Method and apparatus for automatic sorting of unmarked power cells

Abstract

Method for automatic sorting of batteries, the method including generating a fan-shaped X-ray beam; scanning the batteries with the fan-shaped X-Ray beam; for each battery, capturing X-rays that pass though the battery with an X-ray detector and converting the X-rays into first and second digital images, wherein the first digital image represents X-rays at a first energy, and the second digital image represents X-rays at a second energy; automatically analyzing the first and second digital images to determine a type of the battery by identifying characteristic features of each battery type based on a gray spectrum of at least 8 bit resolution that is looked up in a model database branch; and sorting the batteries by type.

Claims

1. A method for automatic sorting of cylindrical batteries, the method comprising: generating a fan-shaped X-ray beam; scanning the cylindrical batteries with the fan-shaped X-Ray beam; for each cylindrical battery, capturing X-rays of the fan-shaped X-Ray beam that pass though the battery with a dual-energy X-ray detector and converting the X-rays into first and second digital images, wherein the first digital image represents X-rays at a first energy, and the second digital image represents X-rays at a second energy; automatically analyzing the first and second digital images to determine a type of the cylindrical battery by identifying characteristic internal features of each battery type, including anode current collectors, based on a gray spectrum of at least 8 bit resolution that is looked up in a model database; and sorting the cylindrical batteries by type.

2. The method of claim 1, wherein the type is determined based on the gray spectrum using size indices and gray indices of the characteristic internal features in the cylindrical batteries.

3. The method of claim 1, wherein the characteristic internal features also include gray contrasts that indicate a porous separator and a plastic cover, which are unique for each individual battery type, and further comprising indexing the characteristic internal features for the model database.

4. The method of claim 1, wherein the gray spectrum includes 65,535 shades of gray.

5. The method of claim 1, wherein the model database includes a branch index used for finding a branch that corresponds to the battery type.

6. The method of claim 5, wherein each branch in the model database has its own index.

7. The method of claim 6, wherein comparisons between the gray spectra for the sample are made within the model database branch having a corresponding index only.

8. An apparatus for automatic sorting of cylindrical batteries, the apparatus comprising: a slanted conveyor with a guiding plank for placing batteries; an X-Ray source generating a fan-shaped X-ray beam; a dual-energy detector for detecting the X-ray beam at two different energies from the X-ray source; a processing computer; a positioning system responsive to the processing computer; and a discharge system with discharging funnels for sorting the cylindrical batteries, wherein the dual-energy detector detects the X-ray beam after it passes through the batteries; for each battery, the processing computer converts X-rays of the fan-shaped X-Ray beam into first and second digital images, wherein the first digital image represents the X-rays at a first energy, and the second digital image represents the X-rays at a second energy; and the processing computer analyzes the first and second digital images to determine a type of the cylindrical battery by identifying characteristic internal features of each battery type, including anode current collectors, based on a gray spectrum of at least 8 bit resolution that is looked up in a model database.

9. The apparatus of claim 8, wherein the positioning system comprises at least one battery sensor, a conveyor position sensor, and a controller.

10. The apparatus of claim 8, wherein the slanted conveyor runs through a scanner tunnel.

11. The apparatus of claim 8, wherein the discharging funnels are mounted at a fixed distance from the scanner and at certain intervals, and wherein dedicated containers are placed along the conveyor.

12. The apparatus of claim 8, further comprising protective screens that block scatter radiation.

13. The apparatus of claim 8, wherein imaging positioning prevents battery image from being superimposed by that of the guiding plank.

14. A method for automatic sorting of batteries, the method comprising: generating a fan-shaped X-ray beam; scanning the batteries with the fan-shaped X-Ray beam; for each battery, capturing X-rays that pass though the battery with an X-ray detector and converting the X-rays into first and second digital images, wherein the first digital image represents X-rays at a first energy, and the second digital image represents X-rays at a second energy; automatically analyzing the first and second digital images to determine a type of the battery by identifying distinctive regions of each battery type based on a gray spectrum of at least 8 bit resolution that is looked up in a database, wherein the distinctive regions include gray contrasts that indicate an anode current collector, a porous separator, and a plastic cover, which are unique for each individual battery type, and then indexing the distinctive regions; and sorting the batteries by type.

Description

BRIEF DESCRIPTION OF THE ATTACHED FIGURES

(1) The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention.

(2) In the drawings:

(3) FIG. 1 illustrates regions within an exemplary battery (type 1 from the five basic types, shown in FIG. 7), as seen on an X-ray.

(4) FIG. 2 illustrates an algorithm for creating a database element.

(5) FIG. 3 illustrates a battery recognition algorithm.

(6) FIG. 4 illustrates a preferred embodiment of the apparatus for sorting power cells.

(7) FIG. 5 shows a diagram of the apparatus for automatic sorting of unmarked power cells.

(8) FIG. 6 shows major functional parts of the X-Ray scanner.

(9) FIG. 7 shows five basic types of the power cells to be sorted.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

(10) Reference will now be made in detail to the preferred embodiments of the present invention, examples of which are illustrated in the accompanying drawings.

(11) To overcome this issue, an expanded gray spectrum algorithm is proposed that separates comparison branches. The gray spectrum is generated from a hi-res grayscale image at 16-bit color depth (65,536 shades of gray). Two images—high-energy image and low-energy image—are used. When a model spectrum is generated, it is first indexed in accordance with the dimensions of the original image and gray values in distinctive regions for a given type.

(12) Each power cell type has its own distinctive regions that can be automatically identified in X-Ray images based on the internal structure of a power cell. FIG. 1, left side and FIG. 1, right side, illustrate examples of such distinctive regions, where FIG. 1, left side illustrates internal structure of a power cell (Type 1), and FIG. 1, right side is the X-Ray image of the same power cell (Type 1).

(13) According to the existing power cell sorting algorithm, first, distinctive regions for each power cell type have to be found, e.g. porous separator 1, anode current collector 2, plastic cover 3, etc. (see FIG. 1).

(14) By comparing distinctive regions of different power cell types, unique regions and combinations thereof are defined for each individual type, which are then indexed. FIG. 7 shows that each type of batteries has: different widths, lengths, place of attachment of the central electrode, which is clearly visible in the X-Ray images. FIG. 7 shows the primary types of batteries: type 1—alcali, type 2—salt (Zn-C, Mn-Zn), type 3—Li-Ion (A—spiral type, B—wound type), type 4—Ni-Cd, type 5—Ni-metal hydride (Ni-Mh). The homogeneous structure of the main body of the electrode can be distinguished from the striped image of the spiral arrangement of the negative-positive electrode group.

(15) The more unique distinctive regions with corresponding indices there are for each power cell type, the more reliable the automatic sorting procedure is when it comes to separating power cells by type).

(16) Such indexation allows to generate a plurality of separate branches in the model database. Therefore, when a sample is compared with the model database, only one branch is used, which greatly increases the speed of analysis, freeing computer resources and allowing to use hi-res gray spectra that have been generated from detailed images at 16-bit color depth (65,536 shades of gray).

(17) This approach allows to maintain recognition speed of up to 50 msec, at 99%+ accuracy. FIG. 2 illustrates an algorithm for creating a database element. FIG. 3 illustrates a recognition algorithm.

(18) An X-Ray dose detector and an apparatus from sorting used-up power cells, as disclosed in EP1132992B1, is closely related to the claimed apparatus. The apparatus according to EP1132992B1 comprises a dosing device, two slanted conveyors moving at different speeds and equipped with a guiding plank, an X-Ray source, two fluorescent X-Ray screens mounted into the guiding plank, two optical lines with phototransistors, and a pneumatic sorting device. This apparatus employs the identification method according to DE19610093A1.

(19) In this apparatus, power cells have to be spaced minimum 10 cm apart, so that they can be captured and then sorted. Therefore, in order to achieve required performance parameters, scanning speed has to be increased, which, in turn, leads to increased source power, increased detector speed, and deteriorated image quality. Also, increased source power imposes additional construction costs to protect personnel from scatter radiation that spreads along the conveyor belt.

(20) To overcome this issue, a power cell positioning system may be installed on the conveyor belt of the apparatus (see FIG. 5), the positioning system comprising an optical sensor 13, a conveyor position sensor 15, and a controller 9. This system locates power cells, thus allowing to accurately capture is image with the detector 7, and then discharging funnels 14 generate a blowing impact that is equal to the power cell length in order to remove the power cell from the scanner. This solution allows to reduce the distance between power cells to 5 cm, thus reducing scanning speed and source power, while maintaining same sorting performance

(21) The apparatus according to EP1132992B1 comprises a detector, which is made of two fluorescent screens mounted apart from each other, with two phototransistors that are connected to the screens via optic fiber cables. This apparatus generates battery images that are two pixels wide, which is not enough to generate an expanded gray spectrum that allows to identify power cell type by means of X-Ray crystallography.

(22) Holes in the fluorescent screens that are mounted into the guiding plank come into direct contact with the power cell that is moved along. Thus, the holes will be filled with a mixture of electrolytes and battery case oxides, which will result in identification errors.

(23) Another problem is that particles of chemically active electrolytes hang in the air, which directly affects performance of both the generator and detector, as well as other electronic parts.

(24) To overcome these issues, an apparatus for automatic sorting of unmarked power cells is proposed. FIG. 4 illustrates a preferred embodiment of the apparatus for sorting power cells.

(25) The main components of the apparatus for automatic sorting of unmarked power cells, as shown in FIG. 4, are: 1. X-Ray scanner; 16. Feeding bin, where used-up power cells are loaded; 17. Feeding conveyor to move power cells from the feeding bin 16 to the aligning conveyor 18; 18. Aligning conveyor; 12. Feeding conveyor; 6. Main slanted conveyor; 11. Protective screens; 14. Discharging funnels that blow sorted power cells from the main conveyor; 19. Guiding grooves; 20. Containers, where sorted power cells are collected; 21. Container for unidentified power cells.

(26) FIG. 5 is a diagram of the apparatus for automatic sorting of unmarked power cells, showing major functional parts. In this figure: 1. X-Ray scanner; 2. X-Ray source; 3. Scanner tunnel; 4. Guiding plank; 5. Used-up power cells; 6. Main slanted conveyor; 7. Detector; 8. Processing computer; 9. Controller; 10. Feeding system that comprises the feeding bin 16, Feeding conveyor 17 and aligning conveyor 18 (see FIG. 4); 11. Protective screens; 12. Feeding conveyor; 13. Optical power cell sensor; 14. Discharging funnels; 15. Conveyor position sensor.

(27) FIG. 6 shows major functional parts of the X-Ray scanner. In this figure: 1. X-Ray scanner; 2. X-Ray source; 3. Scanner tunnel; 4. Guiding plank; 5. Used-up power cells; 6. Slanted conveyor; 7. Detector.

(28) In the exemplary embodiment, main specifications and operation modes of major functional parts of the X-Ray scanner, as shown in FIG. 6, may be as follows: optimum operation modes for the X-Ray source 2 is as follows: anode voltage of 180-200 kV, anode current of 2.0-2.6 mA; focal spot size suitable for identifying the internal structure of power cells and sorting them is 0.8 mm; effective sensor size of the detector 7 may be 0.8 mm; either single-energy or dual-energy scintillation detector, as well as a semiconductor detector for direct counting of photons, with cadmium tellurium (CdTe) active layer, may be used.

(29) Functional features of a semiconductor detector for direct counting of photons may be illustrated by the following example:

(30) In the exemplary embodiment, the active length of the detector array is 100 mm and the active width of each detector array is 6.4 mm Each detector array has 64 pixel lines and the size of each pixel is 0.1 by 0.1 mm.

(31) In order to obtain the required pixel size of 0.8 mm, 8×8 pixel binning may be used.

(32) This detector may operate in both single-energy and dual-energy mode.

(33) The detector array allows using algorithms for generating a high-contrast image, while moving the power cell through the X-Ray beam at a high speed.

(34) A digital image is formed, so that the number of elements (pixels) along one coordinate (in the scanning direction) is determined by the number of scanning steps (number of readings) and the number of detectors in the direction perpendicular to the scanning direction along the other coordinate.

(35) The detector receives X-rays and generates images in TDS (Time Delayed Summation) mode, i.e. when an object moves across the detector in such a way that the same area is imaged one row at a time, thus allowing each row to be summed in order to get a more statistically accurate sample of a given section of the object. These samples form lines in the output image, so as to form a complete scan of the object.

(36) In TDS mode, detector pixel values are shifted along rows during image acquisition to compensate for detector motion. This detector is a non-scintillating detector type (e.g., cadmium-telluride) that allows counting the photons directly.

(37) This detector allows to generate quality X-Ray images using relatively low-current X-ray sources, e.g., up to 2.5 mA of the current requirements for the X-ray source.

(38) Having thus described a preferred embodiment, it should be apparent to those skilled in the art that certain advantages of the described method and apparatus have been achieved.

(39) It should also be appreciated that various modifications, adaptations, and alternative embodiments thereof may be made within the scope and spirit of the present invention. The invention is further defined by the following claims.