Irregular target detection method based on neural network, storage medium and processor

A target detection and neural network technology, applied in storage media and processors, and in the field of neural network-based irregular target detection methods, can solve the problem of not supporting irregular shape target positioning, unable to realize positioning and shape detection, and unable to detect targets The location information and approximate shape of the

Pending Publication Date: 2021-05-11
西安光启智能技术有限公司
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The disadvantage of this kind of discrimination is: the position information and approximate shape of the target cannot be detected; at the same time, the existing technology target detection mainly focuses on the application of regular shape targets, such as face, human body and general object detection
Its main disadvantages are: it only realizes classification, but cannot realize positioning and shape detection; it only supports target detection with common shapes, and does not support irregular shape target positioning

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Irregular target detection method based on neural network, storage medium and processor
  • Irregular target detection method based on neural network, storage medium and processor
  • Irregular target detection method based on neural network, storage medium and processor

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0044] figure 1 It is a flow chart of sample generation in a neural network-based irregular target detection method of the present invention. The sample generation process in the neural network-based irregular target detection method of the present invention includes the steps of: acquiring original data; marking the original data; randomly generating samples. During specific implementation, the original data are pictures of various sizes.

[0045] The sample generation process is responsible for generating positive and negative samples for training. By marking the target, each target is represented by a box (lefttop, rightbottom). On this basis, positive samples are automatically randomly generated based on each box, and uniformly resized to the deep network input. size. Positive samples refer to samples that meet the requirements randomly generated by tools in the marked area based on the marked samples, and negative samples refer to samples outside the marked area. Train...

Embodiment 2

[0071] An embodiment of the present invention also provides a storage medium, the storage medium includes a stored program, wherein, when the above program is running, the procedure of the above method for detecting an irregular object based on a neural network is executed.

[0072] Optionally, in this embodiment, the above-mentioned storage medium may be configured to store program codes for executing the following procedure of the neural network-based irregular target detection method:

[0073] S1. Generate test samples for training;

[0074] S2. Perform model training on the test sample;

[0075] S3. Start the target detection on the trained model, perform detection on the input image of any size according to the zoom ratio, and then summarize the detection results and merge the overlapping frames to generate an irregular target area.

[0076] Optionally, in this embodiment, the above-mentioned storage medium may include but not limited to: U disk, read-only memory (Read-O...

Embodiment 3

[0079] The embodiment of the present invention also provides a processor, the processor is used to run a program, wherein, the program executes the steps in the above-mentioned neural network-based irregular target detection method when running.

[0080] Optionally, in this embodiment, the above program is used to perform the following steps:

[0081] S1. Generate test samples for training;

[0082] S2. Perform model training on the test sample;

[0083] S3. Start the target detection on the trained model, perform detection on the input image of any size according to the zoom ratio, and then summarize the detection results and merge the overlapping frames to generate an irregular target area.

[0084] Optionally, for specific examples in this embodiment, reference may be made to the above-mentioned embodiments and examples described in specific implementation, and details are not repeated in this embodiment.

[0085] It can be seen that by using the processor of the present ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides an irregular target detection method based on a neural network, and the method is characterized in that the method comprises the steps: S1, generating a test sample for training; S2, performing model training on the test sample; and S3, starting target detection for the trained model, detecting the input pictures of any size according to a preset zoom ratio, then summarizing detection results, and merging overlapped frames to generate an irregular target area. Urban management problems can be efficiently and automatically detected, the requirements for automatically identifying delicious food, fruits and vegetables in intelligent agriculture and life are met, and scene identification in visual understanding is achieved. The invention can be universally applied to urban management such as intelligent identification of garbage accumulation, bicycle accumulation, smoke fire detection and the like; food identification in the catering industry; fruits and vegetables are identified in life or smart agriculture; and scene identification and the like are used in all walks of life.

Description

【Technical field】 [0001] The invention relates to the technical field of target detection, in particular to a neural network-based irregular target detection method, a storage medium and a processor. 【Background technique】 [0002] For irregular targets, the common binary classification method is commonly used in the industry to distinguish them. The disadvantage of such discrimination is: the position information and approximate shape of the target cannot be detected; meanwhile, the prior art target detection mainly focuses on the application of regular shape targets, such as human face, human body and general object detection. Its main disadvantages are: it only realizes classification, but cannot realize positioning and shape detection; it only supports the detection of objects with common shapes, and does not support the positioning of irregular shapes. 【Content of invention】 [0003] The technical problem to be solved by the present invention is to provide a neural n...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/62G06N3/04
CPCG06V20/00G06V20/68G06V2201/07G06N3/045G06F18/214
Inventor 刘若鹏栾琳季春霖肖兴贵
Owner 西安光启智能技术有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap