CN217605668U - Global vision detection device for counting machine learning identification - Google Patents
Global vision detection device for counting machine learning identification Download PDFInfo
- Publication number
- CN217605668U CN217605668U CN202220749079.8U CN202220749079U CN217605668U CN 217605668 U CN217605668 U CN 217605668U CN 202220749079 U CN202220749079 U CN 202220749079U CN 217605668 U CN217605668 U CN 217605668U
- Authority
- CN
- China
- Prior art keywords
- visual
- global
- machine learning
- counting machine
- learning identification
- 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.)
- Active
Links
Images
Abstract
The utility model provides a global visual detection device for the machine learning identification of a tablet counting machine, which comprises a visual mechanism and a dropping channel which is vertically arranged; the drop channel has an inlet at an upper end; the area of the falling channel from the inlet to a preset depth away from the inlet is a detection section; the visual mechanism comprises more than two visual catching devices, each visual catching device is arranged on one side of the detection section, and the working direction of each visual catching device faces the detection section. This global vision detection device can supply the product unsettled so that vision mechanism acquires its whole outward appearance through setting up the passageway that drops, utilizes a passageway that drops to set up the mode that a plurality of visuals caught the device, can acquire the global surface information of 360 of waiting to detect the product in the short time, has good implementation convenience.
Description
Technical Field
The utility model relates to the visual detection field, concretely relates to a global visual detection device for tablet counting machine learning discernment.
Background
In the field of packaging, for granular products such as tablets, the appearance of the products is usually required to be detected before being subpackaged and packaged so as to ensure that the appearance structure of the products is intact.
The attached drawing figure 1 is the product detection equipment structure under the prior art, when detecting the product under the prior art, generally spread the product out of arranging, then utilize visual equipment to detect the outward appearance structure of product from the top, visual equipment hardly carries out outward appearance to all surfaces of product in the short time and detects, has very big limitation in the actual implementation.
SUMMERY OF THE UTILITY MODEL
In order to overcome the defect that product detection equipment structure under the prior art exists, the utility model provides a global vision detection device for counting grain machine learning discernment sets up the mode that a plurality of visuals were caught and grabs the device through dropping a passageway to acquire the global surface outward appearance information of 360 of waiting to detect the product in the short time, thereby provide more comprehensive data support for subsequent machine learning discernment, have good implementation convenience.
Correspondingly, the utility model provides a global visual detection device for the machine learning identification of a tablet counting machine, which comprises a visual mechanism and a dropping channel which is vertically arranged;
the drop channel has an inlet at an upper end;
the area of the falling channel from the inlet to a preset depth away from the inlet is a detection section;
the visual mechanism comprises more than two visual catching devices, each visual catching device is arranged on one side of the detection section, and the working direction of each visual catching device faces the detection section.
In an optional implementation manner, the number of the dropping channels is more than two, and all the dropping channels are arranged in sequence along a fixed direction;
any two adjacent falling channels are separated by a solid partition plate;
each is provided with a set of on the detection section of passageway that drops correspondingly vision mechanism.
In an alternative embodiment, in each group of the vision mechanisms, the number of the vision capturing devices is two, and the two vision capturing devices are symmetrically arranged about the corresponding detection segment.
In an alternative embodiment, all the visual capturing devices are arranged in two rows, and the arrangement direction of the visual capturing devices in each row is consistent with the arrangement direction of the falling channel.
In an alternative embodiment, all of the visual capture devices in each row are disposed on a mounting carrier.
In an alternative embodiment, the two or more visual capture devices are located at the same height in the same set of visual mechanisms.
In an alternative embodiment, in the same set of vision mechanisms, the two or more vision capture devices are evenly arranged around the circumference of the corresponding drop chute.
In an alternative embodiment, the vision mechanism further comprises a light source, which is arranged above the drop channel and faces the drop channel.
In an alternative embodiment, the drop channel is formed on the basis of a structural enclosure; the structural member is arranged in a hollow manner in the area corresponding to each visual capture device or the structural member is of a transparent structure in the area corresponding to each visual capture device.
In an alternative embodiment, the visual capture device comprises a CMOS sensor.
To sum up, the utility model provides a global visual inspection device for tablet counting machine learning identification, which can be used for suspending a product by setting a dropping channel so that a visual mechanism can acquire the whole appearance of the product, and can acquire 360 degrees global surface appearance information of the product to be detected in a short time by setting a plurality of visual catching devices by using a dropping channel; the two visual capturing devices symmetrically arranged about the falling channel can efficiently complete the function of acquiring all surface information of a product to be detected, so that more comprehensive data support is provided for subsequent machine learning identification, and the device has good implementation convenience.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 shows a structure of a product inspection apparatus in the prior art.
Fig. 2 is a schematic diagram of a structure of the global vision inspection apparatus according to an embodiment of the present invention.
Fig. 3 is a schematic side view of the global visual inspection device according to an embodiment of the present invention.
Fig. 4 is a schematic cross-sectional view of a front view structure of a global visual inspection device according to an embodiment of the present invention.
Fig. 5 is a schematic diagram of a three-dimensional perspective structure of a product to be detected according to an embodiment of the present invention.
Fig. 6 is the utility model discloses detection mechanism treats the first operation schematic diagram that detects the product.
Fig. 7 is a second operation schematic diagram of the detection mechanism according to the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by a person of ordinary skill in the art without creative efforts belong to the protection scope of the present invention.
The first embodiment is as follows: global vision detection device for counting machine learning identification
Fig. 2 shows the structure schematic diagram of the global visual detection device of the embodiment of the utility model, fig. 3 shows the utility model discloses a global visual detection device's side-looking structure section schematic diagram, fig. 4 shows the utility model discloses a global visual detection device's front-looking structure section schematic diagram.
The embodiment of the utility model provides a global vision detection device for tablet counting machine learning discernment, basic, this global vision detection device includes vision mechanism and the passageway 3 that drops along vertical arrangement.
Specifically, the falling channel 3 is a section of vertical space for the products to be detected to fall, and basically, the falling channel 3 is provided with an inlet at the upper end for the products to be detected to enter; for the sake of illustration, the area of the drop channel 3 from the entrance to a predetermined depth from the entrance is defined (named) as a detection section 11.
Basically, the vision mechanism includes more than two vision capturing devices 5, each vision capturing device 5 is arranged on one side of the detection section 11, the working direction of each vision capturing device 5 faces the detection section 11, and each vision capturing device 5 can capture an image of a product to be detected passing through the detection section 11 to obtain the appearance of the product to be detected.
Specifically, because the vision is caught device 5 and is occupied entity space, when the direction of operation of catching device 5 when two or more vision all towards detection section 11, the scope of the image that two vision caught device 5 acquireed is different certainly, compares in only a vision and catches device 5's the structure that sets up, the utility model discloses the implementation structure of vision mechanism can acquire the more surface image information of product.
Generally, the vision mechanism is connected with an upper computer such as a central control host computer, captured image information is fed back to the upper computer, appearance data about a product to be detected is provided for the upper computer, and accurate data support is carried out on subsequent operations of the upper computer.
Basically, the drop channel 3 is formed on the basis of the structural part 60; the structural member 60 is openly disposed in an area corresponding to each of the visual capture devices 5 or the structural member 60 is a transparent structure 61 in an area corresponding to each of the visual capture devices 5.
The vision mechanism further comprises a light source 4, wherein the light source 4 is arranged above the falling channel 3, and the light source 4 faces the falling channel 3. Typically, the material of the structure 60 is not light absorbing, and in practice, the light reflection can make the falling channel 3 have enough light intensity at the position of the detection section 11 for the visual mechanism to work.
Specifically, the visual capture device 5 includes a CMOS sensor, which may be a commercial grade CMOS, to reduce procurement costs.
Furthermore, the number of the falling channels 3 is more than two, and all the falling channels 3 are arranged in sequence along a fixed direction; any two adjacent dropping channels 3 are separated by a solid partition plate 62; each the detection section 11 of the falling channel 3 is correspondingly provided with a group of vision mechanisms. As the visual mechanism is involved in the detection of the product to be detected, the falling speed of the product to be detected has a threshold value, and more than two falling channels 3 can be arranged in the specific production process for adapting the production line speed.
Basically, in the same set of the vision mechanisms, the two or more vision capturing devices 5 are located at the same height, so that the overlapping range of the image capturing regions of the vision capturing devices 5 in the vision mechanisms is increased, and it is ensured that the acquired image can contain as much appearance information of the tablet as possible.
In the same group of vision mechanisms, the more than two vision catching devices 5 are uniformly arranged around the circumference of the corresponding falling channel 3, so that the vision catching range of each vision catching device 5 can be utilized to the maximum.
Further, in each group of the vision mechanisms, the number of the vision capturing devices 5 is two, and the two vision capturing devices 5 are symmetrically arranged with respect to the corresponding detection section 11.
Specifically, the two visual capturing devices 5 are symmetrically arranged with respect to the corresponding detection segment 11, so that there is no overlapping area between the surface image information of the to-be-detected product acquired by the two visual capturing devices 5, which is beneficial to acquiring the most surface information of the to-be-detected product by using a smaller number of visual capturing devices 5 in the shortest time.
In addition, the image acquisition principle of the visual mechanism is further analyzed.
Fig. 5 shows a schematic three-dimensional perspective structure of a product to be detected according to an embodiment of the present invention, specifically, a tablet which is cylindrical and has a front surface 71, a back surface 72 and a side wall 73 is taken as an example.
Fig. 6 of the accompanying drawings shows the first operation schematic diagram of the product to be detected by the detection mechanism of the embodiment of the present invention, when the tablet is in the posture of the detection section 11 as shown in fig. 4 of the accompanying drawings, because light travels along a straight line, when the front 71 and the back 72 of the tablet are just right-facing with a visual capturing device 5 respectively, the visual capturing device 5 cannot perform image acquisition on the side wall 73 of the tablet, that is, if the visual capturing device 5 in a set of visual mechanism only acquires a visual image about the product to be detected, there is a possibility that the tablet in the posture shown in fig. 4 of the accompanying drawings will be missed.
Fig. 7 of the accompanying drawings shows the second operation schematic diagram of the product to be detected by the detection mechanism of the embodiment of the present invention, and similarly, when the tablet is in the posture of the detection section 11 as shown in fig. 5 of the accompanying drawings, each of the visual capturing devices 5 can capture most of the side wall 73 and the front 71 (the side wall 73 and the back 72), but there are still very small visual blind areas above and below the tablet graphic direction, and if the visual capturing devices 5 in a set of visual mechanisms all only obtain one visual image about the product to be detected, the image about the visual blind area cannot be obtained, and there is a possibility of missing detection.
It should be noted that the tablet positions shown in fig. 4 and 5 are both located near the axis of the visual capture device 5, and in actual practice, since the visual capture range of the visual capture device 5 is fan-shaped, the blind spot position will change accordingly when the tablet position is within the visual capture range of the visual capture device 5 but away from the axis of the visual capture device 5; theoretically, when the posture of the tablet itself is not changed, the visual capturing device 5 can acquire the appearance information of all the surfaces of the tablet by acquiring two images when the tablet is within the visual capturing range of the visual capturing device 5, wherein the tablet position of one image is located above the axis of the visual capturing device 5 and the tablet position of the other image is located below the axis of the visual capturing device 5.
However, in actual operation, because the movement of the tablet in the falling body channel is a free falling body movement, when the visual capturing device 5 respectively acquires a plurality of images of the same tablet, the postures of the tablet in different images are inconsistent, but because the shooting speed of the visual capturing device 5 is high, the posture of the tablet changes, but the change range is small, the probability of generating a blind area is small, and the visual mechanism can acquire all appearance information of the tablet more comprehensively.
For the implementation structure in which a plurality of dropping channels 3 are arranged side by side, all the visual catching devices 5 are arranged in two rows, and the arrangement direction of the visual catching devices 5 in each row of the visual catching devices 5 is the same as the arrangement direction of the dropping channels 3. In addition, all the visual capture devices 5 in each row of the visual capture devices 5 are arranged on a mounting carrier, so that the mounting and assembly are integrally carried out, and the mounting process is simplified.
To sum up, the embodiment of the present invention provides a global visual inspection device for the machine learning identification of a tablet counting machine, which can suspend a product in the air by setting a drop channel so that the visual mechanism can obtain the overall appearance of the product, and can obtain 360 ° global surface appearance information of the product to be detected in a short time by setting a plurality of visual catching devices through one drop channel; the two visual capturing devices symmetrically arranged about the falling channel can efficiently complete the function of acquiring all surface information of a product to be detected, so that more comprehensive data support is provided for subsequent machine learning identification, and the device has good implementation convenience.
The global visual inspection device for the machine learning identification of the tablet counting machine provided by the embodiment of the present invention is introduced in detail, and the specific examples are applied herein to explain the principle and the implementation of the present invention, and the description of the above embodiments is only used to help understand the method and the core idea of the present invention; meanwhile, for the general technical personnel in the field, according to the idea of the present invention, there are changes in the specific implementation and application scope, to sum up, the content of the present specification should not be understood as the limitation of the present invention.
Claims (10)
1. The global visual detection device for the machine learning identification of the tablet counting machine is characterized by comprising a visual mechanism and a dropping channel which is vertically arranged;
the drop channel has an inlet at an upper end;
the area of the falling channel from the inlet to a preset depth away from the inlet is a detection section;
the visual mechanism comprises more than two visual capture devices, each visual capture device is arranged on one side of the detection section, and the working direction of each visual capture device faces the detection section.
2. The global visual inspection device for counting machine learning identification of claim 1, wherein the number of said drop channels is more than two, all said drop channels are arranged in sequence along a fixed direction;
any two adjacent dropping channels are separated by a solid partition plate;
each is provided with a set of on the detection section of passageway that drops correspondingly vision mechanism.
3. The global visual inspection device for counting machine learning identification according to claim 2, wherein in each set of said visual mechanisms, the number of said visual capturing devices is two, and two of said visual capturing devices are symmetrically arranged with respect to the corresponding inspection section.
4. The global visual inspection device for machine learning identification of a particle counting machine according to claim 3, wherein all the visual capture devices are arranged in two rows, and the arrangement direction of the visual capture devices in each row of the visual capture devices is the same as the arrangement direction of the falling passage.
5. The global vision inspection device for counting machine learning identification of claim 4, wherein all the visual capture devices in each row of the visual capture devices are disposed on a mounting carrier.
6. The global visual inspection device for counting machine learning identification according to claim 1, wherein said two or more visual capture devices are located at the same height in the same set of said visual mechanisms.
7. The global visual inspection device for counting machine learning identification according to claim 1, wherein in the same set of said visual mechanisms, said two or more visual grasping devices are uniformly arranged around the circumference of the corresponding drop chute.
8. The global visual inspection device for counting machine learning identification of claim 1, wherein the visual mechanism further comprises a light source disposed above the drop chute with the light source facing in the direction of the drop chute.
9. The global visual inspection device for tablet counting machine learning identification of any one of claims 1 to 8, wherein the drop channel is formed based on structural enclosure; the structural member is arranged in a hollow manner in the area corresponding to each visual capture device or the structural member is of a transparent structure in the area corresponding to each visual capture device.
10. The global visual inspection device for counting machine learning identification of claim 1, wherein the visual capture device comprises a CMOS sensor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202220749079.8U CN217605668U (en) | 2022-04-01 | 2022-04-01 | Global vision detection device for counting machine learning identification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202220749079.8U CN217605668U (en) | 2022-04-01 | 2022-04-01 | Global vision detection device for counting machine learning identification |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN217605668U true CN217605668U (en) | 2022-10-18 |
Family
ID=83564449
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202220749079.8U Active CN217605668U (en) | 2022-04-01 | 2022-04-01 | Global vision detection device for counting machine learning identification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN217605668U (en) |
-
2022
- 2022-04-01 CN CN202220749079.8U patent/CN217605668U/en active Active
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN208915599U (en) | For controlling or the unit of management product and equipment for manufacturing product | |
| CN106040600B (en) | The automatic sorting method of cylindrical parts | |
| US5550583A (en) | Inspection apparatus and method | |
| US11392211B2 (en) | Resource-responsive motion capture | |
| US8780358B2 (en) | Inspection apparatus, system, and method | |
| CN106767423A (en) | Automatic Visual Inspection instrument | |
| BR112020023799A2 (en) | SEED CLASSIFICATION | |
| CN107016373A (en) | The detection method and device that a kind of safety cap is worn | |
| CN109458941A (en) | The automatic load measurement system of mine car and monitoring method | |
| CN110182561A (en) | A kind of logistics equipment of the rapid weighing scanning amount side of collection retrospect one | |
| CN217605668U (en) | Global vision detection device for counting machine learning identification | |
| Chao et al. | Design of a dual-camera system for poultry carcasses inspection | |
| CN101354360A (en) | Method for detecting capsules or tablets packaged by aluminum / aluminum bubbling hood | |
| US11035926B2 (en) | Determining the orientation of objects in space | |
| KR102166287B1 (en) | Apparatus for Checking Packaging Equipment Automation Machine | |
| CN211906310U (en) | Zynq-based machine vision detection system | |
| KR102271342B1 (en) | the capsule inspection system and method | |
| CN109701900A (en) | A kind of article assembly system | |
| CN206631957U (en) | A kind of connecting rod forging vision detection system | |
| JP6270319B2 (en) | X-ray inspection equipment | |
| CN109187593A (en) | A kind of control method and system of detection device | |
| CN108709892A (en) | Detecting system and its method | |
| EP2141486A1 (en) | Device for detecting and classifying residual oxide in metal sheet production lines | |
| US20220334040A1 (en) | Apparatus and Method for Monitoring and Recording Disintegration Times for Pharmaceutical Products | |
| CN111667484A (en) | Rapid counting method for chip scattered components |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| GR01 | Patent grant | ||
| GR01 | Patent grant |