt) under a Caffe framework, so that the model file meets the operator support during the model conversion of a lifting AI processor, the modification of the file content is required to be completed before the model conversion by both model conversion operations, and the AIPP configuration provides different configuration methods for two different conversion modes. The embodiment adopts ATC as an auxiliary development tool for model conversion. When the model conversion is carried out through the ATC command line, the model file, the weight file and the AIPP configuration file are placed under the same folder, the terminal is opened under the folder to carry out environment configuration, and the model conversion is completed through the ATC command line. The mindtudio is used to see if the structure of the om model is consistent with the original caffe model after model conversion using ATC. The above is still another aspect of the present invention.
According to each function analysis in the application scene analysis, a sub-function is respectively established for each function, and the functions of each scene are realized.
1) Initializing and de-initializing resources: when the ACL is used for program development, various resources are initialized, the ACL initialization is realized through an acl.init interface, the application of data preprocessing resources is completed by using an interface under acl.media, the application of resources is loaded by using an interface model under acl.mdl, the problem of subsequent development operation is avoided, and when a certain resource is used, the initialization is needed to be carried out, so that the release of the resource is completed.
2) Model reasoning function: the interface under acl.mdl is mainly called. After the preprocessing operation of the picture is completed, the image can be sent into the model to carry out model loading and execution, and the reasoning task of the model is completed. After model reasoning is finished, model resource release is needed.
3) Data post-processing function: after the model reasoning operation is finished, the picture reasoning result is obtained, and the model reasoning result is analyzed to obtain a final target detection result. Different original models can make the post-processing function different, for example, the YOLOv3 model encapsulates parameters such as confidence level, non-maximum suppression and the like during model conversion, so that the post-processing is not needed to be processed for the part, and only an inference result is needed to be obtained and output. The YOLO-R model needs to analyze the reasoning result in detail, analyze the model reasoning result in the post-processing function, and complete the operations related to the bounding box screening by combining IoU (cross-correlation measurement) with non-maximum suppression and the like, so as to obtain the final target detection result. After the reasoning result is obtained, a reasoning detection result is output, and the reasoning result is marked on the original picture by using opencv or picllow functions.
In summary, the invention provides a spatial target detection method based on Atlas200DK, the method is based on Atlas200DK, resnet18 is adopted as a backbone network to replace an original feature extraction network Darknet53 in a YOLO v3 network model, the combination greatly improves the detection rate of the model, meanwhile, the problem of suitability between an algorithm and a processor is considered, the trained target detection model is subjected to model file conversion processing, the optimized algorithm can be normally executed, the obtained spatial target image is finally input into the target detection model supported by Atlas200DK, a target detection result is obtained through analysis processing, and the reasoning speed of the model can be ensured through optimization and suitability adjustment of the YOLO v3 network model even when facing a large amount of data, and the processing effect of high speed and low power consumption is achieved; the method can improve the processing speed and the detection precision of the space target monitoring, and meets the requirement of the current spaceborne computer system on the high reliability of the space target monitoring.
The above embodiment is only one of the implementation manners capable of implementing the technical solution of the present invention, and the scope of the claimed invention is not limited to the embodiment, but also includes any changes, substitutions and other implementation manners easily recognized by those skilled in the art within the technical scope of the present invention.

Claims (10)

1. The spatial target detection method based on Atlas200DK is characterized by being applied to Atlas200DK and comprising the following steps:
acquiring a space target image on Atlas200 DK;
inputting a space target image into a target detection model supported by Atlas200DK, and obtaining a target detection result through analysis;
the target detection model supported by Atlas200DK is obtained by converting a model file from the target detection model; the target detection model is obtained by training and testing a model structure constructed by YOLO v3 and Resnet18.
2. The Atlas200 DK-based spatial target detection method according to claim 1, wherein the step of constructing the target detection model includes:
and replacing the feature extraction network in the YOLO v3 network model with a Resnet18 network to obtain a YOLO-R network structure.
3. The Atlas200 DK-based spatial target detection method according to claim 1, wherein the specific step of performing model file conversion processing on the target detection model comprises:
and converting the target detection model into an om model file, and outputting the target detection model supported by Atlas200 DK.
4. The spatial target detection method based on Atlas200DK according to claim 3, wherein the target detection model adopts a prototxt model file under a Caffe framework.
5. The Atlas200 DK-based spatial target detection method of claim 4, wherein the custom network modification and AIPP configuration are performed on the prototxt model file under the Caffe framework at the same time during the model file conversion process.
6. The Atlas200 DK-based spatial object detection method of claim 3, wherein the object detection model is converted into an om model file by using an auxiliary development tool ATC in a software stack of a lifting AI processor, wherein the model conversion is completed by executing an ATC command, and the conversion result is collated by using a mini Studio.
7. The Atlas200 DK-based spatial target detection method of claim 1, wherein the acquired spatial target image is input to a target detection model supported by Atlas200DK to obtain an inference detection result, and the inference detection result is analyzed and combined with IoU and non-maximum suppression to perform bounding box screening processing to obtain the target detection result.
8. Atlas200 DK-based spatial target detection system for implementing the steps of the Atlas200 DK-based spatial target detection method of any one of claims 1-7, comprising:
the image acquisition module is used for acquiring a space target image on Atlas200 DK;
the target detection module is used for inputting the acquired space target image into a target detection model supported by Atlas200DK, and obtaining a target detection result through analysis;
the target detection model supported by Atlas200DK is obtained by converting a model file from the target detection model; the target detection model is obtained by training and testing a model structure constructed by YOLO v3 and Resnet18.
9. An apparatus, comprising:
a memory for storing a computer program;
a processor for implementing the Atlas200 DK-based spatial target detection method of any one of claims 1-7 when executing said computer program.
10. A computer readable storage medium storing a computer program, characterized in that the computer program is executed by a processor for implementing the steps of the Atlas200 DK-based spatial target detection method of any of claims 1-7.
CN202410117710.6A 2024-01-26 2024-01-26 Atlas200 DK-based space target detection method and related equipment Pending CN117830857A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410117710.6A CN117830857A (en) 2024-01-26 2024-01-26 Atlas200 DK-based space target detection method and related equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410117710.6A CN117830857A (en) 2024-01-26 2024-01-26 Atlas200 DK-based space target detection method and related equipment

Publications (1)

Publication Number Publication Date
CN117830857A true CN117830857A (en) 2024-04-05

Family

ID=90524211

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410117710.6A Pending CN117830857A (en) 2024-01-26 2024-01-26 Atlas200 DK-based space target detection method and related equipment

Country Status (1)

Country Link
CN (1) CN117830857A (en)

Similar Documents

Publication Publication Date Title
CN109542399B (en) Software development method and device, terminal equipment and computer readable storage medium
CN108830235B (en) Method and apparatus for generating information
CN111062854B (en) Method, device, terminal and storage medium for detecting watermark
CN111899150A (en) Data processing method and device, electronic equipment and storage medium
CN112116567B (en) No-reference image quality evaluation method, device and storage medium
CN113505848A (en) Model training method and device
US20230005171A1 (en) Visual positioning method, related apparatus and computer program product
CN111444986A (en) Building drawing component classification method and device, electronic equipment and storage medium
CN111028279A (en) Point cloud data processing method and device, electronic equipment and storage medium
JP2023526899A (en) Methods, devices, media and program products for generating image inpainting models
CN113780326A (en) Image processing method and device, storage medium and electronic equipment
CN109697083B (en) Fixed-point acceleration method and device for data, electronic equipment and storage medium
CN111815748B (en) Animation processing method and device, storage medium and electronic equipment
CN113327194A (en) Image style migration method, device, equipment and storage medium
CN113033337A (en) TensorRT-based pedestrian re-identification method and device
CN117315758A (en) Facial expression detection method and device, electronic equipment and storage medium
CN117830857A (en) Atlas200 DK-based space target detection method and related equipment
CN117496990A (en) Speech denoising method, device, computer equipment and storage medium
CN116402914A (en) Method, device and product for determining stylized image generation model
CN111144374A (en) Facial expression recognition method and device, storage medium and electronic equipment
CN113139617B (en) Power transmission line autonomous positioning method and device and terminal equipment
CN112766276B (en) Channel adjustment method, device and equipment of convolutional neural network model
CN112801960A (en) Image processing method and device, storage medium and electronic equipment
CN113490954A (en) Neural network, operation method, and program
CN111401560A (en) Inference task processing method, device and storage medium

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