How to copy the opening key from this key image

Abstract

The present invention relates to a method for replicating a key for a locksmith from an image of said key, comprising the steps of determining a) the type of key, b) the key blank, c) the key cutting, wherein steps a) and b) are performed by at least one deep learning algorithm.

Description

[Technical Field]

[0001] The present invention relates to the field of duplication of keys for all kinds of openings. [Background technology]

[0002] In the context of the present invention, an opening may be, for example, a door or a drawer.

[0003] For various reasons in everyday life, it is desirable to be able to copy a key, so that, for example, in case of loss of the original key, one has a duplicate.

[0004] Until recently, copying a key required a visit to a locksmith who could use conventional mechanical tools to create a true copy of the original key.

[0005] To simplify these processes, several solutions are available on the market to automate this classical key duplication process.

[0006] A common feature of these solutions is the proposal to start with an image of the key to be duplicated, which image can be taken, for example, with the camera of a mobile phone.

[0007] This image is then sent to a control center, where it is analyzed and the customer is told whether the key falls into the category of keys that can be remotely copied.

[0008] If it falls into the category of keys that can be copied remotely, after the quote (which is generated manually and not instantaneously) is approved by the customer, the key will be copied remotely and then sent to the customer.

[0009] An example of such a process is given in European Patent Application No. 3992821.

[0010] In this patent application, the blank of the key to be reproduced, i.e. the key model before cutting is performed, is determined from singular points marked on the original key, and these singular points are then used to select a suitable blank from a blank database.

[0011] Although the process described in this patent application has certain advantages compared to other similar solutions, such as reduced computation time, it is not entirely satisfactory in terms of reliability and accuracy of copying the original key. Furthermore, the customer has to wait a while to get a quote, which increases the overall processing time. Summary of the Invention [Problem to be solved by the invention]

[0012] Therefore, the present patent application is directed, inter alia, to improving the process for duplicating keys from one or more images of the original key in a manner that increases reliability and accuracy and reduces the overall time between a customer order and delivery of the duplicate key. [Means for solving the problem]

[0013] This object of the invention, as well as other advantages that will become apparent in the light of the following description, is achieved by a method for replicating a key for an opening from at least one image of said key, comprising: a) the type of key; b) key blanks; c) Key cutting; determining a This is achieved by a method in which steps a) and b) are performed by at least one deep learning algorithm.

[0014] By using such artificial intelligence algorithms, errors made by the algorithms can be corrected over time, increasing reliability and accuracy and reducing overall processing time.

[0015] Thus, thanks to the method according to the invention, keys can be accurately and reliably replicated with virtually no human intervention, making it possible to offer customers a very competitive solution in terms of cost and time.

[0016] According to other optional features of the method according to the invention, it comprises, alone or in combination:

[0017] said algorithm is of the deep convolutional neural network type, such algorithms being known per se and particularly suitable for key duplication methods;

[0018] The background of the image of the key is removed by a neural network algorithm, which allows for the correction of errors associated with the presence of a textured background;

[0019] The orientation of the key relative to the image frame is determined, this step increases reliability when analyzing the key profile;

[0020] Blurred areas of the key image are reduced, this step also contributes to increased reliability;

[0021] The actual dimensions of the key are determined from the distance between the pistons of the relevant lock, this step makes it possible to know the absolute distances between the pixels of the analyzed image of the key and thus generate the information necessary to cut the key;

[0022] The shadows attached to the key are corrected, this correction step making it possible to know the absolute width of the key blank and thus generate the information necessary to cut the key;

[0023] the front and rear views of at least a portion of the key are juxtaposed on the same image, this juxtaposition making it possible to obtain maximum information about the key to be replicated from the analysis of a single image;

[0024] step b) is carried out by analyzing at least one specific area of ​​the key, which analysis of this specific area makes it possible, on the one hand, to reduce the size of the image to be processed and, on the other hand, to avoid contaminating the artificial intelligence algorithm with information that could lead to misunderstandings;

[0025] The analysis is performed from the first derivative of the curve of the key profile, which has been shown to be a simple means of identifying the key profile's characteristic points;

[0026] The specific area includes the joining area of ​​the body of the key with the head, and this joining area has been shown to be characteristic for various key blanks available on the market;

[0027] The characteristics of the grooves of the key in said specific area are analyzed, these characteristics being selected from the group comprising: the number of grooves, the position of these grooves, the depth of these grooves, the shape of the ends of these grooves.

[0028] Furthermore, the invention relates to a computer program product suitable for implementing the method according to the above.

[0029] Furthermore, the invention relates to a key duplicating system comprising means for taking a photograph of a key, means for analyzing the image thus obtained, means for determining the type of key and the blank, and means for cutting said blank, and controlled by the computer program according to the above. [Brief explanation of the drawings]

[0030] Other features and advantages of the present invention will become apparent from a consideration of the following description taken in conjunction with the accompanying drawings.

[0031] [Figure 1] 1 shows a diagram of a flat key to be replicated. [Figure 2] The figure shows a front view and a back view of the joint area between the body of a flat key and the head. [Figure 3] The curve of the profile of the flat key to be replicated and the first derivative of this curve are shown. [Figure 4] A flat key shank highlighting the different cutting depths corresponding to each piston of the associated lock. DETAILED DESCRIPTION OF THE INVENTION

[0032] For clarity, identical or similar elements will have the same or similar reference numbers in all figures.

[0033] In the following, the invention will be described in the context of replicating flat keys, but it will be understood that the principles of the invention are applicable to replicating any other type of key.

[0034] Reference is now made to Figure 1, which shows the flat key to be replicated.

[0035] Such a flat key comprises a head 1, a body 3 in which a cut-out profile 5 is formed, a stopper 7 arranged in the connection area between the body and the head, and longitudinal grooves 9, 11 extending over the length of the body.

[0036] The key head 1 allows for gripping, the key body 3 with its cut-out profile 5 and grooves 9, 11 allows for cooperation with the cylinder and the piston of the lock, and the key stop 7 allows for correct longitudinal positioning of the key when inserted into the lock.

[0037] Such a key is referred to as flat in that, unlike other keys on the market such as pump keys or grooved keys or wave keys (the latter often used to open car door locks), the head 1, body 3, cut-out profile 5 and stopper 7 are actually contained within one plane.

[0038] It is therefore an object of the present invention to be able to replicate such flat keys in a manner that is simple, automated, accurate, reliable and fast.

[0039] This is done by first capturing front and back images of the key to be duplicated, which can be provided by the camera on the user's mobile phone, or by an imaging device installed in a fixed kiosk located, for example, near a supermarket.

[0040] Ideally, the image of the key should be taken against an unstructured, matte background, i.e., a plain background without texture.

[0041] In practice, this condition cannot always be met, so it is useful to use background suppression applications (segmentation) that use neural network algorithms, such as the applications available on the platform https: / / remove.bg / fr.

[0042] Furthermore, algorithms known per se are used to correct the orientation of the key relative to the edges of the image framing and to absorb blurred areas.

[0043] From the image thus obtained, the type of key in question is first determined: for example, there are roughly 16 types of keys in Europe (flat, pump, grooved, wave, etc.).

[0044] This decision step is performed by an artificial intelligence algorithm, more specifically a deep convolutional neural network algorithm.

[0045] Such algorithms use a large number of previously verified images to provide accurate and reliable results regarding the type of key.

[0046] Once the key type (flat, pump, grooved, wave, etc.) is determined, the next step is to determine the key model, i.e., the blank selected to cut the appropriate profile.

[0047] The key blank to be copied is in fact a key suitable for insertion into the same lock as the key to be copied, i.e. a key with a body having dimensions (length, width) and shape (cross section) that allow it to be inserted into the relevant lock, but without a cut-out profile.

[0048] For example, in Europe there are over 2,000 different blanks for flat keys.

[0049] In the method according to the invention, a deep learning algorithm, preferably of the deep convolutional neural network type, is again used to determine a suitable key blank from an image of the key: therefore, unlike the prior art, no blank database is used which is queried by the feature descriptors of the key to be duplicated.

[0050] This deep learning algorithm can be gradually corrected by an operator to improve accuracy and reliability.

[0051] One option that has proven particularly effective allows the size of the image to be analyzed to be reduced to a specific region of the key.

[0052] More specifically, as can be seen in FIG. 2, the image analyzed may relate to the front R and rear V of the joining area with the head 1 of the body 3 of the key, where the stopper 7 of this key is located.

[0053] In fact, this junction region turned out to be highly characteristic of the key to be replicated.

[0054] FIG. 3 shows the curve C1 of the profile of the key to be duplicated and the curve C2 of the first derivative of this profile curve.

[0055] The curve C2 of the first derivative contains two characteristic peaks P1, P2 corresponding to the position of the stopper 7 of the key and the inflection point I of the head of the key, respectively.

[0056] The positions of these two peaks P1, P2 on the longitudinal axis of the key (x-axis in Figure 3) make it possible to define a band B on the key that extends from a position a slightly in front of the stopper 7 to a position b corresponding to the inflection point I.

[0057] This band B corresponds to the area of ​​the key where the key grooves 9a, 11a (front) and 9b, 11b (back) end and, depending on the type of milling used, includes the full range of possible shapes for the ends of these grooves, such as conical, or pointed, as shown for example in Figure 2, or beveled.

[0058] A review of commercially available flat key models reveals that the number of grooves in the band B region on both sides of the key, the location of these grooves, the depth of these grooves, and the shape of the edges of these grooves are characteristic information of each key model.

[0059] Therefore, by analyzing images of both sides of the key in the area of ​​band B by a deep learning algorithm, preferably of the deep convolutional neural network type, it becomes possible to achieve a reliable separation between different flat key models.

[0060] In particular, it should be noted that the fact that band B starts a little before stopper 7 (part a of band B in Figure 3) makes it possible to provide a margin to cover the case where grooves 9, 11 stop before this stopper.

[0061] Once the key blank to be duplicated has been determined in this way, the parameters required for physically cutting the body 3 of the key are determined.

[0062] For this purpose it is essential to know the actual dimensions of the key to be replicated, i.e. the absolute distance between pixels of the image of the key to be replicated.

[0063] For this purpose, this image is calibrated from the distance d separating two consecutive grooves G1, G2, as shown in Figure 4: this distance d, which corresponds to the distance separating the two pistons of the associated lock, is in fact known for a given key blank.

[0064] For greater accuracy, it is also possible to take the average of all the distances between grooves to be replicated from the key.

[0065] Another important parameter is the width of the barrel 3 of the key to be duplicated, i.e. the width of the blank.

[0066] This width determination may be subject to error due to shadows when imaging the key to be duplicated.

[0067] For this purpose, once a suitable blank has been determined according to the principles described above, the width of the accompanying shadow is estimated by difference: this makes it possible to accurately determine the depth of each groove to be cut into the blank, which can be represented by a coding ranging from 1 to 10, for example, as shown in Figure 4.

[0068] Once the key type, model, and notches to be made in the blank are determined, the customer can be immediately sent an automated quote for a duplicate key, for example to an email address.

[0069] Upon receiving the customer's approval of the quote, an automated machine, under operator supervision, can cut the key profile into the body of the blank from the key encoding determined above.

[0070] A finishing (polishing) and inspection step can then complete the saw.

[0071] As can be seen from the above, the method according to the present invention allows for accurate, reliable and fast automation of the process of reproducing keys, such as flat keys.

[0072] The method according to the invention can be implemented by an integrated machine comprising means for imaging the key, software means for carrying out the steps described above and means for cutting the key profile: in this case the user places the key to be duplicated in the machine and receives the duplicated key a few moments later.

[0073] Of course, a human-machine interface, such as a smart phone application, is provided to ask the user questions (e.g., how many copies of the duplicate key to make), submit quotes, and bill for the services.

[0074] Alternatively, the user can take a photo of the key to be duplicated themselves, for example using the camera on their mobile phone.

[0075] An application downloaded to the phone can then transmit these images to a remote server where the above-mentioned analysis and processing means are implemented.

[0076] An automated machine, either at the same location as this remote server or at another location, supervised by an operator, can cut the profile onto an appropriate key blank and then send the duplicate key to the user by mail.

[0077] In the process of determining the key type and appropriate blank, questions asked of the user at the machine and / or operator intervention allow the deep learning algorithm to gradually correct errors it makes, improving the accuracy and reliability of the process over time.

[0078] Of course, the present invention has been described above by way of example only, and it will be appreciated that those skilled in the art can create various alternative embodiments of the invention without departing from the scope of the invention.

Claims

1. 1. A method for replicating a key for an opening from at least one image of said key, comprising:

1. A method comprising the steps of: a) determining a key type; b) determining a key blank; and c) determining a key cutting, wherein steps a) and b) are performed by at least one deep learning algorithm.

2. The method of claim 1 , wherein the algorithm is of the deep convolutional neural network type.

3. The method of claim 1 or 2, wherein the background of the image of the key is removed by a neural network algorithm.

4. The method according to any one of claims 1 to 3, wherein the orientation of the key relative to the frame of the image is determined.

5. The method according to any one of claims 1 to 4, wherein blurred areas of the image of the key are reduced.

6. A method according to any one of claims 1 to 5, wherein the actual dimensions of the key are determined from the piston-to-piston distance (d) of the associated lock.

7. The method according to any one of claims 1 to 6, wherein shadows associated with the key are corrected.

8. The method of any one of claims 1 to 7, wherein a front view and a back view of at least a portion of the key are juxtaposed on the same image.

9. The method according to any one of claims 1 to 8, wherein step b) is performed by analysing at least one particular region of the key.

10. 10. The method according to claim 9, wherein the analysis is carried out from the first derivative (C2) of the curve (C1) of the profile (5) of the key.

11. 11. The method according to claim 9 or 10, wherein the specific area includes a joining area of ​​the body (3) of the key with the head (1).

12. 12. The method of claim 11, wherein groove characteristics of the key in the particular area are analyzed, the characteristics being selected from the group including: number of grooves, location of the grooves, depth of the grooves, shape of the ends of the grooves.

13. A computer program product adapted for carrying out the method according to any one of claims 1 to 12.

14. 14. A key duplicating system comprising means for taking a photograph of a key, means for analyzing the image thus obtained, means for determining the type of key and the blank, and means for cutting said blank, controlled by the product of claim 13.

Images