Factory operation picture recognition device based on miniGPT

Abstract

The utility model relates to the technical field of picture recognition devices, in particular to a factory operation picture recognition device based on miniGPT, which comprises four vertical beams arranged in a matrix mode, a bottom plate is fixedly mounted at the bottom ends of the vertical beams, a bottom rubber pad is arranged at the bottom of the bottom plate, a bearing plate is fixedly mounted at the top ends of the vertical beams, and the bearing plate is fixedly mounted at the top ends of the vertical beams. A rectangular rod is fixedly installed at the center position of the upper surface of the bearing plate, the rectangular rod is sleeved with a damping pad and further sleeved with rectangular sleeves, the rectangular sleeves are fixedly installed on the bearing plate, a top plate is fixedly installed among the four rectangular sleeves, an image acquisition module is arranged at the bottom of the top plate, and a U-shaped sleeve is fixedly installed on the vertical beam. A fixing rod is fixedly installed between the two U-shaped sleeves which are right opposite in the front-back direction, a supporting rod is fixedly installed on the image collecting module, and a support is fixedly installed at the top end of the supporting rod. The supporting device is firm and stable in structure and can play a role in stable supporting.

Description

Technical Field

[0001] The utility model relates to the technical field of picture recognition devices, and in particular to a factory operation picture recognition device based on miniGPT. Background Art

[0002] With the rapid development of science and technology, especially breakthroughs in artificial intelligence and deep learning, image recognition technology has been widely used in various fields, significantly improving production efficiency and safety. In the field of industrial automation, image recognition technology plays an irreplaceable role. It helps improve product quality, optimize production processes, and ensure production safety by real-time detection and analysis of image data in the production process.

[0003] Currently, image recognition technologies widely used in industrial automation systems mostly rely on traditional computer vision algorithms, such as convolutional neural networks. These algorithms perform well in specific scenarios, but still have many limitations when dealing with complex and changeable factory operating environments. For example, traditional algorithms often lack recognition accuracy and robustness when dealing with complex situations such as low-quality images, light changes, and obstructions.

[0004] miniGPT is a kind of multimodal recognition function for images by combining pre-trained ViT and Q-Former visual encoders. MiniGPT not only has powerful text generation capabilities, but also can recognize and analyze image content, realizing the deep fusion of text and image. Applying miniGPT to factory operation picture recognition devices has the advantage of being able to efficiently handle complex and changeable factory operation environments.

[0005] Since the image recognition device needs to use a camera to capture the picture, most factory operation image recognition devices use a bracket to fix the camera in the corresponding position. This type of bracket lacks the corresponding shock-absorbing component and cannot achieve shock-absorbing operation. In addition, this type of bracket is generally a vertical rod, which lacks limit components around it, which will cause the structure of the vertical rod itself to be not strong and stable enough, which is not conducive to the effect of stable support. It will cause the camera part to shake easily with external collisions, affecting the stability of its picture capture. In view of this, we propose a factory operation image recognition device based on miniGPT. Utility Model Content

[0006] The purpose of the utility model is to provide a factory operation picture recognition device based on miniGPT to solve the defects mentioned in the above background technology.

[0007] In order to achieve the above purpose, the utility model provides the following technical solutions:

[0008] A factory operation picture recognition device based on miniGPT comprises four vertical beams arranged in a matrix, a bottom plate is fixedly installed at the bottom end of the vertical beam, a bottom rubber pad is arranged at the bottom of the bottom plate, a bearing plate is fixedly installed at the top end of the vertical beam, a rectangular rod is fixedly installed at the center position of the upper surface of the bearing plate, a shock-absorbing pad is sleeved on the rectangular rod, a rectangular sleeve is also sleeved on the rectangular rod, the rectangular sleeve is fixedly installed on the bearing plate, a top plate is fixedly installed between the four rectangular sleeves, an image acquisition module is arranged at the bottom of the top plate, a U-shaped sleeve is fixedly installed on the vertical beam, and a fixing rod is fixedly installed between two U-shaped sleeves facing each other front and back.

[0009] Preferably, a plurality of positioning tubes arranged in a matrix are fixedly mounted on the bottom plate, a plurality of through holes arranged in a matrix are provided on the bottom rubber pad, and the positioning tubes are plug-fitted into the through holes.

[0010] Preferably, the number of the positioning cylinders and the number of the through holes are both 3 to 6, and the size of the positioning cylinders is matched with the size of the through holes.

[0011] Preferably, a support rod is fixedly mounted on the image acquisition module, a support is fixedly mounted on the top end of the support rod, and the support is fixedly mounted on the bottom surface of the top plate.

[0012] Preferably, a plurality of bolt holes arranged in a matrix and connected to the outside are provided in the top plate, and the bolts are fixed by passing through the supports and the bolt holes.

[0013] Preferably, the rectangular sleeve presses the shock-absorbing pad onto the upper surface of the bearing plate, and the thickness of the shock-absorbing pad is 1.8 cm to 3 cm.

[0014] Preferably, a processing unit is fixedly mounted on one of the vertical beams, and a storage module and a communication module are provided on the processing unit.

[0015] Compared with the prior art, the beneficial effects of the utility model are:

[0016] 1. The utility model can perform image capture operations through the image acquisition module. In addition, it supports through four vertical beams and limits through fixed rods, so that the structure between the four vertical beams is more solid and stable. In addition, the shock-absorbing pads and bottom rubber pads can provide a certain shock-absorbing protection effect, so that the structure of the image acquisition module is more stable and not easy to shake, and a stable image capture effect can be achieved.

[0017] 2. The utility model realizes further fixing operation between the four vertical beams by providing a top plate and a rectangular sleeve, thereby ensuring that the structure between the four vertical beams is more solid and stable. In addition, by providing a plurality of bolt holes, the position of the support can be adjusted for fixed installation operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic diagram of the overall structure of the utility model;

[0019] Figure 2 This is one of the explosion structure schematic diagrams of the utility model;

[0020] Figure 3 This is the second schematic diagram of the explosion structure of the utility model;

[0021] Figure 4 For this utility model Figure 3 The enlarged view of point A in the middle;

[0022] Figure 5 This is a system module block diagram of the utility model.

[0023] The meaning of each number in the figure is:

[0024] 1. vertical beam; 10. bearing plate; 11. rectangular rod; 12. shock-absorbing pad; 13. fixing rod; 131. U-shaped sleeve; 14. bottom plate; 141. positioning cylinder; 15. bottom rubber pad; 151. through hole;

[0025] 2. Image acquisition module; 20. Support rod; 21. Support;

[0026] 3. Top plate; 30. Bolt hole; 31. Rectangular sleeve;

[0027] 4. Processing unit; 40. Storage module; 41. Communication module. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the utility model to clearly and completely describe the technical solutions in the embodiments of the utility model. Obviously, the described embodiments are only part of the embodiments of the utility model, not all of the embodiments. Based on the embodiments in the utility model, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the utility model.

[0029] See also Figure 1-Figure 5The utility model provides a technical solution: a factory operation picture recognition device based on miniGPT, comprising four vertical beams 1 arranged in a matrix, a bottom plate 14 is fixedly installed at the bottom end of the vertical beam 1, a bottom rubber pad 15 is arranged at the bottom of the bottom plate 14, a plurality of positioning cylinders 141 arranged in a matrix are fixedly installed on the bottom plate 14, a plurality of through holes 151 arranged in a matrix are arranged on the bottom rubber pad 15, the positioning cylinders 141 and the through holes 151 are plugged and matched, the number of the positioning cylinders 141 and the through holes 151 are 3 to 6, the size of the positioning cylinders 141 is adapted to the size of the through holes 151, the bottom plate 14 and the bottom rubber pad 15 are fixedly installed on the external foundation by a plurality of fastening bolts, so as to realize stable support operation;

[0030] Specifically, a bearing plate 10 is fixedly installed on the top of the vertical beam 1, a rectangular rod 11 is fixedly installed at the center of the upper surface of the bearing plate 10, a shock-absorbing pad 12 is sleeved on the rectangular rod 11, a rectangular sleeve 31 is also sleeved on the rectangular rod 11, the rectangular sleeve 31 is fixedly installed on the bearing plate 10, and a top plate 3 is fixedly installed between the four rectangular sleeves 31, so as to further limit and fix the four vertical beams 1, so that the overall structure is more firm and stable;

[0031] Specifically, a U-shaped sleeve 131 is fixedly installed on the vertical beam 1 , and a fixing rod 13 is fixedly installed between two U-shaped sleeves 131 facing each other at the front and rear ends, so as to fix the front and rear vertical beams 1 .

[0032] In this embodiment, an image acquisition module 2 is provided at the bottom of the top plate 3, a support rod 20 is fixedly installed on the image acquisition module 2, a support 21 is fixedly installed on the top of the support rod 20, and the support 21 is fixedly installed on the bottom surface of the top plate 3; a plurality of bolt holes 30 arranged in a matrix and connected to the outside are provided in the top plate 3, and the bolts are fixed by passing through the support 21 and the bolt holes 30, so as to facilitate the adjustment of the position of the support 21 for fixing operation.

[0033] Furthermore, the rectangular sleeve 31 presses the shock-absorbing pad 12 onto the upper surface of the bearing plate 10 . The thickness of the shock-absorbing pad 12 is 1.8 cm to 3 cm, so that the shock-absorbing pad 12 can achieve a certain shock-absorbing protection effect.

[0034] In addition, a processing unit 4 is fixedly installed on one of the vertical beams 1, and a storage module 40 and a communication module 41 are provided on the processing unit 4. The image acquisition module 2 is used to capture high-definition pictures or video frames of the factory operation site; the processing unit 4 is equipped with a high-performance processor and GPU to run the miniGPT model and necessary image processing algorithms; the storage module 40 is used to store model parameters, recognition results and historical data; the communication module 41 supports wireless or wired network connection to realize remote transmission and reception of data.

[0035] It is worth noting that a software architecture is set up in the processing unit 4, which includes a miniGPT model. It adopts a lightweight GPT architecture and has the ability to deeply understand factory operation pictures through pre-training and fine-tuning; the image processing algorithm includes image preprocessing, such as denoising, enhancement, feature extraction and matching, etc., to optimize the recognition effect; the recognition engine is based on the miniGPT model and image processing algorithm, and intelligently identifies and analyzes the input picture, and outputs the recognition result; monitoring and management system: integrated in the factory information platform, receives the recognition results, and realizes real-time monitoring of the operation process, abnormal warning and data analysis.

[0036] It is worth noting that the image acquisition module 2 is used to capture pictures of the factory operation site in real time. The processing unit 4 receives the pictures, and after pre-processing by the image processing algorithm, it is input into the miniGPT model for recognition. The miniGPT model intelligently recognizes and analyzes the operation scenes, personnel behaviors, equipment status, etc. in the pictures based on pre-trained knowledge and task requirements. The recognition results are transmitted to the monitoring and management system through the communication module 41 to realize real-time monitoring of the operation process and abnormal warning.

[0037] Finally, it should be noted that the processing unit 4, storage module 40, communication module 41, image acquisition module 2 and other components involved in the utility model are all universal standard parts or components known to technical personnel in this field, and their structures and principles are all known to technical personnel in this field through technical manuals or through conventional experimental methods. In the idle space of this device, all the above-mentioned electrical components, which refer to power elements, electrical components, and adapted controllers and power supplies, are connected through wires. The specific connection means should refer to the working principle of the utility model. The electrical connection between each electrical component is completed in a sequential working order, and the detailed connection means are all well-known technologies in the field.

[0038] When the miniGPT-based factory operation picture recognition device of the utility model is in use, the positioning cylinder 141 is inserted into the through hole 151, and then the bottom plate 14 and the bottom rubber pad 15 are fixedly installed on the concrete base with fastening bolts, and then the U-shaped sleeve 131 is fixedly installed on the corresponding vertical beam 1, and then the shock-absorbing pad 12 is sleeved on the rectangular rod 11, and the rectangular sleeve 31 is fixedly installed on the bearing plate 10, and finally, the support 21 is fixedly installed on the top plate 3 with fastening bolts to complete the fixing operation.

[0039] The above shows and describes the basic principle, main features and advantages of the utility model. Those skilled in the art should understand that the utility model is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the utility model and are not used to limit the utility model. Without departing from the spirit and scope of the utility model, the utility model may have various changes and improvements, which fall within the scope of the utility model to be protected. The scope of protection of the utility model is defined by the attached claims and their equivalents.

Claims

1. A factory operation picture recognition device based on miniGPT, characterized in that: The invention comprises four vertical beams (1) arranged in a matrix, wherein a bottom plate (14) is fixedly mounted at the bottom end of the vertical beam (1), a bottom rubber pad (15) is arranged at the bottom of the bottom plate (14), a bearing plate (10) is fixedly mounted at the top end of the vertical beam (1), a rectangular rod (11) is fixedly mounted at the center position of the upper surface of the bearing plate (10), a shock-absorbing pad (12) is sleeved on the rectangular rod (11), a rectangular sleeve (31) is also sleeved on the rectangular rod (11), the rectangular sleeve (31) is fixedly mounted on the bearing plate (10), a top plate (3) is fixedly mounted between the four rectangular sleeves (31), an image acquisition module (2) is arranged at the bottom of the top plate (3), a U-shaped sleeve (131) is fixedly mounted on the vertical beam (1), and a fixed rod (13) is fixedly mounted between two U-shaped sleeves (131) facing each other front and back.

2. The factory operation picture recognition device based on miniGPT according to claim 1 is characterized in that: A plurality of positioning cylinders (141) arranged in a matrix are fixedly mounted on the bottom plate (14), a plurality of through holes (151) arranged in a matrix are provided on the bottom rubber pad (15), and the positioning cylinders (141) are plug-fitted into the through holes (151).

3. The factory operation picture recognition device based on miniGPT according to claim 2 is characterized in that: The number of the positioning cylinders (141) and the number of the through holes (151) are both 3 to 6, and the size of the positioning cylinders (141) is matched to the size of the through holes (151).

4. The factory operation picture recognition device based on miniGPT according to claim 1 is characterized in that: A support rod (20) is fixedly mounted on the image acquisition module (2), a support seat (21) is fixedly mounted on the top end of the support rod (20), and the support seat (21) is fixedly mounted on the bottom surface of the top plate (3).

5. The factory operation picture recognition device based on miniGPT according to claim 4 is characterized in that: The top plate (3) is provided with a plurality of bolt holes (30) arranged in a matrix and connected to the outside, and the bolts pass through the support (21) and the bolt holes (30) for fixing.

6. The factory operation picture recognition device based on miniGPT according to claim 1 is characterized in that: The rectangular sleeve (31) presses the shock-absorbing pad (12) onto the upper surface of the bearing plate (10); the thickness of the shock-absorbing pad (12) is 1.8 cm to 3 cm.

7. The factory operation picture recognition device based on miniGPT according to claim 1 is characterized in that: A processing unit (4) is fixedly mounted on one of the vertical beams (1), and a storage module (40) and a communication module (41) are arranged on the processing unit (4).

Images