CN114331967A - Visual collection image quality inspection method and system - Google Patents

Abstract

The invention provides a method and system for quality inspection of visually collected images. The method includes: establishing a pure noise data benchmark based on a carrier material; receiving original image data, synthesizing the original image data and the pure noise data benchmark into standard comparison data; The image data to be tested is compared with the image data to be tested and the standard comparison data, and the quality inspection judgment result is obtained. First establish a pure noise data benchmark for the carrier material, which includes optical noise in the visual acquisition process, and the original image data contains correct information, neither environmental noise nor product defects, and the original image data and pure noise data benchmarks are synthesized as a standard comparison Data, which contains the correct information of the original image data and unavoidable optical noise, compares the standard comparison data with the image data to be tested, and can clearly identify product defects in the image data to be tested, avoid misjudgment, and the standard comparison data. The establishment method is simple, the labor cost is saved, the processing efficiency is high, and the misjudgment rate is low.

Description

A kind of visual acquisition image quality inspection method and system

technical field

The invention belongs to the technical field of intelligent manufacturing, and in particular relates to a method and system for quality inspection of visually collected images.

Background technique

In the production process of printed products, visual inspection is an important link in determining the degree of quality completion. At present, the visual inspection method for printed matter is as follows: first, perform machine vision collection on the finished printed matter to obtain a small batch of picture data, and then manually select a good product from this batch of picture data as a standard part picture, and use the Pictures of standard parts are entered into the processing system and used as comparison standards for subsequent inspections.

This method is suitable for mass production. A standard part picture can meet the quality inspection of mass printing. Compared with mass production, the manual screening when determining the standard part picture is acceptable, even if the standard part is determined. The efficiency of the image processing is very low. With the increasing demand for small batches and customization of printed materials, the above-mentioned quality inspection methods with high labor input cost and low efficiency seem out of place.

SUMMARY OF THE INVENTION

The purpose of the present invention is to overcome the above-mentioned deficiencies in the prior art, and to provide a visual acquisition image quality inspection method and system, which is mainly used to solve the problems of low efficiency, high labor cost, and incompatibility with small batch customization in the prior art printed product inspection process. production issues.

In order to achieve the above-mentioned purpose, in the first aspect, the present invention provides a quality inspection method for visually captured images, comprising the following steps:

Establish a pure noise data benchmark based on carrier material;

receiving original image data, and synthesizing the original image data and the pure noise data benchmark into standard comparison data;

The image data to be tested is collected visually, and the image data to be tested is compared with the standard comparison data to obtain a quality inspection judgment result.

Further, before establishing the pure noise data benchmark based on the carrier material, environmental factors are determined first, and the environmental factors include but are not limited to visual focus information and visual auxiliary light information.

Further, when establishing a pure noise data benchmark based on carrier materials, M pieces of carrier material noise maps are collected under certain environmental factors, and an N1×N2 unit cell lattice map is established for each carrier material noise map. The same divided area of the noise map of the carrier material is calculated by the matrix analysis algorithm, and the pure noise data benchmark of the carrier material is established based on the calculation result of the matrix analysis algorithm.

Further, the same divided area is a unit cell area with a size of A×A, where A≧1.

Further, the establishment of pure noise data benchmarks is carried out for different carrier material types, and a carrier material data benchmark database is established.

Further, when receiving the original image data, and synthesizing the original image data and the pure noise data benchmark into standard comparison data, the following steps are included:

Receive a quality inspection task data package, the quality inspection task data package includes at least original image data and carrier material information, the original image data is the original design image data, and retrieved from the carrier material data benchmark database according to the carrier material information For the corresponding pure noise data benchmark, the original image data is combined with the extracted pure noise data benchmark to obtain standard comparison data.

Further, when comparing the image data to be measured and the standard contrast data, the following steps are included:

Establish N1×N2 unit cell bitmaps for the image data to be measured and the standard comparison data, respectively, perform matrix value calculation for the same divided area of the two unit cell bitmaps, and calculate the matrix values that belong to the same divided area. values are compared.

Further, when the matrix value calculation is performed, the unit cell bitmap of the image data to be tested and the standard comparison data is divided according to the A×A unit cell area, and the matrix value calculation is performed for each unit cell area. .

Further, when using vision to collect the image data to be tested, the product to be tested is firstly collected by machine vision, and the collected image is cropped to obtain the image data to be tested.

In a second aspect, the present invention also provides a quality inspection system applied to the above-mentioned visually captured images, comprising:

Data benchmark database, including several pure noise data benchmarks based on carrier materials;

a preprocessing unit, configured to receive original image data, and synthesize the original image data and the pure noise data benchmark into standard comparison data;

The acquisition unit is used to visually collect the image data to be measured;

A comparison unit, configured to compare the image data to be tested and the standard comparison data to obtain a quality inspection judgment result.

Compared with the prior art, the beneficial effects of the present invention at least include:

First, establish a pure noise data benchmark for the carrier material. This pure noise data benchmark includes optical noise in the visual acquisition process, while the original image data contains correct information, neither environmental noise nor product defects. Therefore, the original image data and After the pure noise data benchmark is synthesized into the standard comparison data, the standard comparison data contains the correct information of the original image data and unavoidable optical noise. The establishment of standard comparison data only needs to synthesize the original image data with the pure noise data benchmark system, saving labor costs, high processing efficiency, low false positive rate, and good quality inspection.

Description of drawings

The present invention will be further described by using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. For those of ordinary skill in the art, under the premise of no creative work, other Attached.

FIG. 1 is a schematic flowchart of a quality inspection method for visually captured images provided by the present invention.

FIG. 2 is a schematic flow chart of the establishment of the carrier material data reference database in the present invention.

FIG. 3 is a specific flow chart of comparing the image data to be tested with the standard comparison data in the present invention.

FIG. 4 is a schematic diagram of a visual acquisition image quality inspection system provided by the present invention.

Detailed ways

The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are a part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. The indicated orientation or positional relationship is based on the orientation or positional relationship shown in the accompanying drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the indicated device or element must have a specific orientation or a specific orientation. construction and operation, and therefore should not be construed as limiting the invention. Furthermore, the terms "first", "second", and "third" are used for descriptive purposes only and should not be construed to indicate or imply relative importance.

Referring to FIGS. 1 to 3 , in a first aspect, the present invention provides a quality inspection method for visually captured images, comprising the following steps:

Establish a noise-only data benchmark based on carrier material, which includes optical noise during visual acquisition;

Receive the original image data, the information contained in the original image data is correct information, neither environmental noise nor product defects, the original image data and the pure noise data benchmark are synthesized into standard comparison data, so far, the standard comparison data Contains the correct information and inevitable optical noise of the original image data;

Use vision to collect the image data to be measured, and compare the image data to be measured with the standard comparison data. Since the same carrier material is used, the optical noise generated by the carrier material during the visual acquisition process has been included in the standard comparison data. , the part of the optical noise in the image data to be tested will not cause misjudgment in the comparison process. Only when the image data to be tested includes product defects, will it be exposed during the comparison process with the standard comparison data, so as to achieve Identify the optical noise and product defects on the product to be tested, and obtain the quality inspection result after the comparison.

The above methods can clearly identify product defects in the image data to be tested, and avoid misjudgment. The establishment of standard comparison data only needs to synthesize the original image data with the pure noise data benchmark system, which saves labor costs and has high processing efficiency. The misjudgment rate is low, and the quality of inspection is good.

With reference to FIG. 2, in this embodiment, before establishing the pure noise data benchmark based on the carrier material, the environmental factors are determined first, and the environmental factors include but are not limited to visual focus information and visual auxiliary light information. Depending on the lighting environment, camera focus settings, etc., environmental factors will be different for different carrier materials. After a certain carrier material is determined, the environmental factors related to it should be determined accordingly, and the subsequent collection and testing will be carried out. Image data while maintaining the same environmental factors.

Referring to FIG. 2 , as an implementation manner, when establishing a pure noise data benchmark based on carrier materials, under certain environmental factors, M pieces of carrier material noise maps are collected, and a unit of N1×N2 is established for each carrier material noise map. Element lattice map, in which the values of N1 and N2 can be the same or different; perform matrix analysis algorithm calculation on the same divided area of M noise maps of carrier materials, and establish pure noise data of carrier materials based on the calculation results of matrix analysis algorithm benchmark.

It should be noted that, for each carrier material noise image, first establish a unit element lattice map with a size of N1×N2, which can be a rectangle or a square according to the size of the carrier material. In this unit element lattice map, press The area is divided according to a certain rule. The rule of dividing the area can be divided according to the unit area of A×B, where the values of A and B can be the same or different; they can also be divided according to the area set artificially; for example, in 10× The 20 unit element lattice map is divided into 2×4 size, each divided area does not overlap, a total of 25 divided areas can be divided; under the same division rule, each carrier material noise map can be divided The same divided area is obtained, and the matrix analysis algorithm is calculated for the same divided area of the noise map of different carrier materials to obtain the average value of this divided area. Based on the calculation results of the matrix analysis algorithm, the pure noise data benchmark of the carrier material is established to calculate the accuracy. High, the coverage of noise calculation results is true and accurate.

In other embodiments, the same divided area is a unit cell area with a size of A×A, where A≥1, that is, the unit cell area is a square. Of course, it can be divided by a size of 1×1, that is Taking each unit element as the minimum division area, the accuracy is the highest.

In contrast, after knowing the general layout of the product to be tested, you can focus on the layout of the area where the label has more concentrated actual content. Relatively speaking, the grid area is thinner, and for the corner area, the grid is relatively The area is larger, so it can be divided according to the artificially set area.

In some embodiments, in order to adapt to the scenarios of various carrier materials, the establishment of pure noise data benchmarks is performed for different types of carrier materials, a carrier material data benchmark database is established, and the pure noise data benchmarks corresponding to various carrier materials are organized into Database, if you need to use the specific carrier material and then call it out.

3, more specifically, when receiving the original image data, and synthesizing the original image data and the pure noise data benchmark into standard comparison data, the following steps are included:

Receive a quality inspection task data package, the quality inspection task data package includes at least original image data and carrier material information, the original image data is the original design image data, and retrieved from the carrier material data benchmark database according to the carrier material information For the corresponding pure noise data benchmark, the original image data and the extracted pure noise data benchmark are data synthesized to obtain standard comparison data; using the data drive, the appropriate pure noise data benchmark is automatically retrieved, and then combined with the original design Image data, can quickly get standard comparison data.

In this embodiment, when comparing the image data to be measured and the standard comparison data, the following steps are included:

Establish N1×N2 unit cell bitmaps for the image data to be measured and the standard comparison data, respectively, perform matrix value calculation for the same divided area of the two unit cell bitmaps, and calculate the matrix values that belong to the same divided area. In the same way, the image data to be tested and the standard comparison data are divided into regions according to the same rule, and each divided region may have different situations in the corresponding two data. Judging the quality inspection results.

As an embodiment, when the matrix value calculation is performed, the unit cell bitmap of the image data to be tested and the standard comparison data is divided according to A×A unit cell area, that is, a square unit cell area As a comparison object, a matrix value calculation is performed for each unit cell area.

In this embodiment, when using vision to collect the image data to be tested, the product to be tested is firstly collected by machine vision, the collected image is cropped, and cut to a suitable size to obtain the image data to be tested, and then compared with the standard data. The comparison can improve the data processing efficiency and improve the comparison efficiency.

4, in the second aspect, the present invention also provides a quality inspection system applied to the above-mentioned visually captured images, including:

The data benchmark database includes several pure noise data benchmarks based on carrier materials, and correspondingly, a pure noise data establishment unit is also provided for establishing corresponding pure noise data benchmarks for different carrier materials;

a preprocessing unit, configured to receive original image data, and synthesize the original image data and the pure noise data benchmark into standard comparison data;

The acquisition unit is used to visually collect the image data to be measured;

A comparison unit, configured to compare the image data to be tested and the standard comparison data to obtain a quality inspection judgment result.

Compared with the prior art, the present invention provides a visual acquisition image quality inspection method and system. First, a pure noise data benchmark is established for the carrier material. The pure noise data benchmark includes the optical noise in the visual acquisition process, and the original image data. It contains correct information, neither environmental noise nor product defects. Therefore, after synthesizing the original image data and the pure noise data benchmark into standard comparison data, the standard comparison data contains the correct information of the original image data and unavoidable optical noise. After the standard comparison data is compared with the image data to be tested, product defects in the image data to be tested can be clearly identified to avoid misjudgment. The establishment of the standard comparison data only needs to synthesize the original image data with the pure noise data benchmark system , saving labor costs, high processing efficiency, low false positive rate, and good quality inspection.

Finally, it should be emphasized that the present invention is not limited to the above-mentioned embodiments, and the above are only preferred embodiments of the present invention, and are not intended to limit the present invention. Substitutions, improvements, etc., should all be included within the protection scope of the present invention.

The above description is the main process steps of the invention, in which other functional steps can be interspersed, and the above-mentioned logical sequence and process steps can be disrupted. be protected.

Claims (10)

1. A visual collection image quality inspection method is characterized by comprising the following steps:

establishing a pure noise data reference based on a carrier material;

receiving original image data, and synthesizing the original image data and the pure noise data into standard contrast data;

and acquiring image data to be detected by using vision, and comparing the image data to be detected with the standard comparison data to obtain a quality inspection judgment result.

2. The method of claim 1, wherein prior to establishing the pure noise data reference based on the carrier material, environmental factors are determined, wherein the environmental factors include, but are not limited to, visual focus information and visual auxiliary light information.

3. The method of claim 2, wherein in establishing the carrier-material-based pure noise data reference, M carrier-material noise maps are acquired under certain environmental factors, a single-element dot map of N1 × N2 is established for each carrier-material noise map, matrix analysis algorithm calculations are performed on the same divided regions of the M carrier-material noise maps, and the carrier-material-based pure noise data reference is established based on the matrix analysis algorithm calculations.

4. The method of claim 3, wherein said identically demarcated regions are unit-cell regions of size AxA, where A ≧ 1.

5. The method of claim 4, wherein the establishing of the pure noise data reference is performed for different types of carrier materials respectively to establish the carrier material data reference database.

6. The method of claim 5, wherein the step of receiving raw image data and combining the raw image data with the pure noise data reference to form standard contrast data comprises the steps of:

receiving a product inspection task data packet, wherein the product inspection task data packet at least comprises original image data and carrier material information, the original image data is original design image data, calling a corresponding pure noise data reference in a carrier material data reference database according to the carrier material information, and performing data synthesis on the original image data and the called pure noise data reference to obtain standard comparison data.

7. The visual inspection method of claim 6, wherein when comparing the image data to be inspected with the standard comparison data, the method comprises the following steps:

respectively establishing an N1 multiplied by N2 unit-element bitmap for the image data to be detected and the standard comparison data, respectively calculating matrix values of the same divided areas of the two unit-element bitmaps, and comparing the matrix values of the same divided areas.

8. The method as claimed in claim 7, wherein in the calculation of matrix values, the unit cell bitmap of the image data to be measured and the standard contrast data is divided into a unit cell area of a x a, and the calculation of matrix values is performed for each unit cell area.

9. The visual collection image quality inspection method according to claim 8, wherein when the image data to be inspected is collected visually, the machine vision collection is performed on the product to be inspected, and the collected image is cut to obtain the image data to be inspected.

10. A visual inspection system for use in any one of claims 1 to 9, comprising:

a data reference database comprising a plurality of pure noise data references based on a carrier material;

the preprocessing unit is used for receiving original image data and synthesizing the original image data and the pure noise data into standard contrast data;

the acquisition unit is used for visually acquiring image data to be detected;

and the comparison unit is used for comparing the image data to be detected with the standard comparison data to obtain a quality inspection judgment result.

Images