阅读说明：本技术 一种用于带式输送机的非金属异物检测系统 (Non-metal foreign matter detection system for belt conveyor ) 是由 林贺 于 2020-08-06 设计创作，主要内容包括：本发明公开了一种用于带式输送机的非金属异物检测系统,主要涉及带式输送机智能检测相关技术领域；包括：图像采集仓；工业相机；工业相机光源；相机防尘罩；雨刷；喷水装置；高性能图像处理服务器,所述高性能图像处理服务器通过网口连接工业相机,用于读取工业相机采集到的图像,部署基于Darknet目标检测框架的异物检测服务；客户端,所述客户端用于检测状态和检测结果展示；控制板,所述客户端、高性能图像处理服务器、喷水装置均与控制板信号连接；本发明采用了基于计算机视觉的深度学习目标检测方法,从方法上解决了识别准确率和识别效率的问题,从配套硬件上解决了光源和镜头清洁问题。(The invention discloses a non-metallic foreign body detection system for a belt conveyor, which mainly relates to the technical field related to intelligent detection of the belt conveyor; the method comprises the following steps: an image acquisition bin; an industrial camera; an industrial camera light source; a camera dust cover; a wiper; a water spraying device; the high-performance image processing server is connected with the industrial camera through a network port and used for reading images acquired by the industrial camera and deploying foreign matter detection service based on a Darknet target detection framework; the client is used for displaying the detection state and the detection result; the client, the high-performance image processing server and the water spraying device are in signal connection with the control panel; the invention adopts a deep learning target detection method based on computer vision, thereby solving the problems of recognition accuracy and recognition efficiency and the problems of light source and lens cleaning from matched hardware.)

1. A non-metallic foreign object detection system for a belt conveyor, comprising:

the image collecting bin (1) is provided with a channel (12) for a belt conveyor (11) to pass through, and the channel (12) is not in contact with the belt conveyor (11);

an industrial camera (2), the industrial camera (2) being located on the upper side of the belt conveyor (11), the industrial camera (2) being mounted in the center of the top inside the image acquisition bin (1);

the industrial camera light sources (3), the industrial camera light sources (3) are arranged on the left side and the right side of the top in the image acquisition bin (1);

the camera dustproof cover (4) is sleeved on the industrial camera (2), and a transparent cover (41) is arranged at the front end of the camera dustproof cover (4);

the windshield wiper comprises a windshield wiper (5), wherein the windshield wiper (5) is controlled by a steering engine (51), and the windshield wiper (5) is used for wiping the front end face of a transparent cover (41);

a water spraying device (6), wherein the water spraying device (6) is used for spraying water to the front end surface of the transparent cover (41);

the high-performance image processing server (7) is connected with the industrial camera (2) through a network port, and is used for reading images acquired by the industrial camera (2) and deploying foreign body detection service based on a Darknet target detection framework;

the client (8), the client (8) is used for detecting the state and displaying the detection result;

the client (8), the high-performance image processing server (7) and the water spraying device (6) are in signal connection with the control board (8).

2. A non-metallic foreign matter detection system for a belt conveyor according to claim 1, characterized in that: the water spraying device (6) comprises a water pump (61), a water tank (62) and a spray head (63), wherein the spray head (63) points to the front end face of the transparent cover (41).

3. A non-metallic foreign matter detection system for a belt conveyor according to claim 1, characterized in that: the foreign matter detection service based on the Darknet target detection framework adopts a YOLOv4 network model, and the foreign matter detection service migrates and learns on the basis of a pre-training model, trains a sample set sampled in a material loading operation scene of an actual belt conveyor (11), and obtains a foreign matter detection network model.

4. A non-metallic foreign matter detection system for a belt conveyor according to claim 3, characterized in that: the sample set is a sample set which is made in a YOLO support format by simulating various postures of field foreign matters in coal materials in a laboratory environment, traversing images shot in multiple directions and distances, and then manually calibrating detection frames and categories.

5. A non-metallic foreign matter detection system for a belt conveyor according to claim 3, characterized in that: the foreign body detection network model is applied and the sample set is expanded and fine-tuned, so that the model with high accuracy is obtained step by step.

6. A non-metallic foreign matter detection system for a belt conveyor according to claim 1, characterized in that: the detection state refers to the connection state of the industrial camera (2) and the connection state of the control panel (9), the high-performance image processing server (7) can directly read the state of the industrial camera (2) through the driving of the industrial camera (2), the working state of the control panel (9) is monitored by adopting heartbeat communication with the high-performance image processing server (7), and the state of the control panel (9) and the state of the water spraying device (6) connected with the control panel (9) are judged according to the communication state and the communication message.

7. A non-metallic foreign matter detection system for a belt conveyor according to claim 1, characterized in that: the industrial camera light source (3) comprises a strip led light source and a light homogenizing plate.

8. A non-metallic foreign matter detection system for a belt conveyor according to claim 1, characterized in that: the included angle between the industrial camera (2) and the horizontal plane is 60 degrees.

Technical Field

The invention mainly relates to the technical field related to intelligent detection of belt conveyors, in particular to a non-metallic foreign matter detection system for a belt conveyor.

Background

As is well known, foreign matters on a conveying belt of a belt conveyor are main factors causing damage and even tearing of the belt, particularly, the belt under a coal mine is expensive to manufacture, and if the belt is longitudinally torn, great economic loss is generated, and even the personal safety of field workers is damaged.

At present, most of the existing anti-tearing protection systems are controlled by detecting faults after the faults occur, and although part of loss can be reduced, the problems are not fundamentally solved. The existing foreign matter detection means from the source comprises manual detection, metal detectors, radar detection, visual identification and the like.

Wherein, the manual detection efficiency is low, and the personnel safety risk is increased; the metal detector method is a common method, has good effect on removing metal foreign matters, but has no effect on non-metal foreign matters; the radar detection cost is high and the maintenance is difficult; the visual identification overcomes the defects of the former methods, but a high-definition camera is needed, so that the difficult problems of harsh environments in coal mines, light rays and lens cleaning are solved, and various complex environments cause poor anti-interference capability, low identification accuracy or low identification efficiency of a visual identification algorithm.

Disclosure of Invention

The invention aims to solve the problems in the prior art, and provides a non-metallic foreign matter detection system for a belt conveyor.

In order to achieve the purpose, the invention is realized by the following technical scheme:

a non-metallic foreign object detection system for a belt conveyor, comprising:

the image acquisition bin is provided with a channel for the belt conveyor to pass through, and the channel is not in contact with the belt conveyor;

an industrial camera located on an upper side of the belt conveyor, the industrial camera being mounted in the center of the top within the image acquisition bin;

the industrial camera light sources are arranged on the left side and the right side of the top in the image acquisition bin;

the camera dustproof cover is sleeved on the industrial camera, and a transparent cover is arranged at the front end of the camera dustproof cover;

the windshield wiper is controlled by the steering engine and used for wiping the front end face of the transparent cover;

the water spraying device is used for spraying water to the front end face of the transparent cover;

the high-performance image processing server is connected with the industrial camera through a network port and used for reading images acquired by the industrial camera and deploying foreign matter detection service based on a Darknet target detection framework;

the client is used for displaying the detection state and the detection result;

and the client, the high-performance image processing server and the water spraying device are in signal connection with the control panel.

Preferably, the water spraying device comprises a water pump, a water tank and a spray head, and the spray head points to the front end face of the transparent cover.

Preferably, the foreign matter detection service based on the Darknet target detection framework adopts a YOLOv4 network model, and the foreign matter detection service migrates and learns on the basis of a pre-training model to train a sample set sampled in an actual belt conveyor material loading operation scene to obtain a foreign matter detection network model.

Preferably, the sample set is a sample set which is made in a YOLO support format by simulating various postures of field foreign matters in coal materials in a laboratory environment, traversing images shot in multiple directions and distances, and manually calibrating detection frames and categories.

Preferably, the foreign object detection network model is applied while expanding the sample set and fine tuning, so as to gradually obtain a model with high accuracy.

Preferably, the detection state refers to a connection state of an industrial camera and a connection state of a control panel, the high-performance image processing server can directly read the state of the industrial camera through driving of the industrial camera, and the working state of the control panel is monitored by heartbeat communication with the high-performance image processing server, and the state of the control panel and the state of a water spraying device connected with the control panel are judged according to the communication state and the communication message.

Preferably, the industrial camera light source comprises a strip led light source and a light homogenizing plate.

Preferably, the angle between the industrial camera and the horizontal plane is 60 degrees.

Compared with the prior art, the invention has the beneficial effects that:

1. the invention solves the dustproof problem of computer vision in the operation scene of the belt conveyor, and has practical application significance;

2. the invention enables the detection accuracy and efficiency of the computer vision detection method to reach higher level, thus having competitiveness;

3. the invention is a lightweight solving method for solving the problem of longitudinal tearing of the belt, has low cost and simple deployment, and has higher popularization value.

Drawings

FIG. 1 is a schematic structural view of the present invention;

FIG. 2 is a schematic diagram of the internal hardware deployment of an image acquisition bin;

FIG. 3 is a schematic block diagram of an industrial camera deployment;

FIG. 4 is a diagrammatic, schematic view of a camera dust cover;

FIG. 5 is a schematic diagram of the hardware device connections of the present invention;

fig. 6 is a flow diagram of a detection mode implementation of the present invention.

The reference numbers in the drawings: 1. an image acquisition bin; 11. a belt conveyor; 12. a channel; 2. an industrial camera; 3. an industrial camera light source; 4. a camera dust cover; 41. a transparent cover; 5. a wiper; 51. a steering engine; 6. a water spraying device; 61. a water pump; 62. a water tank; 63. a spray head; 7. a high performance image processing server; 8. a client; 9. and a control panel.

Detailed Description

The invention will be further illustrated with reference to the following specific examples. It should be understood that these examples are for illustrative purposes only and are not intended to limit the scope of the present invention. Further, it should be understood that various changes or modifications of the present invention may be made by those skilled in the art after reading the teaching of the present invention, and these equivalents also fall within the scope of the present application.