Device and method for optical inspection

Abstract

The invention relates to a device (10, 10', 10'', 10''') for optically inspecting one or more objects (12, 12a, 12b), in particular food, having the following features: an optical sensor (16), which is designed to measure an optical property of the one or more objects (12, 12a, 12b) in order to obtain at least a first measurement value associated with a first feature; a further sensor (18s) or further data sources (18d), which is / are designed to measure a further property of the one or more objects (12, 12a, 12b) in order to obtain at least a second value associated with a second feature; and a processor (20), which is designed to compare the first measurement value associated with the first feature and the second value associated with the second feature with a model or trained model in order to derive state information for the one or more objects (12, 12a, 12b).

Description

[0001] DEVICE AND METHOD FOR OPTICAL ASSESSMENT

[0002] Description

[0003] Embodiments of the present invention relate to a device and a method for the optical inspection of one or more objects, in particular foodstuffs. Further embodiments relate to a corresponding computer program.

[0004] The digital capture of as many product characteristics as possible using optical sensors paves the way for the automated capture of commercial data in goods logistics (identification, quantity, weight, etc.) as well as for the assessment of their quality (freshness, sweetness, acidity and water content, and much more). The principles of these processes are generally well-known and are already widely used for specific products. What is new about the concept, in addition to the possibility of creating digital twins (e.g., of a box or pallet of identical products), is the inclusion of additional product characteristics. These characteristics can also be features invisible to the human eye, which are captured spectroscopically in the near-infrared and microwave ranges (spectral footprint) to obtain more detailed information, particularly about the internal properties of products.

[0005] The quality of fruit and vegetables is determined by a variety of influencing factors and described by (occasionally) changing evaluation criteria. Growth period, origin, and environmental conditions make it necessary to keep these changing parameters up-to-date for product assessment along the food chain, especially at the point of sale (POS). Manual inspection can only capture and document these changes to a limited extent, partly due to the sometimes subjective and varying perceptions from employee to employee.

[0006] Modern image analysis across a broad spectral range not only enables the capture of information such as quantity, type, and various properties, but can also detect variances within one or more batches, distinguishing expected from unexpected changes, and alerting the user accordingly. The added value for the user arises from the automated comparison of current and historical data using machine learning methods, on the one hand, and from information captured "incidentally" for commercial use, on the other (traditionally: data from incoming goods (GD), etc.).

[0007] Some approaches already exist in the prior art, such as those disclosed in the article "Food Analysis Using MI R Spectroscopy" or in the article "Optical Analysis" by I PMS. These two publications explain spectroscopy and hyperspectral imaging, respectively. Other sensors for food assessment can also be found in the prior art. The publication entitled "Application of Microwave Spectroscopy Analysis on Determining Quality of Vegetable Oil" demonstrates spectroscopic analysis in microwave frequency ranges. EP 3926333 A1 discloses X-ray examinations of fruit. An alternative sensor is the so-called sniffing sensor, e.g., for the assessment of fats in food chemistry, as disclosed in DE 19947669 A1.

[0008] The processing of sensor data is described in a BMEL publication titled "Movie-Q - Visual Quality Control Intelligently Optimized - How to Make Small Food Safer." In this context, reference should also be made to a patent application filed by Biometic with publication number EP 3967414 A1.

[0009] What all approaches have in common is that the data collection and / or analysis are suboptimal. Therefore, there is a need for an improved approach.

[0010] The object of the present invention is to create a concept for the optical inspection of objects, such as food, which has an improved compromise between detection accuracy, applicability in industrial environments, such as food retail, and cost efficiency.

[0011] The problem is solved by the subject matter of the independent patent claims.

[0012] Embodiments of the present invention provide a device for the optical assessment of one or more objects, in particular foodstuffs, having an optical sensor in combination with a further sensor, e.g. a non-optical sensor or a further data source, and a processor. The optical sensor is designed to detect an optical property of the one or more objects in order to obtain at least a first measured value associated with a first feature. The further sensor or the further data source is designed to detect a further, e.g. non-optical property of the one or more objects in order to obtain at least a second value associated with a second feature.The processor is configured to compare the first measured value associated with the first feature and the second value associated with the second feature with a model or trained model to derive state information for the one or more objects. According to embodiments, the trained model may be a model trained using K1.

[0013] Embodiments of the present invention are based on the realization that the use of different values ​​belonging to different characteristics offers a good starting point for assessing objects, such as fruit. A measured value alone is often not meaningful enough, since, for example, one measured value can be interpreted differently for different fruits, e.g. in the sense of different genera or different varieties. A particular spectrum can be interpreted differently for an apple of variety 1 than for an apple of variety 2. Therefore, for example, the information about the variety to be examined can be loaded via another data source. The same spectrum can also indicate different qualities for the same breeding and variety, namely depending on the degree of ripeness.In this respect, the degree of ripeness can be determined using the additional sensor, so that a quality assessment can then be made based on the degree of ripeness when viewed in conjunction with the optical measurement value. The advantage of this system is that a non-contact sensor in the form of an optical sensor is used in combination with additional data sources, which makes it possible to inspect the objects or food items in packaging, such as a vegetable crate. Furthermore, such optical sensors are cost-effective and can be flexibly integrated into any warehouse management system. A further advantage is the high flexibility of the system due to the fact that either an additional sensor, such as a microwave sensor, or a data source, such as the warehouse management software, can be used to extract the second feature.

[0014] According to embodiments, the processor is designed to train the model based on a training data set which comprises at least a first measured value, a second measured value and additional information. As already mentioned, the model can be a Kl model, so that the training functions with a Kl algorithm. The additional information is, for example, the result of the data set comprising additional information, measured value and second value. With regard to the sensors, it should be noted that according to embodiments, the further sensor can be a non-optical sensor, such as a microwave sensor. Alternatively, the further sensor can comprise an ultrasonic sensor, X-ray sensor or gas sensor. All of these further sensors have in common that they are sensors of a different type and detect a property of a different type, e.g. an odor or generally a non-optical property.Examples of gas sensors include the so-called sniffer sensor or the gas spectrometer. According to embodiments, the optical sensor can be a stationary camera, an RGB sensor, an HSI sensor, or an HSI camera. Alternatively, a 3D camera or multiple cameras can be used as the optical sensor. According to embodiments, the optical sensor is arranged at a defined distance from the one or more objects or the examination position of the one or more objects or from a container for the one or more objects or the examination position starting from a container for the one or more objects.

[0015] With regard to the optical sensor, it should further be noted that, according to embodiments, it is designed to take a spectroscopy image of the one or more objects as a measured value and / or to determine a spectral footprint of the one or more objects as a measured value or as a basis for determining the feature. For example, an optical spectrum or characteristics of the optical spectrum, i.e. the spectral footprint, can form the feature. According to embodiments, a plurality of first measured values ​​assigned to a plurality of 2D positions across the plurality of objects can be determined for the one or more objects. If, for example, it is assumed that the objects are only apples that are transported in an apple crate and are examined therein, a plurality of measured values ​​can be determined for a plurality of apples at a plurality of 2D positions. Of course, this also applies to any other fruits orObjects. According to a further embodiment, the further sensor can be designed to determine a plurality of second measured values ​​assigned to 2D positions or the 2D positions across a plurality of objects or even assigned to 3D positions along a penetration depth into the one or more objects (DGF also assigned to the previously determined 2D positions). These two embodiments therefore advantageously make it possible for a plurality of objects of a batch to be examined simultaneously or for an object to be examined at a plurality of positions. By assigning them via the 2D positions, first measured values ​​and second values ​​can be assigned to one another. By means of the special embodiment of determining the second value assigned to 3D positions, ie along a penetration depth, an object can be examined not only on the surface but also on the inside.Here, the first measured value can be linked to the second value based on the 2D positions. A conceivable variant for 3D determination is an ultrasonic sensor that, for example, determines second values ​​at different penetration depths. A different type of sensor can also be designed to determine second values ​​at different penetration depths.

[0016] In the above embodiments, it was always assumed that the second value is determined by a second sensor. According to a further embodiment, a value already stored in advance or a value determined in advance can also be used as the second value. This second value can be obtained, for example, via an interface to a merchandise management system that is available in accordance with further embodiments. According to embodiments, an additional value is also obtained in addition to the second value. The additional value can, for example, be information relating to the type of object, so that the first measured value can be classified. The additional value can be a previously determined measured value for the objects or information relating to the origin, age, harvest time or similar.In order to be able to assign this second value or the further additional value to the objects, the device can have an interface for integrating the one or more objects, in particular an interface for an RFID or a barcode reader, so that the second value and / or the further additional value is obtained via this interface or another interface. For example, fruit can be packaged in fruit crates equipped with an RFID tag or a barcode label. At this point, it should be noted that the optical sensor can also be designed, according to exemplary embodiments, to support or perform the identification, for example by scanning the barcode.

[0017] According to embodiments, the processor is configured to determine a characteristic quality value or a characteristic manufacturing specification as additional information, so that the characteristic quality value or the characteristic manufacturing specification can be compared with the information from an inventory management system. This serves for quality control, in that the quality value can be compared with a value stored in the inventory management system using sensor data, additional information, and the model. According to embodiments, the characteristics can come from the group comprising size, shape, color, and structure. The condition information can come from a group comprising variety, ripeness, freshness, quality, and origin.

[0018] As already mentioned above, a preferred embodiment envisions fruit, vegetables, and other foodstuffs as the object. According to the embodiments, the fruit, vegetables, or foodstuffs can be stored in boxes.

[0019] Another embodiment provides a method for inspecting one or more objects or food items. The method comprises the following steps:

[0020] - detecting an optical property of the one or more objects in order to obtain at least a first measured value associated with a first feature;

[0021] - detecting a further (e.g. non-optical) property of the one or more objects by means of a further sensor or from a further data source in order to obtain at least a second value associated with a second feature; and

[0022] - comparing the first measured value associated with the first feature and the second value associated with the second feature with a model or trained model to derive state information for the one or more objects.

[0023] According to further embodiments, the method can be computer-implemented.

[0024] Embodiments of the present invention are explained with reference to the accompanying drawings. They show:

[0025] Fig. 1 is a schematic representation of a system according to a basic embodiment;

[0026] Fig. 2 is a schematic representation of a camera, in particular a camera with a light source and a spectral sensor, for use as an optical sensor according to embodiments; Fig. 3 is a schematic representation of a concept of an inspection station at the incoming goods inspection according to further embodiments; and

[0027] Fig. 4 is a schematic representation of a concept of a testing station at the incoming goods area with different sensors.

[0028] Before exemplary embodiments of the present invention are explained below with reference to the accompanying drawings, it should be noted that elements and structures with the same function are provided with the same reference numerals, so that the description of them is applicable to one another or interchangeable.

[0029] Fig. 1 shows a system 10 for the optical inspection of one or more objects 12, here food items 12a and 12b. The system, which can also be integrated into a single device, comprises an optical sensor 16, another sensor 18s or a data source 18d, and a processor 20.

[0030] The optical sensor can, for example, be a camera configured to take a spectroscopic image of the objects 12a and 12b to be examined. The sensor data can be assigned to the individual objects 12a and 12b based on their position, according to embodiments. For example, the objects 12a and 12b can be located in a common food packaging, so that the camera 16 then performs a surface scan. The result of this scan is a measured value assignable to the one or more objects. The measured value can be in the form of a spectroscopic image or a spectral footprint, according to embodiments. This measured value(s) is / are then fed to the processor 20, where it is combined with other values, such as, for example, further measured values ​​from the sensor 18s.Sensor 18s can, for example, be a sniffer sensor that creates a gas spectroscopy image of objects 12a, 12b or of objects 12a, 12b as a value. These measured values ​​are also fed to processor 20 and can be assigned to the first measured values ​​via position. As an alternative to the measured values ​​from sensor 18s, a value can also be loaded directly from a database, e.g., an inventory management system. Database 18d then transmits the corresponding values ​​for objects 12a and 12b to the processor, which then combines them. The first measured value can be interpreted using the second value from sensor 18s or database 18d. This is done according to embodiments using a model or a Kl model.

[0031] From the first measured value of the sensor 16, a characteristic such as a size, a shape, a color, a structure, etc. can be determined. A characteristic is also derived from the second value, e.g., an origin, determined from the merchandise management system 18d. The first characteristic, in combination with the second characteristic, provides information about a condition, e.g., a quality grade. If, for example, one assumes a color for the first characteristic, one and the same color can be interpreted differently for different objects 12a and 12b of different origins. In this respect, only the second characteristic, in combination with the first characteristic, makes it possible to derive information about the quality.

[0032] For this purpose, a model or a pre-trained model is used. The model can be determined using Kl and can also be further trained during operation. A possible application is explained below with reference to Fig. 2.

[0033] Fig. 2 shows a device 10' with two cameras 16a and 16b as sensors. The two cameras 16a and 16b enable a 3D image of the objects 12 in the object carrier 12t. Optionally, in addition to the cameras 16a and 16b, a light source 17 can also be provided to illuminate the objects 12.

[0034] Cameras 16a and 16b can be implemented as follows, according to exemplary embodiments: The technology to be used for image acquisition is essentially based on standard camera technology in the RGB and / or HSI range, or, as a potential alternative to HSI cameras, spectral single sensors. This technology is widespread and widely used.

[0035] According to exemplary embodiments, one camera 16a can be used as the first sensor and the second camera 16b as the second sensor. Furthermore, it would also be conceivable for additional features to be determined via a database (not shown).

[0036] The targeted metrology concept is based on classical image processing techniques. In addition to comparing features (size, shape, color, structure, etc.), segmentation—that is, distinguishing known objects from one another—is a core functionality.

[0037] In a first project step, the recording situation will be designed so that stationary cameras with lighting, each positioned from above and at a defined distance, capture an image of a predetermined fruit / vegetable crate (product). The crate will be placed at a defined distance and position.

[0038] The basis for evaluations is historical data, which may be further enhanced by new measurements. Their evaluation and comparison with current measurements enable appropriate assessment of the products under consideration; whether to identify the correct apple variety for payment at the checkout, to determine ripeness shortly before delivery to the retailer or at the end of the week, to check specifications and indicate origin; of course, this data is always integrated into the company's IT system via digital interfaces and thus potentially part of further transactions.

[0039] For the first step in implementation, we use the customer's warehouse as a starting point, as much data is collected there anyway, and the step to related commercial data is easy to implement. Furthermore, automated collection can also be useful for other purposes (rationalization, etc.). Data from earlier steps in the food chain will be imported at a later stage.

[0040] Referring to Fig. 3, another configuration is explained. Fig. 3 shows a device 10" for examining an object 12, wherein the objects are individually detected. Sensor 16 is provided for this purpose, while a second sensor 18 is provided in the object carrier or object stage 12t. The second sensor 18 can be a spectral sensor. According to embodiments, the camera 16 can be equipped with an integrated light source.

[0041] By optically recording incoming goods quality using camera / sensor technology, individual products and product deliveries become more comparable with each other, whether due to the possible consideration of natural (e.g., seasonal) influences or due to different origins. In addition, knowledge of the corresponding product / characteristic history is accessible and easily referenced. Broadening the information base by recording the spectral footprint as a basis for characteristic determination reduces the effort required for other measurements (e.g., sweetness => spectral identifier instead of Brix) and, in the foreseeable future, opens up additional information for product characteristics (e.g., fruit fibrousness) that is not yet available today, without additional effort.

[0042] Furthermore, the collected data can form the basis for further / new applications. For example, simplified product identification at the checkout or customer measurements of freshness or flavor at the POS. In the medium term, this could make it possible to distinguish sour apples from sweeter ones, or to identify sweeter strawberries in direct comparison.

[0043] Collecting product data along the entire food chain opens up the possibility for continuous product monitoring. The temporal context potentially provides not only indications of changes due to external influences such as storage or transport, but also indications of alternative supply sources or food fraud.

[0044] Since the technology can, in principle, be used whenever it comes to detecting characteristics with visual relevance, other applications are also conceivable. The advantages become apparent whenever it comes to detecting changes or deviations at different points in time and / or in comparison to other measurements. In the medical field, for example, the technology is used to monitor wound healing or to diagnose changes in melanoma.

[0045] Fig. 4 shows a further development of the embodiment from Fig. 2, namely the device 10'". The device 10"" has, in addition to the sensor 16a and the light source 17, a further sensor 18s, here an ultrasonic sensor. The sensors 16a and 16s scan the objects 12 or food items 12 in the transport container 12t, for example, from above. The surface of the objects 12 is examined by means of the camera 16a, while the ultrasonic sensor 18s is designed to enable depth resolution and to scan the objects 12 in the depth direction. Since there are several objects 12 in the transport container 12t, the measured values ​​from the sensor 16a (first measured values) can be assigned to the measured values ​​of the sensor 18s (second values ​​or second measured values) via the position in the transport container 12t. For example, the surface can be divided into a grid, so that each grid unit has a Measured value using sensor 16a or18s and can be assigned to one another. The assignment is performed via the processor 20. The processor 20 is also connected to a database 18d, from which further data on the objects 12 is obtained. For example, this can be a categorization of type, genus, and origin of the objects 12 and / or a quality specification received from the manufacturer. The quality specification is verified by the processor 20 by determining further quality information based on the first and second measured values ​​(if necessary with the aid of the information regarding genus, type, and origin) and comparing this with the quality information taken from the database 18d.

[0046] According to embodiments, the quality value obtained by the database 18d can be understood as part of a training data set. In this respect, the quality value is used as the expected result, and the model or the class model is trained by the processor 20 such that the first and second values ​​in the model are assigned to the result, such as the quality grade.

[0047] As already described above, the technology of optically capturing object characteristics (here, food) to describe their quality has been widely used for some time. The basic principle is essentially comparable.

[0048] Regarding the model, the model to be trained, or the AI ​​model, it should be noted that this concerns the path to "comparable" measurements; that is, issues such as sensor calibration, storage of results, comparable lighting, etc., which are used by many manufacturers of feature recognition applications. According to the exemplary embodiments, the control of the lighting 17 is also understood as a feature.

[0049] In general, it should be noted that the number of measured values ​​or general features, which also include lighting, is not limited, since the model can have any number of dimensions. The Kl algorithm determines which dimension is relevant for the evaluation. Differences often arise from the type or genre of the objects 12, so that the information determined via the database 18d is essential. According to embodiments, the objects 12 can be assigned to the correct data record in the database 18d via a reading device that is designed to receive identification information about the objects 12, for example, from the transport box 12t. A barcode or RFID tag can be used for this purpose. According to embodiments, the barcode can be read by the camera 16s. The correct data record is then retrieved from the database 18d using this read identification information.According to the implementation examples, this serves as input information for the model or trained model.

[0050] Even though the above examples focused on food as an object, it should be noted that the same approach can also be used for other types of objects, such as for sorting bulk goods.

[0051] In the above embodiments, it was assumed that objects 12 are located in a fixed position, e.g., in a slide. According to embodiments, the objects, such as food items (fruit, vegetables, etc.), can also be provided as bulk goods. For example, the food items are conveyed through the scanning area of ​​the device 10, 10', 10", 10'" by means of transport means, such as a conveyor belt.

[0052] With regard to the sensors, it should be noted that, according to embodiments, they scan an identical or at least comparable scanning area in which the objects are to be expected, in order to obtain the first measured values ​​and second values.

[0053] With regard to the measured values, it should be noted that, according to the exemplary embodiments, multiple measured values ​​can be assigned to multiple objects. Of course, multiple measured values ​​can also be assigned to one object. According to the exemplary embodiments, a first measured value in combination with a second value for an object is sufficient to apply the concept described above.

[0054] In the above embodiments, it was assumed that the device 10 exists as a unit. According to embodiments, the device 10 can also exist as a system, ie, with distributed components.

[0055] Although some aspects have been described in the context of a device, it should be understood that these aspects also represent a description of the corresponding method, so that a block or component of a device can also be understood as a corresponding method step or as a feature of a method step. Analogously, aspects described in the context of or as a method step also represent a description of a corresponding block, detail, or feature of a corresponding device. Some or all of the method steps may be performed by (or using) a hardware apparatus, such as a microprocessor, a programmable computer, or an electronic circuit. In some embodiments, some or more of the key method steps may be performed by such an apparatus.

[0056] Depending on specific implementation requirements, embodiments of the invention may be implemented in hardware or software. The implementation may be performed using a digital storage medium, such as a floppy disk, a DVD, a Blu-ray Disc, a CD, a ROM, a PROM, an EPROM, an EEPROM, or a FLASH memory, a hard disk, or other magnetic or optical storage device storing electronically readable control signals that can interact or cooperate with a programmable computer system to perform the respective method. Therefore, the digital storage medium may be computer-readable.

[0057] Some embodiments according to the invention thus comprise a data carrier having electronically readable control signals capable of interacting with a programmable computer system such that one of the methods described herein is carried out.

[0058] In general, embodiments of the present invention may be implemented as a computer program product having a program code, wherein the program code is effective to perform one of the methods when the computer program product is run on a computer.

[0059] The program code can, for example, also be stored on a machine-readable medium.

[0060] Other embodiments include the computer program for performing one of the methods described herein, wherein the computer program is stored on a machine-readable medium. In other words, one embodiment of the method according to the invention is thus a computer program that has program code for performing one of the methods described herein when the computer program runs on a computer. Another embodiment of the method according to the invention is thus a data carrier (or a digital storage medium or a computer-readable medium) on which the computer program for performing one of the methods described herein is recorded.

[0061] A further embodiment of the method according to the invention is thus a data stream or a sequence of signals that represents the computer program for carrying out one of the methods described herein. The data stream or the sequence of signals can be configured, for example, to be transferred via a data communication connection, for example, via the Internet.

[0062] A further embodiment comprises a processing device, for example a computer or a programmable logic device, which is configured or adapted to carry out one of the methods described herein.

[0063] A further embodiment comprises a computer on which the computer program for performing one of the methods described herein is installed.

[0064] A further embodiment according to the invention comprises a device or system designed to transmit a computer program for performing at least one of the methods described herein to a recipient. The transmission can be electronic or optical, for example. The recipient can be, for example, a computer, a mobile device, a storage device, or a similar device. The device or system can, for example, comprise a file server for transmitting the computer program to the recipient.

[0065] In some embodiments, a programmable logic device (e.g., a field-programmable gate array, an FPGA) may be used to perform some or all of the functionalities of the methods described herein. In some embodiments, a field-programmable gate array may interact with a microprocessor to perform any of the methods described herein. In general, in some embodiments, the methods are performed by any hardware device. This may be general-purpose hardware such as a computer processor (CPU) or method-specific hardware such as an ASIC.

[0066] The above-described embodiments are merely illustrative of the principles of the present invention. It is understood that modifications and variations of the arrangements and details described herein will be apparent to others skilled in the art. Therefore, it is intended that the invention be limited only by the scope of the following claims and not by the specific details presented in the description and explanation of the embodiments herein.

[0067] Reference symbol

[0068] Device (10, 10', 10", 10'")

[0069] Object (12, 12a, 12b)

[0070] Optical sensor (16)

[0071] Additional sensor (18s)

[0072] Data source (18d)

[0073] Processor (20)

Claims

Patent claims 1. A device (10, 10', 10", 10'") for the optical inspection of one or more objects (12, 12a, 12b), in particular foodstuffs, comprising: an optical sensor (16) configured to detect an optical property of the one or more objects (12, 12a, 12b) in order to obtain at least a first measured value associated with a first feature; a further sensor (18s) or further data sources (18d) configured to detect a further property of the one or more objects (12, 12a, 12b) in order to obtain at least a second value associated with a second feature; and a processor (20) configured to compare the first measured value associated with the first feature and the second value associated with the second feature with a model or trained model in order to derive state information for the one or more objects (12, 12a, 12b).

2. Device (10, 10', 10", 10'") according to claim 1, wherein the trained model is a Kl-trained model.

3. Device (10, 10', 10", 10'") according to claim 2, wherein the processor (20) is designed to train the Kl model based on a training data set which comprises at least a first measured value, a second value and a state information item, or based on a plurality of training data sets which each comprise first measured values, second value and the state information.

4. Device (10, 10', 10", 10'") according to one of the preceding claims, wherein the further sensor (18s) comprises a microwave sensor; and / or wherein the further sensor (18s) comprises a non-optical sensor; and / or wherein the second feature represents a non-optical feature.

5. Device (10, 10', 10", 10'") according to one of the preceding claims, wherein the further sensor (18s) comprises an ultrasonic sensor, X-ray sensor, gas sensor / sniffer sensor / gas spectroscope.

6. Device (10, 10', 10", 10'") according to one of the preceding claims, wherein the optical sensor (16) comprises a stationary camera, RGB camera and / or HSI camera; and / or wherein the optical sensor (16) comprises a 3D camera and / or multiple cameras; and / or the optical sensor (16) is arranged at a defined distance from the one or more objects (12, 12a, 12b) and / or from a container for the one or more objects (12, 12a, 12b).

7. Device (10, 10', 10", 10'") according to one of the preceding claims, wherein the optical sensor (16) is designed to take a spectroscopy image of the one or more objects (12, 12a, 12b) as a measured value and / or to determine a spectral footprint of the one or more objects (12, 12a, 12b) as a measured value or as a basis for the feature determination.

8. Device (10, 10', 10", 10'") according to one of the preceding claims, wherein the optical sensor (16) is designed to determine a plurality of first measured values ​​associated with a plurality of 2D positions over the plurality of objects (12, 12a, 12b).

9. Device (10, 10', 10", 10'") according to one of the preceding claims, wherein the further sensor (18s) is designed to determine a plurality of second measured values ​​associated with 2D positions over a plurality of objects (12, 12a, 12b) or associated with 3D positions along a penetration depth into the one or more objects (12, 12a, 12b).

10. Device (10, 10', 10", 10'") according to one of the preceding claims, wherein the device (10, 10', 10", 10'") has an interface to means for identifying the one or more objects (12, 12a, 12b), in particular an interface for an RFID or barcode reader, in order to obtain the second value or a further additional value via the interfaces.

11. Device (10, 10', 10", 10'") according to one of the preceding claims, wherein the device (10, 10', 10", 10'") comprises an interface to a merchandise management system in order to receive the second value or a further additional value via the interface.

12. Device (10, 10', 10", 10'") according to one of the preceding claims, wherein the processor (20) is designed to determine a characteristic quality value and / or a characteristic origin indication as status information, so that the characteristic quality value or the characteristic origin indication can be compared with information from a merchandise management system.

13. Device (10, 10', 10", 10'") according to one of the preceding claims, wherein the features come from the group comprising size, shape, color, structure, etc.; and / or wherein the condition information comes from the group comprising a variety, ripeness, freshness, and origin.

14. Device (10, 10', 10", 10'") according to one of the preceding claims, wherein the objects (12, 12a, 12b) come from the group comprising fruit, vegetables, food, fruit stored in boxes, vegetables stored in boxes, food stored in boxes.

15. A system for the optical inspection of one or more objects (12, 12a, 12b), in particular foodstuffs, comprising: an optical sensor (16) configured to detect an optical property of the one or more objects (12, 12a, 12b) in order to obtain at least one first measured value associated with a first feature; another sensor (18s) or another data source (18d) configured to detect a further property of the one or more objects (12, 12a, 12b) in order to obtain at least one second value associated with a second feature; and a processor (20) configured to compare the first measured value associated with the first feature and the second value associated with the second feature with a model or trained model in order to derive state information for the one or more objects (12, 12a, 12b).

16. Method for the optical inspection of one or more objects (12, 12a, 12b), in particular foodstuffs, comprising the following steps: Detecting an optical property of the one or more objects (12, 12a, 12b) by means of an optical sensor (16) in order to obtain at least a first measured value associated with a first feature; detecting a further property of the one or more objects (12, 12a, 12b) by means of a further sensor (18s) or from a further data source (18d) in order to obtain at least a second value associated with a second feature; and Comparing the first measured value associated with the first feature and the second value associated with the second feature with a model or trained model to derive state information for the one or more objects (12, 12a, 12b).

17. The method of claim 16, wherein detecting the further property comprises detecting a non-optical property; and / or wherein the second feature represents a non-optical feature.

18. A computer program comprising a program code for carrying out the method according to claim 16 or 17.

Images