Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Target detection method in mining truck environment based on improved tiny-yolov3

A target detection and truck technology, which is applied in the direction of instruments, biological neural network models, character and pattern recognition, etc., can solve the problems of reducing the depth of the convolutional layer, increasing the running speed, and unsatisfactory real-time performance, so as to improve the accuracy of network detection, The effect of loss reduction

Pending Publication Date: 2019-09-06
NORTHEASTERN UNIV
View PDF1 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although the detection accuracy of yolov3 and SSD is very high, their real-time performance on embedded devices and PCs with poor performance is not ideal. tiny-yolov3 is a simplified model of YOLOV3, which reduces the depth of the convolutional layer
Although the detection accuracy has decreased, the running speed has been greatly improved

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
  • Target detection method in mining truck environment based on improved tiny-yolov3
  • Target detection method in mining truck environment based on improved tiny-yolov3
  • Target detection method in mining truck environment based on improved tiny-yolov3

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0046] like figure 1 Shown: A target detection method based on the improved tiny-yolov3 mining truck environment, including the following steps:

[0047] S1. Acquiring target image data.

[0048] The target object image data described here are pictures of people and vehicles taken by the camera on the truck from a bird's-eye view.

[0049] S2. Preprocessing the acquired image data of the target object.

[0050] It should be noted that the preprocessing described here at least includes data enhancement on the acquired image data of the target object and calculation of the size of anchors suitable for the dataset of the tiny-yolov3 model in step S3 by using the k-means clustering algorithm.

[0051] S3. Input the preprocessed object image data into the tiny-yolov3 model, and obtain the pixel position coordinates of the object in the image through the processing of the tiny-yolov3 model.

[0052] Wherein, the tiny-yolov3 model is an improved model combined with a residual netw...

Embodiment 2

[0076] The data set used in this embodiment consists of: extracting two types of people and cars in VOC2007 and VOC2012, and randomly selecting 50% of each as a part of the data set; pictures of people and cars taken by a camera at a bird’s-eye view; Pictures of mining trucks, people and vehicles taken at the scene. Among them, it is necessary to do data enhancement on the pictures containing mining trucks to expand the number of training for such targets as mining trucks.

[0077] The hardware configuration used in the experiment in this embodiment is a computer with Intel Core i5-7500 3.40GHz processor, NVIDIA GTX 1050Ti graphics card, 16GB RAM, 500GB Western Digital mechanical hard disk, the system is Windows10, 64-bit system, the programming language is Python, GPU acceleration Adopt CUDA9.0 and CUDNN7.0 to realize.

[0078] The training method used in this embodiment is: randomly select 90% of the pictures in the above data set as the training set, and the remaining 10% ...

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 relates to a target detection method in a mining truck environment based on improved tiny-yolov3. The method comprises the following steps: S1, acquiring image data of a target object; S2, preprocessing the acquired image data of the target object; S3, inputting the preprocessed image data of the target object into the tiny-yolov3 model, processing through the tiny-yogov3 model to obtain pixel position coordinates of the target object in the image; wherein the tiny-yolov3 model is a model improved by combining a residual network structure, and the target object image data is thetarget object image data obtained under the overlook condition. According to the detection method provided by the invention, the detection precision of the target can be greatly improved on the premise that the operation speed is not reduced.

Description

technical field [0001] The invention belongs to the technical field of computer vision detection, and in particular relates to an improved tiny-yolov3-based target detection method in a mining truck environment. Background technique [0002] In most open-pit mines, due to labor shortage and high labor costs, on the one hand, larger mining trucks are sought to reduce the need for personnel, and on the other hand, better solutions are also sought. With the continuous advancement of technology, unmanned mining trucks have emerged as the times require. Unmanned mining trucks save labor costs while improving safety and production efficiency, becoming an integral part of mine digitization. In the unmanned driving of mining trucks, active collision avoidance technology is a very important part. Due to the characteristics of height, width, and large size, mining trucks have large blind spots and long braking distances, and truck drivers will inevitably have traffic accidents due t...

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
CPCG06V20/58G06N3/045G06F18/23213
Inventor 肖冬单丰李泽黎霸俊李雪娆刘爔文
Owner NORTHEASTERN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com