Method and system for estimation of fruitlet drop

    公开(公告)号:US11131657B2

    公开(公告)日:2021-09-28

    申请号:US16933479

    申请日:2020-07-20

    Abstract: A system comprising: at least one hardware processor; and a non-transitory computer-readable storage medium having stored thereon program instructions, the program instructions executable by the at least one hardware processor to: receive spectral data acquired from a plurality of fruits, wherein the spectral data is obtained within a specified range of wavelengths, at a training stage, train a machine learning model on a training set comprising: (i) the spectral data, and (ii) labels indicating, with respect to each of the fruits, a drop status within a specified time period subsequent to the acquiring, and at an inference stage, apply the machine learning model to target spectral data acquired from a target fruit, to predict the drop status of the target fruit within a specified time range subsequent to the acquiring.

    HYPERSPECTRAL SENSING SYSTEM AND PROCESSING METHODS FOR HYPERSPECTRAL DATA

    公开(公告)号:US20210231497A1

    公开(公告)日:2021-07-29

    申请号:US17230386

    申请日:2021-04-14

    Abstract: A hyperspectral sensing device may include an optical collector configured to collect light and to transfer the collected light to a sensor having spectral resolution sufficient for sensing hyperspectral data. In some examples, the sensor comprises a compact spectrometer. The device further comprises a power supply, an electronics module, and an input/output hub enabling the device to transmit acquired data (e.g., to a remote server). In some examples, a plurality of hyperspectral sensing devices are deployed as a network to acquire data over a relatively large area. Methods are disclosed for performing dark-current calibration and/or radiometric calibration on data obtained by the hyperspectral sensing device, and/or another suitable device. Data obtained by the device may be represented in a functional basis space, enabling computations that utilize all of the hyperspectral data without loss of information.

    Hyperspectral sensing system and processing methods for hyperspectral data

    公开(公告)号:US11073423B2

    公开(公告)日:2021-07-27

    申请号:US16583042

    申请日:2019-09-25

    Abstract: A hyperspectral sensing device may include an optical collector configured to collect light and to transfer the collected light to a sensor having spectral resolution sufficient for sensing hyperspectral data. In some examples, the sensor comprises a compact spectrometer. The device further comprises a power supply, an electronics module, and an input/output hub enabling the device to transmit acquired data (e.g., to a remote server). In some examples, a plurality of hyperspectral sensing devices are deployed as a network to acquire data over a relatively large area. Methods are disclosed for performing dark-current calibration and/or radiometric calibration on data obtained by the hyperspectral sensing device, and/or another suitable device. Data obtained by the device may be represented in a functional basis space, enabling computations that utilize all of the hyperspectral data without loss of information.

    Subject identification device and subject identification method

    公开(公告)号:US10989596B2

    公开(公告)日:2021-04-27

    申请号:US16747597

    申请日:2020-01-21

    Abstract: A subject identification device includes: an illuminator configured to generate illumination light including components at a plurality of wavelength bands, each of the components having a characteristic in accordance with a respective one of settings; an imager configured to generate an image signal by capturing light from a subject under the illumination light having the illumination characteristic; and a processor including hardware. The processor is configured to: define an illumination characteristic of the illumination light; analyze the image signal to acquire spectral information of the subject; and cross check the spectral information of the subject with subject identification information in order to identify the subject. When the subject is not identified, the processor is configured to define another illumination characteristic that causes spectral information of potentials for the subject to be identified, and subsequently each of the imager and the processor performs a process.

    DEVICE FOR RECORDING SPECTRAL INFORMATION
    116.
    发明申请

    公开(公告)号:US20190301935A1

    公开(公告)日:2019-10-03

    申请号:US16442254

    申请日:2019-06-14

    Applicant: SPECTRICITY

    Abstract: A user device including a camera, a spectrometer module, and a processing unit is disclosed. In one aspect, the camera is adapted to acquire at least one image of a scenery which falls within a field of view of the camera. The spectrometer module is adapted to acquire spectral information from a region within the scenery which region falls within a field of view of the spectrometer module. The processing unit is adapted to determine, based on information relating the field of view of the spectrometer module to the field of view of the camera, a spectrometer module target area, within the at least one image, corresponding to the region. The processing unit is adapted to output display data to a screen of the user device for providing an indication of the target area on the display.

    Object recognition apparatus using spectrometer and method thereof
    119.
    发明授权
    Object recognition apparatus using spectrometer and method thereof 有权
    使用光谱仪的物体识别装置及其方法

    公开(公告)号:US09217670B2

    公开(公告)日:2015-12-22

    申请号:US14306866

    申请日:2014-06-17

    Abstract: Provided are an apparatus and method for recognizing an object on the basis of property information on an object obtained using a multi-wavelength spectrometer. An apparatus for recognizing an object using a multi-wavelength spectrometer includes an image processing unit configured to extract an region of interest from an input three-dimensional image and output shape information on the region of interest, a light irradiation unit configured to irradiate light of a plurality of wavelengths to an arbitrary position of an object corresponding to the detected region of interest, a light receiving unit configured to measure a spectrophotometric value for each light of the plurality of wavelengths, and a light processing unit configured to generate a differential spectrophotometric map using a differential value between spectrophotometric values of different wavelengths measured at the same light irradiation position, and recognize the object using the differential spectrophotometric map and the shape information.

    Abstract translation: 提供了一种用于基于使用多波长光谱仪获得的对象的属性信息来识别对象的装置和方法。 一种使用多波长光谱仪识别物体的装置,包括:图像处理单元,被配置为从输入的三维图像中提取感兴趣的区域,并输出关于感兴趣区域的形状信息;光照射单元, 多个波长到对应于所检测的感兴趣区域的物体的任意位置;光接收单元,被配置为测量所述多个波长中的每个光的分光光度值;以及光处理单元,被配置为生成差分分光光度图 使用在相同光照射位置测量的不同波长的分光光度值之间的差值,并使用差分分光光度图和形状信息识别物体。

Patent Agency Ranking