Recycled aggregate self-adaptive crushing system based on machine vision and preparation method

Abstract

A recycled aggregate self-adaptive crushing system based on machine vision relates to the technical field of solid waste resourceful treatment equipment and comprises a feeding and crushing module, a machine vision module, a central control and decision module and a multi-stage aggregate distribution and collection module. The feeding and crushing module sends out construction waste in a single-particle or thin-layer flow mode and conducts energy-controllable impact crushing. The machine vision module captures an aggregate image and identifies particle size distribution in real time; the central control module compares the real-time granularity with a preset target gradation, and dynamically adjusts crushing parameters; and the shunting and collecting module distributes aggregates with different particle sizes through high-pressure gas. The preparation method comprises the following steps: inputting target gradation, feeding raw material single particles for crushing, analyzing the particle size in real time by machine vision, adjusting the crushing particle size by a central module as required, and shunting to corresponding channels according to the particle size to realize continuous production. The technical problems that in the prior art, product grading fluctuation is large, the process is long, efficiency is low, energy consumption is high, accurate grading cannot be achieved, and recycling efficiency is low can be solved.

Description

Technical Field

[0001] This invention relates to the field of equipment technology for the resource utilization of solid waste, specifically to an adaptive crushing system and preparation method for recycled aggregate based on machine vision. Background Technology

[0002] Currently, the production of recycled coarse aggregate from construction waste mainly relies on traditional mining crushing and screening equipment, such as jaw crushers and impact crushers, combined with multi-layer vibrating screens. These traditional methods have the following inherent drawbacks: 1. The process is extensive and lacks controllability. The parameters of the crusher (such as speed and chamber shape) are usually fixed and cannot be dynamically adjusted according to the characteristics of the raw materials or product requirements, resulting in large fluctuations in product gradation and unstable quality. 2. The passive "crush first, then screen" model means that the aggregate undergoes a complete crushing process before being graded by a vibrating screen. For scenarios requiring multiple gradations, the crushed mixture must be screened multiple times, resulting in a long process, low efficiency, high energy consumption, and the potential for dust generation and secondary crushing. 3. Precise grading is impossible. Vibrating screens are based on physical aperture size, which is ineffective for separating flaky and needle-shaped aggregates and cannot achieve precise separation of discontinuous particle size ranges (such as only taking 5-10mm and 20-25mm). 4. Low resource utilization efficiency: For the same batch of raw materials, it is difficult to efficiently obtain a variety of high-quality recycled aggregates of different specifications in a single processing. Summary of the Invention

[0003] The purpose of this invention is to provide a machine vision-based adaptive crushing system and preparation method for recycled aggregates, which can solve the technical problems of large product gradation fluctuations, long process / low efficiency / high energy consumption, inability to achieve accurate grading, and low resource utilization efficiency in the existing technology.

[0004] To achieve the above objectives, the present invention adopts the following technical solution.

[0005] An adaptive crushing system for recycled aggregates based on machine vision includes a feeding and crushing module, a machine vision module, a central control and decision-making module, and a multi-stage aggregate diversion and collection module.

[0006] The feeding and crushing module can deliver construction waste raw materials in the form of single particles or thin laminar flow and perform energy-controlled impact crushing.

[0007] The machine vision module can capture images of crushed aggregate falling and identify and statistically analyze the particle size distribution of aggregate in real time.

[0008] The central control and decision-making module has a built-in target gradation database. It can receive the particle size recognition results generated by the machine vision module and compare them with the preset target gradation. Based on the comparison results, it dynamically adjusts the working parameters of the crushing module so that the particle size distribution of the crushed aggregate meets the preset target.

[0009] The multi-stage aggregate diversion and collection module can use high-pressure gas to blow aggregates of different particle size ranges to the corresponding positions based on the real-time particle size recognition results of a single aggregate or aggregate group by the machine vision module.

[0010] Furthermore, the feeding and crushing module includes an electromagnetic vibrating feeder and an impact crusher. The vibration frequency and amplitude of the electromagnetic vibrating feeder can be adjusted by the central control and decision module. The impact crusher is a vertical shaft impact crusher, and the rotor speed of the impact crusher can be steplessly adjusted by the central control and decision module.

[0011] Furthermore, the machine vision module includes an imaging dark box, which contains a high-speed camera and a light source. The high-speed camera is used to capture images of the aggregate falling freely after crushing, and by running an embedded processor based on a deep learning-based semantic segmentation model, it identifies the projected area and equivalent particle size of each aggregate in the image in real time.

[0012] Furthermore, the camera is a linear array or area array camera, and the light source is a high-brightness LED light source.

[0013] Furthermore, the central control and decision-making module can preset multiple target gradation curves, generate corresponding crushing control logic for each target gradation through a single crushing process, and drive the multi-stage aggregate diversion and collection module to blow the generated aggregates to the corresponding positions according to their particle size.

[0014] Furthermore, the multi-stage aggregate diversion and collection module includes a diverter located at the end of the aggregate falling path. The diverter includes multiple guide channels corresponding to different particle size ranges. Each guide channel has a high-pressure airflow nozzle at its inlet. The airflow nozzle can be controlled by the central control and decision module to achieve high-frequency opening and closing, and is used to blow the aggregate into the corresponding guide channel according to the particle size recognition result of the machine vision module.

[0015] Furthermore, the diverter includes a return channel. The multi-stage aggregate diversion and collection module can send oversized aggregates exceeding the maximum particle size range corresponding to all guide channels back to the crushing module for cyclic crushing until the particle size range requirements are met.

[0016] Furthermore, the multi-stage aggregate diversion and collection module includes multiple collection bins, the number of which corresponds to the number of guide channels and is used to collect aggregates from the corresponding guide channels and return channels.

[0017] A machine vision-based method for preparing recycled aggregate, based on the above-mentioned crushing system, includes the following steps: S1. Input one or more target gradations of recycled coarse aggregates into the central control and decision-making module; S2. The system starts up. Construction waste raw materials are fed into the crushing module in the form of single particles through the feeding module for preliminary crushing. S3. During the free fall of the crushed aggregate, the machine vision module captures images and analyzes its particle size distribution in real time. S4. The central control and decision-making module compares the real-time particle size distribution with the target gradation. If the real-time average particle size is too large, the crushing module is controlled to reduce the particle size of the produced aggregate. If the real-time average particle size is too small, the crushing module is controlled to increase the particle size of the produced aggregate. S5. For falling aggregates or aggregate groups, based on the particle size data identified by the machine vision module, the multi-stage aggregate diversion and collection module generates a corresponding high-pressure airflow to blow the falling aggregates or aggregate groups into the guide channel or return channel that matches the target gradation. S6. Continue with steps S2 to S5 to achieve continuous operation.

[0018] By adopting the above technical solution, the present invention has the following beneficial effects: 1. This invention, through real-time feedback and control, can accurately control the crushed products within the target particle size range, significantly improving the pass rate and stability of product gradation; 2. This invention can produce multiple recycled aggregates with different gradations in one go and in parallel from the same batch of feed, which greatly improves production flexibility and resource utilization efficiency. 3. This invention can separate qualified aggregates in a timely manner, avoiding "over-crushing" and "secondary crushing" in the crushing chamber, reducing micro-cracks and fine powder content, and helping to improve the quality of recycled aggregates; 4. The system of the present invention has learning capabilities and can automatically find and maintain the optimal crushing parameters according to the construction waste raw materials of different sources and compositions, thereby reducing the dependence on the experience of operators. 5. The enclosed visual inspection and pneumatic diversion process in this invention generates almost no dust, making it more environmentally friendly than vibrating screening. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the overall structure of the crushing system in this invention.

[0020] Figure 2 This is a schematic diagram of the preparation method in this invention.

[0021] Figure descriptions: 11. Electromagnetic vibrating feeder, 12. Impact crusher, 13. Raw material bin, 21. Imaging dark box, 22. Camera, 23. Light source, 3. Central control and decision-making module, 41. Diverter, 42. Guide channel, 43. Airflow nozzle, 44. Return channel, 45. Collection bin, 46. Return conveyor. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the present invention clearer, the features and performance of a machine vision-based adaptive crushing system and preparation method for recycled aggregates will be further described in detail below with reference to the accompanying drawings and embodiments.

[0023] Example 1 Please see the appendix Figures 1-2 A machine vision-based adaptive crushing system for recycled aggregates includes a feeding and crushing module, a machine vision module, a central control and decision-making module, and a multi-stage aggregate diversion and collection module.

[0024] The feeding and crushing module can feed construction waste raw materials in the form of single particles or thin laminar flow and perform energy-controlled impact crushing. Preferably, in this embodiment, the construction waste raw materials are fed in the form of thin laminar flow. Specifically, the feeding and crushing module includes an electromagnetic vibrating feeder 11 and an impact crusher 12. The vibration frequency and amplitude of the electromagnetic vibrating feeder 11 can be adjusted by the central control and decision module 3. The impact crusher 12 is a vertical shaft impact crusher, and the rotor speed of the impact crusher 12 can be steplessly adjusted by the central control and decision module 3.

[0025] The machine vision module can capture images of crushed aggregate falling freely and identify and statistically analyze the particle size distribution of the aggregate in real time. Specifically, the machine vision module includes an imaging dark box 21, which contains a high-speed camera 22 and a light source 23. The camera 22 is a linear array or area array camera, and the light source 23 is a high-brightness LED light source. The camera 22 is used to capture images of the crushed aggregate falling freely, and by running an embedded processor based on a deep learning-based semantic segmentation model, it identifies the projected area and equivalent particle size of each aggregate in the image in real time.

[0026] The central control and decision-making module 3 has a built-in target gradation database. It can receive particle size recognition results generated by the machine vision module and compare them with preset target gradations. Based on the comparison results, it dynamically adjusts the working parameters of the crushing module to ensure that the particle size distribution of the crushed aggregate meets the preset target. The central control and decision-making module 3 can preset multiple target gradation curves. Through a single crushing process, it generates corresponding crushing control logic for each target gradation and drives the multi-stage aggregate diversion and collection module to blow the generated aggregate to the corresponding positions according to particle size.

[0027] The multi-stage aggregate diversion and collection module, based on the real-time particle size recognition results of a single aggregate or aggregate group by a machine vision module, uses high-pressure gas to blow aggregates of different particle size ranges to their corresponding positions. Specifically, the multi-stage aggregate diversion and collection module includes a diverter 41, located at the end of the aggregate falling path. The diverter 41 includes multiple guide channels 42 corresponding to different particle size ranges. Each guide channel 42 has a high-pressure airflow nozzle 43 at its inlet. The airflow nozzle 43 can be controlled by the central control and decision module 3 to achieve high-frequency opening and closing, used to blow aggregates into the corresponding guide channel 42 according to the particle size recognition results of the machine vision module. The diverter 41 also includes a return channel 44. The multi-stage aggregate diversion and collection module can send oversized aggregates exceeding the maximum particle size range corresponding to all guide channels 42 back to the crushing module for cyclic crushing until the particle size range requirements are met. The multi-stage aggregate diversion and collection module also includes multiple collection bins 45. The number of collection bins 45 is consistent with the number of guide channels 42 and corresponds one-to-one. They are used to collect aggregates in the corresponding guide channels 42 and return channels 44.

[0028] A machine vision-based method for preparing recycled aggregate, based on the aforementioned crushing system, includes the following steps.

[0029] S1. Input one or more target gradations of recycled coarse aggregates into the central control and decision-making module 3.

[0030] S2. The system starts up. Construction waste raw materials are fed into the crushing module in a thin laminar flow form through the feeding module for preliminary crushing.

[0031] S3. During the free fall of the crushed aggregate, the machine vision module captures images and analyzes its particle size distribution in real time.

[0032] S4. The central control and decision-making module 3 compares the real-time particle size distribution with the target gradation. If the real-time average particle size is too large, the crushing module is controlled to reduce the particle size of the produced aggregate. If the real-time average particle size is too small, the crushing module is controlled to increase the particle size of the produced aggregate.

[0033] S5. For falling aggregates or aggregate groups, based on the particle size data identified by the machine vision module, the multi-stage aggregate diversion and collection module generates a corresponding high-pressure airflow to blow the falling aggregates or aggregate groups into the guide channel 42 or return channel 44 that matches the target gradation.

[0034] S6. Continue with steps S2 to S5 to achieve continuous operation.

[0035] In practical implementation, the core concept of this invention is to construct an integrated intelligent closed-loop system of "perception-decision-execution". The system uses crushed aggregate as an "information carrier", instantly acquires its particle size information through machine vision, and compares this information with the target value by the central control and decision-making module 3 to make a decision, thereby adjusting the crushing intensity and directing the diversion mechanism to complete the precise classification and collection of aggregate.

[0036] like Figure 1 As shown, the feeding and crushing module includes an electromagnetic vibrating feeder 11 and a vertical shaft impact crusher 12. The feeder sends the raw materials into the crusher in the form of a thin layer of material flow. The raw material bin 13 is used to store the construction waste raw materials to be processed.

[0037] The machine vision module includes a closed imaging chamber 21, inside which is a high-speed camera 22 and a linear light source 23. The crushed aggregate falls freely within the field of view.

[0038] The central control and decision-making module 3 includes an industrial computer and integrated control software, which can display real-time images, granular distribution curves, and control parameters.

[0039] The multi-stage aggregate diversion and collection module includes a diverter 41 and collection bins 45. The diverter 41 includes three guide channels 42 corresponding to different particle size ranges. The diverter 41 also includes a return channel 44. Four high-pressure airflow nozzles 43 are arranged around the inlets of the guide channels 42 and the return channel 44 to blow the aggregate into the guide channels 42 or the return channel 44 of the corresponding size. Three collection bins 45 are connected below the diverter 41. The three collection bins 45 correspond to the three guide channels 42 respectively and are used to collect finished products with different gradations.

[0040] The multi-stage aggregate diversion and collection module also includes a return conveyor, which is used to send the aggregate discharged from the return channel 44 back to the raw material silo 13.

[0041] In practical operation, the user first sets two target gradations on the software interface of the central control and decision-making module 3: gradation A (particle size 5-10 mm) and gradation B (particle size 10-20 mm). After the system starts, construction waste raw materials are fed from the raw material bin 13 into a uniform thin layer via the electromagnetic vibrating feeder 11 and then into the vertical shaft impact crusher 12 for primary crushing. The rotational speed of the primary crushing is set to a conservative initial value.

[0042] The crushed aggregate is discharged from the crusher outlet and falls freely inside the imaging dark box 21. A high-speed camera 22, in conjunction with a linear light source 23, captures clear images of the aggregate. These images are transmitted in real-time to the central control and decision-making module 3, whose built-in AI particle size analysis algorithm immediately calculates the particle size distribution of the current batch. For example: particles > 20 mm account for 30%, particles 10–20 mm account for 40%, particles 5–10 mm account for 25%, and particles < 5 mm account for 5%.

[0043] The central control and decision-making module 3 compares the real-time particle size distribution of the aggregate with the target gradations A and B. If the logic determines that the proportion of "extra-large particles" (>20mm) is too high, a control command is generated to appropriately increase the rotor speed of the vertical shaft impact crusher 12 to enhance the crushing effect. If the logic determines that the proportion of "ultra-fine particles" (<5mm) is too high, a control command is generated to appropriately decrease the rotor speed of the vertical shaft impact crusher 12 to weaken the crushing effect.

[0044] Simultaneously, for each falling aggregate, the AI ​​model identifies its equivalent particle size in real time. When an aggregate is identified as belonging to gradation A, the control system immediately triggers the high-pressure airflow nozzle 43 of the corresponding guide channel 42 with a particle size of 5-10 mm. A precise pulse of airflow blows the aggregate into the corresponding guide channel 42 and then into the corresponding collection bin 45. Similarly, aggregates identified as gradation B are blown into the corresponding guide channel 42 with a particle size of 10-20 mm and then into the corresponding collection bin 45. Aggregates identified as "oversized" (particle size > 20 mm) are blown into the return channel 44 and sent back for re-crushing by the return conveyor 46. Fine particles with a particle size < 5 mm enter the corresponding guide channel 42 with a particle size < 5 mm and then fall into the corresponding collection bin 45.

[0045] This process continues in a loop. The speed of the crusher is dynamically adjusted according to the real-time monitoring of the "oversized particles" ratio until the system reaches a stable state, that is, to continuously produce qualified A and B graded recycled coarse aggregates with the highest efficiency.

[0046] Example 2 Unlike Embodiment 1, the impact crusher in this embodiment is a hammer crusher, and the energy of the impact hammer of the impact crusher can be steplessly adjusted by the central control and decision-making module.

[0047] It should be noted that the parts not described in detail in this solution are all prior art. The above embodiments are only used to illustrate the present invention, but the present invention is not limited to the above embodiments. Any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. A machine vision based adaptive crushing system for recycled aggregates, characterized by: It includes a feeding and crushing module, a machine vision module, a central control and decision-making module (3), and a multi-stage aggregate diversion and collection module. The feeding and crushing module can deliver construction waste raw materials in the form of single particles or thin laminar flow, and perform energy-controlled impact crushing. The machine vision module can capture images of crushed aggregate falling to the ground, and identify and analyze the particle size distribution of the aggregate in real time. The central control and decision-making module (3) has a built-in target gradation database. It can receive the particle size recognition results generated by the machine vision module and compare them with the preset target gradation. Based on the comparison results, it dynamically adjusts the working parameters of the crushing module so that the particle size distribution of the crushed aggregate meets the preset target. The multi-stage aggregate diversion and collection module can use high-pressure gas to blow aggregates of different particle size ranges to the corresponding positions based on the real-time particle size recognition results of a single aggregate or aggregate group by the machine vision module.

2. A machine vision based recycled aggregate self-adaptive crushing system as claimed in claim 1, wherein: The feeding and crushing module includes an electromagnetic vibrating feeder (11) and an impact crusher (12). The vibration frequency and amplitude of the electromagnetic vibrating feeder (11) can be adjusted by the central control and decision module (3). The impact crusher (12) is a vertical shaft impact crusher. The rotor speed of the impact crusher (12) can be steplessly adjusted by the central control and decision module (3).

3. A machine vision based recycled aggregate self-adaptive crushing system as claimed in claim 1, wherein: The machine vision module includes an imaging dark box (21), which is equipped with a high-speed camera (22) and a light source (23). The high-speed camera (22) is used to capture images of the aggregate falling freely after crushing, and by running an embedded processor based on a deep learning-based semantic segmentation model, it identifies the projected area and equivalent particle size of each aggregate in the image in real time.

4. A machine vision based recycled aggregate self-adaptive crushing system as claimed in claim 3, wherein: The camera (22) is a linear array or area array camera, and the light source (23) is a high-brightness LED light source.

5. A machine vision based recycled aggregate adaptive crushing system as claimed in claim 1, wherein: The central control and decision-making module (3) can preset multiple target gradation curves, generate corresponding crushing control logic for each target gradation through a single crushing process, and drive the multi-stage aggregate diversion and collection module to blow the generated aggregates to the corresponding positions according to their particle size.

6. The adaptive crushing system for recycled aggregate based on machine vision as described in claim 1, characterized in that: The multi-stage aggregate diversion and collection module includes a diverter (41), which is located at the end of the aggregate falling path. The diverter (41) includes multiple guide channels (42) corresponding to different particle size ranges. Each guide channel (42) has a high-pressure airflow nozzle (43) at its inlet. The airflow nozzle (43) can be controlled by the central control and decision module (3) to achieve high-frequency opening and closing, and is used to blow the aggregate into the corresponding guide channel (42) according to the particle size recognition result of the machine vision module.

7. The adaptive crushing system for recycled aggregate based on machine vision as described in claim 6, characterized in that: The diverter (41) includes a return channel (44). The multi-stage aggregate diversion and collection module can send ultra-large aggregates that exceed the maximum particle size range corresponding to all guide channels (42) back to the crushing module for cyclic crushing through the return channel (44) until the particle size range requirement is met.

8. The machine vision-based adaptive crushing system for recycled aggregates as described in claim 7, characterized in that: The multi-stage aggregate diversion and collection module includes multiple collection bins (45). The number of collection bins (45) is consistent with the number of guide channels (42) and corresponds one-to-one. They are used to collect aggregates in the corresponding guide channels (42) and return channels (44).

9. A method for preparing recycled aggregate based on machine vision, based on the crushing system as described in claims 1-8, characterized in that: Includes the following steps, S1. Input one or more target gradations of recycled coarse aggregates into the central control and decision-making module (3); S2. The system starts up. Construction waste raw materials are fed into the crushing module in the form of single particles through the feeding module for preliminary crushing. S3. During the free fall of the crushed aggregate, the machine vision module captures images and analyzes its particle size distribution in real time. S4. The central control and decision-making module (3) compares the real-time particle size distribution with the target gradation. If the real-time average particle size is too large, the crushing module is controlled to reduce the particle size of the produced aggregate. If the real-time average particle size is too small, the crushing module is controlled to increase the particle size of the produced aggregate. S5. For the falling aggregate or aggregate group, based on the particle size data identified by the machine vision module, the falling aggregate or aggregate group is blown into the guide channel (42) or return channel (44) that matches the target gradation by controlling the multi-stage aggregate diversion and collection module to generate the corresponding high-pressure airflow. S6. Continue with steps S2 to S5 to achieve continuous operation.

Images