OA20142A - Device for sorting physical defects on objects. - Google Patents

Abstract

The present inspection or sorting device (also called intelligent sieve) consists of: - A module making it possible to control the movement of the conveyor belt; - An object image acquisition chamber (containing an on-board camera, an appropriate lighting system depending on the targeted application); - A module allowing the rotation of objects under the conveyor belt; - A module allowing their analysis (Computer, Laptop or FPGA card); - A module allowing the ejection of objects. It allows at lower cost, in a completely modular way, by replacing one or more electronic modules while modifying the programs (of the FPGA card, those of the microcontroller), to perform one or more tasks simultaneously. Equipped with a unit for real-time data transmission by wifi, it automates the sorting of objects.

Description

DEVICE FOR SORTING PHYSICAL FAULTS ON OBJECTS (Smart sieve)

FIELD OF THE INVENTION TO WHICH THE INVENTION RELATES

The present invention relates to a device for inspecting and / or sorting by artificial vision of quality defects on objects (using physical or chemical characteristics) in several fields (food, medicine, mining, etc.). It can be sorting of Penja pepper, rice, beans, soybeans, corn, dates, cherries, tablets, metals, etc.

PRIOR ART

The work of "hal-00466401" presents an artificial vision system allowing us to discriminate between seeds falling in free fall from their digital images. However, the very rapid movement of the object in front of the cameras prevents the capture and analysis of certain grains and therefore reduces the performance of the device. Likewise the performance of the device is based on the morphology which can simply be regulated by a mechanical device. In addition, the synchronization between the scrolling of the grains and the processing speed of the algorithms is not reassured. Finally, the number of cameras used (three) to analyze a single seed increases the costs of the device and especially if we want to perform a simultaneous analysis of several seeds in free fall scrolling with several colors / textures.

The work of "WO / 2008/150903" presents a seed sorter which uses a turntable on which there are holes which can accommodate the seeds in blocks of 24 and whose purpose is to carry the seeds to convey them to two cameras which analyze their morphology. Defective seeds are then ejected by a pneumatic system. The synchronization between the movement of the grains and the ejection speed is not always reassured and the cost of the device is very high due to the number of cameras used.

The work of "WO / 2020/116096" presents an optical sorter for several types of objects. This optical sorter is provided with a plurality of optical units for sorting, by quality, moving objects. A first optical sorting unit transfers poor quality objects to a second optical sorting unit using a discharge medium. The device developed is a very bulky mechanism and increases its production cost.

The works of "WO / 2019/155764" present an invention the aim of which is to provide an optical sorter with which the positional precision of a constituent element fixed to a frame can be simply calculated without any welding distortion occurring. produce in the frame. This device is for exclusive application in object welding.

The work of "WO / 2002/074457" presents a device and method for the automatic inspection of objects moving in an essentially single-layer flow. The device cannot allow the analysis of the different surfaces of the object.

The aforementioned previous work does not make it possible to combine efficiency and cost reduction in the inspection or sorting of quality defects on objects.

BRIEF DESCRIPTION OF THE INVENTION

The object of the invention is to inspect, sift or sort by an intelligent system objects of different sizes passing on a conveyor belt and having quality defects. Advantageously, the device first comprises an on-board camera which captures the images of the objects in a manner synchronized with the conveying mechanism so that the speed of the algorithms dedicated to the analysis of the grains is adjusted with the ejection device. The objects are arranged in rows and columns so that there is space between them and that several can be collected in a camera image capture. The number of rows and columns (or the number of objects to analyze in an image capture) depends on the resolution of the camera. The slope imposed on the conveyor belt before the image acquisition chamber keeps only the objects to be analyzed in the cell where they are housed by bringing other foreign objects below. Then a capture is actuated in the image acquisition chamber through an object detector ("WO / 1998/036393") or a simple synchronization (between the image acquisition sequence and the speed of the conveyor) . Finally, the inspection and / or sorting method (“EP / A1 / 0054598”) uses the algorithms developed (for segmentation, characterization and classification of images) which make it possible to extract the objects from the scene (by isolating or by eliminating the background), to characterize them through their physical properties (texture and color) while allowing them to be discriminated (by supervised or unsupervised learning). Segmentation algorithms (thresholding, fuzzy logic, etc.) such as characterization (wavelets, histograms, etc.) and discrimination (wide-margin separators, neural networks, etc.) depend on the complexity of the digital image obtained.

Thus, by way of non-limiting example, the device can analyze, with different meshes of the conveyor belt, the quality defects on the beans, on the grains of rice, on the

Penja pepper. The same process, without limitation, is applicable to sorts of several other objects of different sizes (dates, gold, etc.).

According to the different characteristics of the invention, the device comprises: a camera, a reducer motor or 220V stepper motor (the voltage here is not limiting), the speed of which is controlled by an electronic card, a belt conveyor, lighting lamps, Arduinos boards, an FPGA board (or Laptop or computer), 5V stepper motors, ejection pipes.

The advantages are numerous:

- Reduced device costs by the ability to analyze multiple objects in a single image sequence;

- Better control of the scrolling of an object and its ejection;

- Possibility of multi-classifying objects (using supervised learning [neural networks, fuzzy algorithms, wide-margin separators, etc.] or unsupervised) with reduction of calculation time by eliminating calculations linked to morphological parameters (taking into account the texture and color of objects);

- Increased precision (approximately 10 times more precise than conventional devices).

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will be better understood by studying a particular embodiment taken by way of non-limiting example and illustrated by the appended drawings, in which:

- Floor 1/4 presents an embodiment of the device (figure 1)

- Floor 2/4 presents a conveyor belt model respectively for the transport of rice (figure 2), the transport of Penja pepper (figure 3) and the transport of beans (figure 4).

- Floor 3/4 presents image processing or image analysis for decision making (figures 5 and 6)

- Floor 4/4 presents the device for rotating the objects (figure 7) below the conveyor belt and allowing them to be deposited in the ejection chamber

DETAILED DESCRIPTION OF THE INVENTION

Referring to Figure 1, the device according to the invention comprises a reduction motor or a stepping motor (8), a conveyor belt (7), carrying the object to be analyzed (6), an electronic card of control of the conveyor (5), a camera (4), an acquisition chamber for the images to be analyzed (3), a card actuating the rotation of the objects (2) and an object ejection chamber (1). The laptop or laptop computer (Figure 1 (9)) is used for image analysis.

The reduction motor or stepping motor (figure 1 (8)), allows to control the movement of the conveyor belt. The speed of the stepper motor is set from an electronic board (figure 1 (5)) which contains a microcontroller / microprocessor.

The acquisition chamber (Figure 1 (3)) is equipped with an illumination system which facilitates the capture of images of the objects.

The camera (Figure 1 (4)) above the acquisition chamber captures images of objects on the moving conveyor belt at a determined frequency.

Images (Figures 2-4) of the objects are sent to a computer or FPGA board for analysis (Figures 5 and 6) and classification. The device can be used for object inspection (removing all the ejection device). In this sense, it provides the statistics of the defective objects contained in the packaging without, however, sorting them out.

The device can be used for sorting objects. In this sense, an electronic card (containing a microcontroller) placed below the conveyor belt allows the objects to be rotated (figure 1 (2)) in order to film the two facets of the objects. Two 5V motors are used for the rotation and translation of said device. An iron bar installed on the first sliding device (which supports the rack) prevents objects from falling (Figure 7 (23)) as they rotate (Figure 7 (25)). The rotation of objects (Figure 7 (25)) can be simply actuated through the translational movement of the rack (Figure 7 (24)) which rotates the toothless wheels or directly through the rotation of the toothless wheels (Figure 7 ( 25)). A 5V motor rotates the toothless wheel directly (or rotates it through a vice-nut system which drives the translational movement of the rack-and-pinion system). Another 5 V motor translates the sliding device. This sliding device frees objects that will fall into the ejection chamber provided with the outlet openings.

The conveyor belt (figure 2 (10)) can contain several types of grains: grains with serious defects (figure 5 (16)), slight defects (figure 5 (15)) or conforming grains (figure 5 ( 17)). Indeed, when the two faces of the grains are analyzed (Figure 6), the system releases the grains at the level of the ejection chamber by a combined translational movement of the wheel-rack system and the sliding device. Another 5V motor below the bent pipes of the ejection chamber allows the grain to be directed towards the outlet of the specific pipes according to their physical / qualitative parameter (texture, color).

INDUSTRIAL APPLICATION

The device according to the invention is particularly intended for quality control in the packaging of objects.

By way of example and without limitation:

- inspection of quality defects on food grains by artificial vision (which meets the needs of the food industries which must comply with the marketing standards laid down by the Codex Alimentarius). This may involve approving the quality of the packaging (presence of quality defects) of rice grains (figure 2 (11)), white peppers (figure 3 (13)), black peppers (figure 3 (12) ), beans (Figure 4 (14)).

- inspection of the quality of drugs by artificial vision. This may involve approving the quality before packaging (presence of quality defects) on all types of tablets.

- sorting of quality defects on food grains and / or seeds by artificial vision (which meets the needs of the food industries which must comply with the marketing standards laid down by the Codex Alimentarius). This can be the sorting of quality defects on rice grains (figure 2 (11)), white peppers (figure 3 (13)), black peppers (figure 3 (12)), beans ( figure 4 (14)).

- sorting of quality defects on metals in mining industries. It can be gold, iron, etc. The principle of conveyor belt delivery remains the same. The type of conveyor belt varies depending on the material to be analyzed.

2?

Claims (10)

1- Inspection and / or sorting device for physical defects on objects (allowing inspection, sifting or sorting by an intelligent system of objects of different sizes), composed of a conveyor belt (figure 1 (7 )) which allows the objects to scroll, a reduction motor or a stepping motor (figure 1 (8)) which drives the conveyor belt carrying the object to be analyzed (figure 1 (6)), a card conveyor control electronics (figure 1 (5)), a camera (figure 1 (4)), an acquisition chamber of the images to be analyzed (figure 1 (3)), a card activating the rotation of the objects (figure 1 (2)) and an object ejection chamber (Figure 1 (1)). 2- Device, which can analyze and / or sort the objects according to claim 1, in the fields of seeds to seeds, food, mining and medical. 3- Device for inspecting physical defects on objects according to claim 1, characterized in that it provides the statistics of the defective objects contained in the packaging. 4- Device for sorting physical defects on objects according to claim 1, characterized in that it proceeds by an electronic card (containing a microcontroller) placed below the conveyor belt to rotate the objects (Figure 1 (2)) in order to film the two facets of the objects (grains, drugs, etc.). 5- Device for sorting physical defects on objects according to claims 1 and 4, comprising an iron bar installed on the first sliding device (which supports the rack), said iron bar prevents the objects from falling (figure 7 (23 )) during their rotation (Figure 7 (25)). 6- Device for inspecting and / or sorting physical defects on the objects according to claims 1 and 2, where said objects can be Penja pepper, rice, beans, soybeans, corn, dates, cherries, tablets or metals. 7- Device for inspecting physical defects on objects according to claim 1, characterized in that it allows the conveying of objects in a "pipeline / checkerboard" arrangement; 8- Device for inspecting and sorting physical defects on objects according to claims 1 and 2, which allows the ejection of objects through the rotation and / or the controlled translation of the bent pipes (or materials having similar arrangements) in order to lead the analyzed objects to exit orifices; 9. Device for inspecting physical defects on objects according to claim 1, provided with a method for analyzing images containing objects arranged in a pipeline / checkerboard for sorting; 10-Device for inspecting physical defects on objects according to claim 1, which allows the capture of images (of objects arranged in a pipeline) in a chamber illuminated by LED bulbs (single LED, UVC LED, etc.) .