A computer vision-based item sorting system

Abstract

The present invention discloses an article sorting system based on computer vision, which specifically relates to the technical field of data processing systems, including an article sorting division module, an article data collection module, an article data analysis module, a comprehensive analysis module, a sorting strategy formulation module, and an automatic sorting module. The present invention determines the target article through the article sorting division module, and marks it according to the way of dividing the single article. The article data collection module collects the size data and pattern data of each article, and the article data analysis module processes and analyzes these data. The comprehensive analysis module compares the data analysis results with the preset sorting rules to formulate the optimal sorting strategy. The automatic sorting module automatically sorts the target article according to the sorting strategy. The present invention can quickly and accurately complete the sorting task of the article, improve production efficiency, reduce manual intervention and reduce costs, and is of great significance for realizing an automated and intelligent production line.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing systems, and more specifically, to an article sorting system based on computer vision. Background Art

[0002] With the rapid development of the economy and the continuous changes in consumption patterns, the logistics and warehousing industries are facing increasing challenges. In order to meet consumers' demand for fast, accurate and efficient logistics services, enterprises need to improve the automation and intelligence level of logistics and warehousing. The sorting system is a key link in the logistics and warehousing process, and its performance and efficiency directly affect the operation of the entire logistics chain. Therefore, the research and improvement of the sorting system is of great significance.

[0003] The existing item sorting system mainly uses conveyor belts as the main transportation tool to transport items from the entrance to the designated classification area. During the transportation process, the system uses sensors and controllers to detect various information of the items in real time, such as size, weight, color, etc., and classifies the items according to preset rules. Once the items arrive at the designated classification area, the conveyor belt will automatically divert them to ensure that each classification area only receives items of the same category.

[0004] However, there are still some shortcomings in its actual use. For example, some sorting systems rely on manual operation and cannot fully realize unmanned operation, resulting in low efficiency and high error rate. Secondly, although some sorting systems use technologies such as machine learning, they still face problems such as large data volume and poor real-time performance in actual applications, making it difficult to meet the needs of large-scale sorting. In addition, existing sorting systems usually require specialized technicians to maintain and manage, which is complicated for ordinary users to operate and has high maintenance costs. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides an article sorting system based on computer vision, which solves the problems raised in the above-mentioned background technology through the following scheme.

[0006] To achieve the above object, the present invention provides the following technical solution: an article sorting system based on computer vision, comprising:

[0007] Sorting item classification module: used to identify the items to be sorted as target items, and mark the target items as 1, 2...n in sequence by classifying them according to individual items;

[0008] Item data collection module: used to collect the size data, pattern data, material data and quality data of each item, and transmit the collected data to the item data analysis module;

[0009] Item data analysis module: including size data analysis unit, pattern data analysis unit, material data analysis unit and quality data analysis unit. Each analysis unit is used to establish a corresponding mathematical model, import the data transmitted by the item data acquisition module into the mathematical model, calculate the size characteristic value, pattern characteristic value, material characteristic value and quality characteristic value of each item, and transmit the analysis results to the comprehensive analysis module;

[0010] Comprehensive analysis module: used to establish a comprehensive analysis model for item information, import the data transmitted by the item data analysis module into the comprehensive analysis model for item information, calculate the comprehensive characteristic value of each item, and transmit the comprehensive characteristic value of each item to the sorting strategy formulation module;

[0011] Sorting strategy formulation module: used to classify and sort items according to their comprehensive feature values, and transmit the sorting information of each item after classification to the automated sorting module;

[0012] Automated sorting module: The items are sorted by the robotic arm on the side of the conveyor belt. When the items are transmitted to the bottom of the robotic arm, the robotic arm sorts the items on the conveyor belt according to the sorting category information of each item transmitted by the sorting strategy formulation module.

[0013] Preferably, the dimension data includes perimeter, area, roundness and rectangularity, which are marked as Sc i , Sa i , Sr i and Se i , the pattern data includes RGB value, primary color ratio, secondary color ratio and color quantity, which are marked as Pc i 、Pm i 、Po i and Pq i , material data includes glossiness, transparency, reflectivity and texture characteristics, marked as Mg i , Mt i ,Ma i and Mf i , quality data includes weight, dynamic characteristics, eccentricity and center of gravity position, which are marked as Qw i , Qd i , Qe i and Qg i , where i=1, 2,...n, i represents the i-th item.

[0014] Preferably, the object data acquisition module continuously photographs the objects by installing a CMOS camera above the conveyor belt to extract the circumference, area, roundness and rectangularity of each object; a laser scanner is installed on the side of the conveyor belt to emit a laser beam and receive the reflected signal to obtain point cloud data of the object surface and extract the RGB value, main color ratio, secondary color ratio and color quantity of each object; a spectrometer is installed on the side of the conveyor belt to measure the gloss, transparency, reflectivity and texture characteristics of the object surface by analyzing the reflected light on the object surface; a pressure sensor is installed under the conveyor belt to collect the weight, dynamic characteristics, eccentricity and center of gravity position of each object.

[0015] Preferably, the dimension data analysis unit is used to establish a dimension data analysis model, which is specifically expressed as: GS i represents the size characteristic value of the i-th item, Sc i represents the perimeter of the i-th item, Sa i represents the area of ​​the i-th item, Sr i Indicates the roundness of the i-th object, Se i represents the rectangularity of the i-th item, and μ1 represents other influencing factors of the size characteristic value of the item.

[0016] Preferably, the pattern data analysis unit is used to establish a pattern data analysis model, which is specifically expressed as: GP i represents the pattern feature value of the i-th item, Pc i represents the RGB value of the i-th item, Pm i represents the main color ratio of the i-th item, Po i represents the secondary color ratio of the i-th item, Pq i represents the number of colors of the i-th object, and μ2 represents other influencing factors of the pattern feature value of the object.

[0017] Preferably, the material data analysis unit is used to establish a material data analysis model, which is specifically expressed as: GM i Represents the material characteristic value of the i-th item, Mg i Indicates the glossiness of the i-th item, Mt i Indicates the transparency of the i-th item, Ma i represents the reflectivity of the i-th object, Mf i represents the texture feature of the i-th item, and μ3 represents other influencing factors of the material feature value of the item.

[0018] Preferably, the quality data analysis unit is used to establish a quality data analysis model, which is specifically expressed as: GQi represents the quality characteristic value of the i-th item, Qw i represents the weight of the i-th item, Qd i represents the dynamic characteristics of the ith item, Qe i represents the eccentricity of the i-th item, Qg i represents the center of gravity of the i-th item, and μ4 represents other influencing factors of the quality characteristic value of the item.

[0019] Preferably, the item information comprehensive analysis model is specifically expressed as: i =GS i ×GP i ×GM i ×GQ i ×λ,η i Represents the comprehensive characteristic value of the i-th item, GS i represents the size characteristic value of the i-th item, GP i represents the pattern feature value of the i-th item, GM i represents the material characteristic value of the i-th item, GQ i represents the quality characteristic value of the i-th item, and λ represents other influencing factors of the comprehensive characteristic value of the item.

[0020] Preferably, the sorting strategy formulation module is used to determine the sorting category of each item, divide the comprehensive feature value of each item into comprehensive feature value thresholds of each item in an equal amount according to the preset classification number, and mark them as 1, 2...m in sequence, and match the comprehensive feature value of each item with each comprehensive feature value threshold and mark them as Where i=1, 2...n, i represents the i-th item, where j=1, 2...m, j represents the j-th sorting category.

[0021] Technical effects and advantages of the present invention:

[0022] The present invention sequentially numbers the items to be sorted through a sorting item classification module, collects size data, pattern data, material data and quality data of each item through an item data collection module, calculates the size characteristic value, pattern characteristic value, material characteristic value and quality characteristic value of each item through an item data analysis module, establishes an item information comprehensive analysis model through a comprehensive analysis module to calculate the comprehensive characteristic value of each item, classifies and sorts the items according to their comprehensive characteristic values ​​through a sorting strategy formulation module, and sorts the items through an automated sorting module;

[0023] The present invention uses computer vision technology, and the system can quickly and accurately identify the characteristics of objects, such as size, pattern, material and quality, so as to formulate the best sorting strategy according to these characteristics, realize automatic sorting, improve sorting efficiency, accurately identify the characteristics of objects, avoid errors caused by human factors, improve sorting accuracy, and formulate personalized sorting strategies to meet the needs of different users. It has high practical value. The sorting system of the present invention can be modularly configured according to actual needs, and can be applied to large-scale production lines as well as small warehouse management. It has strong flexibility, avoids safety hazards caused by human operation, and improves the safety of the sorting process. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic diagram of the overall structure of the present invention. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] refer to Figure 1 The computer vision-based item sorting system shown includes an item sorting classification module, an item data collection module, an item data analysis module, a comprehensive analysis module, a sorting strategy formulation module and an automatic sorting module.

[0027] The sorting item classification module is used to determine the items to be sorted as target items, and mark the target items as 1, 2...n in sequence by classifying them according to individual items.

[0028] The item data collection module is used to collect size data, pattern data, material data and quality data of each item, and transmit the collected data to the item data analysis module.

[0029] The dimension data include perimeter, area, roundness and rectangularity, which are marked as Sc i , Sa i , Sr i and Se i , the pattern data includes RGB value, primary color ratio, secondary color ratio and color quantity, which are marked as Pc i 、Pm i 、Po i and Pq i, material data includes glossiness, transparency, reflectivity and texture characteristics, marked as Mg i , Mt i ,Ma i and Mf i , quality data includes weight, dynamic characteristics, eccentricity and center of gravity position, which are marked as Qw i , Qd i , Qe i and Qg i , where i=1, 2,...n, i represents the i-th item.

[0030] The object data acquisition module continuously photographs the objects by installing a CMOS camera above the conveyor belt to extract the circumference, area, roundness and rectangularity of each object. A laser scanner is installed on the side of the conveyor belt to emit a laser beam and receive the reflected signal to obtain the point cloud data of the object surface and extract the RGB value, main color ratio, secondary color ratio and color quantity of each object. A spectrometer is installed on the side of the conveyor belt to measure the gloss, transparency, reflectivity and texture characteristics of the object surface by analyzing the reflected light on the object surface. A pressure sensor is installed under the conveyor belt to collect the weight, dynamic characteristics, eccentricity and center of gravity position of each object.

[0031] The item data analysis module includes a size data analysis unit, a pattern data analysis unit, a material data analysis unit and a quality data analysis unit. Each analysis unit is used to establish a corresponding mathematical model, import the data transmitted by the item data acquisition module into the mathematical model, calculate the size characteristic value, pattern characteristic value, material characteristic value and quality characteristic value of each item, and transmit the analysis results to the comprehensive analysis module.

[0032] The dimension data analysis unit is used to establish a dimension data analysis model, which is specifically expressed as: GS i represents the size characteristic value of the i-th item, Sc i represents the perimeter of the i-th item, Sa i represents the area of ​​the i-th item, Sr i Indicates the roundness of the i-th object, Se i represents the rectangularity of the i-th item, and μ1 represents other influencing factors of the item's size characteristic value.

[0033] The pattern data analysis unit is used to establish a pattern data analysis model, which is specifically expressed as: GP i represents the pattern feature value of the i-th item, Pc i represents the RGB value of the i-th item, Pm i represents the main color ratio of the i-th item, Poi represents the secondary color ratio of the i-th item, Pq i represents the number of colors of the i-th object, and μ2 represents other influencing factors of the pattern feature value of the object.

[0034] The material data analysis unit is used to establish a material data analysis model, which is specifically expressed as follows: GM i Represents the material characteristic value of the i-th item, Mg i Indicates the glossiness of the i-th item, Mt i Indicates the transparency of the i-th item, Ma i represents the reflectivity of the i-th object, Mf i represents the texture feature of the i-th item, and μ3 represents other influencing factors of the material feature value of the item.

[0035] The quality data analysis unit is used to establish a quality data analysis model, which is specifically expressed as: GQ i represents the quality characteristic value of the i-th item, Qw i represents the weight of the i-th item, Qd i represents the dynamic characteristics of the ith item, Qe i represents the eccentricity of the i-th item, Qg i represents the center of gravity of the i-th item, and μ4 represents other influencing factors of the quality characteristic value of the item.

[0036] The comprehensive analysis module is used to establish a comprehensive analysis model for item information, import the data transmitted by the item data analysis module into the comprehensive analysis model for item information, calculate the comprehensive feature value of each item, and transmit the comprehensive feature value of each item to the sorting strategy formulation module.

[0037] The item information comprehensive analysis model is specifically expressed as: i =GS i ×GP i ×GM i ×GQ i ×λ,η i Represents the comprehensive characteristic value of the i-th item, GS i represents the size characteristic value of the i-th item, GP i represents the pattern feature value of the i-th item, GM i represents the material characteristic value of the i-th item, GQ i represents the quality characteristic value of the i-th item, and λ represents other influencing factors of the comprehensive characteristic value of the item.

[0038] The sorting strategy formulation module is used to classify and sort the items according to their comprehensive feature values, and transmit the sorting information of each item after classification to the automatic sorting module.

[0039] The sorting strategy formulation module is used to determine the sorting category of each item, divide the comprehensive feature value of each item into comprehensive feature value thresholds of each item in equal amounts according to the preset classification number, and mark them as 1, 2...m in sequence, and match the comprehensive feature value of each item with each comprehensive feature value threshold and mark them as n i j , where i=1, 2...n, i represents the i-th item, and j=1, 2...m, j represents the j-th sorting category.

[0040] The automated sorting module sorts the items through the robotic arm on the side of the conveyor belt. When the items are transferred to the bottom of the robotic arm, the robotic arm sorts the items on the conveyor belt according to the sorting category information of each item transmitted by the sorting strategy formulation module.

[0041] The present invention sequentially numbers the items to be sorted through a sorting item classification module, collects size data, pattern data, material data and quality data of each item through an item data collection module, calculates the size characteristic value, pattern characteristic value, material characteristic value and quality characteristic value of each item through an item data analysis module, establishes an item information comprehensive analysis model through a comprehensive analysis module to calculate the comprehensive characteristic value of each item, classifies and sorts the items according to their comprehensive characteristic values ​​through a sorting strategy formulation module, and sorts the items through an automated sorting module.

[0042] The present invention uses computer vision technology, and the system can quickly and accurately identify the characteristics of objects, such as size, pattern, material and quality, so as to formulate the best sorting strategy according to these characteristics, realize automatic sorting, improve sorting efficiency, accurately identify the characteristics of objects, avoid errors caused by human factors, improve sorting accuracy, and formulate personalized sorting strategies to meet the needs of different users. It has high practical value. The sorting system of the present invention can be modularly configured according to actual needs, and can be applied to large-scale production lines as well as small warehouse management. It has strong flexibility, avoids safety hazards caused by human operation, and improves the safety of the sorting process.

[0043] Secondly: In the drawings of the embodiments disclosed in the present invention, only the structures related to the embodiments disclosed in the present invention are involved, and other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of the present invention can be combined with each other;

[0044] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An item sorting system based on computer vision, characterized in that: include: Sorting item classification module: used to identify the items to be sorted as target items, and mark the target items as 1, 2...n in sequence by classifying them according to individual items; Item data collection module: used to collect the size data, pattern data, material data and quality data of each item, and transmit the collected data to the item data analysis module; Item data analysis module: including size data analysis unit, pattern data analysis unit, material data analysis unit and quality data analysis unit. Each analysis unit is used to establish a corresponding mathematical model, import the data transmitted by the item data acquisition module into the mathematical model, calculate the size characteristic value, pattern characteristic value, material characteristic value and quality characteristic value of each item, and transmit the analysis results to the comprehensive analysis module; The dimension data analysis unit is used to establish a dimension data analysis model, which is specifically expressed as: GS i represents the size characteristic value of the i-th item, Sc i represents the perimeter of the i-th item, Sa i represents the area of ​​the i-th item, Sr i Indicates the roundness of the i-th object, Se i represents the rectangularity of the i-th item, μ1 represents other influencing factors of the item's size characteristic value; The pattern data analysis unit is used to establish a pattern data analysis model, which is specifically expressed as: GP i represents the pattern feature value of the i-th item, Pc i represents the RGB value of the i-th item, Pm i represents the main color ratio of the i-th item, Po i represents the secondary color ratio of the i-th item, Pq i represents the number of colors of the i-th object, μ2 represents other influencing factors of the pattern feature value of the object; The material data analysis unit is used to establish a material data analysis model, which is specifically expressed as follows: GM i Represents the material characteristic value of the i-th item, Mg i Indicates the glossiness of the i-th item, Mt i Indicates the transparency of the i-th item, Ma i represents the reflectivity of the i-th object, Mf i represents the texture feature of the i-th item, μ3 represents other influencing factors of the material feature value of the item; The quality data analysis unit is used to establish a quality data analysis model, which is specifically expressed as: GQ i represents the quality characteristic value of the i-th item, Qw i represents the weight of the i-th item, Qd i represents the dynamic characteristics of the ith item, Qe i represents the eccentricity of the i-th item, Qg i represents the center of gravity of the i-th item, μ4 represents other influencing factors of the quality characteristic value of the item; Comprehensive analysis module: used to establish a comprehensive analysis model for item information, import the data transmitted by the item data analysis module into the comprehensive analysis model for item information, calculate the comprehensive characteristic value of each item, and transmit the comprehensive characteristic value of each item to the sorting strategy formulation module; The item information comprehensive analysis model is specifically expressed as: i =GS i ×GP i ×GM i ×GQ i ×λ,η i Represents the comprehensive characteristic value of the i-th item, GS i represents the size characteristic value of the i-th item, GP i represents the pattern feature value of the i-th item, GM i represents the material characteristic value of the i-th item, GQ i represents the quality characteristic value of the i-th item, and λ represents other influencing factors of the comprehensive characteristic value of the item; Sorting strategy formulation module: used to classify and sort items according to their comprehensive feature values, and transmit the sorting information of each item after classification to the automated sorting module; Automated sorting module: The items are sorted by the robotic arm on the side of the conveyor belt. When the items are transmitted to the bottom of the robotic arm, the robotic arm sorts the items on the conveyor belt according to the sorting category information of each item transmitted by the sorting strategy formulation module.

2. The computer vision-based item sorting system according to claim 1, characterized in that: The dimension data include perimeter, area, roundness and rectangularity, which are marked as Sc i , Sa i , Sr i and Se i , the pattern data includes RGB value, primary color ratio, secondary color ratio and color quantity, which are marked as Pc i 、Pm i 、Po i and Pq i , material data includes glossiness, transparency, reflectivity and texture characteristics, marked as Mg i , Mt i ,Ma i and Mf i , quality data includes weight, dynamic characteristics, eccentricity and center of gravity position, which are marked as Qw i , Qd i , Qe i and Qg i , where i=1, 2,...n, i represents the i-th item.

3. The computer vision-based item sorting system according to claim 1, characterized in that: The object data acquisition module continuously photographs the objects by installing a CMOS camera above the conveyor belt to extract the circumference, area, roundness and rectangularity of each object. A laser scanner is installed on the side of the conveyor belt to emit a laser beam and receive the reflected signal to obtain the point cloud data of the object surface and extract the RGB value, main color ratio, secondary color ratio and color quantity of each object. A spectrometer is installed on the side of the conveyor belt to measure the gloss, transparency, reflectivity and texture characteristics of the object surface by analyzing the reflected light on the object surface. A pressure sensor is installed under the conveyor belt to collect the weight, dynamic characteristics, eccentricity and center of gravity position of each object.

4. The computer vision-based item sorting system according to claim 1, characterized in that: The sorting strategy formulation module is used to determine the sorting category of each item, divide the comprehensive feature value of each item into comprehensive feature value thresholds of each item in equal amounts according to the preset classification number, and mark them as 1, 2...m in sequence, and match the comprehensive feature value of each item with each comprehensive feature value threshold and mark them as Where i=1, 2...n, i represents the i-th item, where j=1, 2...m, j represents the j-th sorting category.

Images