Method for tracking and characterizing perishable goods in a store

Abstract

One variation of a method for tracking fresh produce in a store includes: accessing a first hyper-spectral image, of a produce display in a store, recorded at a first time; extracting a first spectral profile from a first region of the first hyper-spectral image depicting a first set of produce units in the produce display; identifying a first varietal of the first set of produce units; characterizing qualities (e.g., ripeness, bruising, spoilage) of the first set of produce units in the produce display based on the first spectral profile; and, in response to qualities of the first set of produce units in the produce display deviating from a target quality range assigned to the first varietal, generating a prompt to audit the first set of produce units in the produce display.

I claim: 
     
         1 . A method for tracking fresh produce in a store comprising:
 accessing a first hyper-spectral image, of a produce display in the store, recorded at a first time;   detecting a first produce unit of a first varietal in a first region of the first hyper-spectral image;   extracting a first spectral profile from the first region of the first hyper-spectral image;   characterizing a quality of the first produce unit in the produce display based on the first spectral profile; and   in response to the quality of the first produce unit deviating from a target quality range, generating a prompt to audit the first produce unit in the produce display.   
     
     
         2 . The method of  claim 1 :
 wherein detecting the first produce unit of the first varietal in the first hyper-spectral image comprises detecting a first group of produce units in the first hyper-spectral image, the first group of produce units comprising the first produce unit;   wherein extracting the first spectral profile from the first region of the first hyper-spectral image comprises extracting the first spectral profile from the first region of the first hyper-spectral image;   wherein characterizing the quality of the first produce unit in the produce display based on the first spectral profile comprises characterizing the quality of the first group of produce units in the produce display based on the first spectral profile; and   wherein generating the prompt to audit the first produce unit in the produce display comprises, in response to the quality of the first group of produce units deviating from the target quality range, generating the prompt to audit the first group of produce units in the produce display.   
     
     
         3 . The method of  claim 1 , wherein detecting the first produce unit of the first varietal in the first region of the first hyper-spectral image comprises:
 detecting a set of edges in the first hyperspectral image;   accessing a template edge geometry of the first varietal; and   identifying a subset of edges, in the set of edges, as the first produce unit of the first varietal in response to matching the subset of edges to the template edge geometry.   
     
     
         4 . The method of  claim 1 , further comprising:
 detecting a second produce unit in a second region of the first hyper-spectral image;   extracting a second spectral profile from the second region of the first hyper-spectral image; and   generating a second prompt to investigate mixed objects in the produce display in response to the second spectral profile differing from the first spectral profile.   
     
     
         5 . The method of  claim 1 , further comprising:
 identifying the first varietal of the first produce unit based on the first spectral profile;   detecting a second produce unit in a second region of the first hyper-spectral image;   extracting a second spectral profile from the second region of the first hyper-spectral image;   identifying a second varietal of the second produce unit based on the second spectral profile; and   generating a second prompt to remove the second produce unit from the produce display in response to the second varietal differing from the first varietal.   
     
     
         6 . The method of  claim 1 , wherein generating the prompt to audit the produce display comprises:
 generating a recommendation to re-sort the set of produce units occupying the produce display based on ripeness; and   transmitting the recommendation to a computing device affiliated with an associate of the store.   
     
     
         7 . The method of  claim 1 :
 wherein characterizing the quality of the first produce unit comprises identifying the first produce unit as overripe based on the first spectral profile; and   wherein generating the prompt to audit the produce display comprises:
 generating a recommendation to remove the first produce unit from the produce display; and 
 transmitting the recommendation to a computing device affiliated with an associate of the store. 
   
     
     
         8 . The method of  claim 1 :
 further comprising accessing the target quality range defining a target ripeness range;   wherein characterizing the quality of the first produce unit comprises identifying the first produce unit as underripe and outside of the target ripeness range based on the first spectral profile;   wherein generating the prompt to audit the produce display comprises generating a recommendation to return the first produce unit from the produce display to back-of-store inventory; and   further comprising:
 estimating a time duration for the first produce unit to reach the target ripeness range; and 
 scheduling a second prompt to return the first produce unit to the produce display after the time duration. 
   
     
     
         9 . A method for tracking fresh produce in a store comprising:
 accessing a first hyper-spectral image, of a first produce bin, recorded at a first time by an imaging system located within the store;   detecting a first set of produce units of a first varietal in a first region of the first hyper-spectral image;   extracting a first spectral profile from the first region depicting the first set of produce units;   characterizing a quality of the first set of produce units in the first produce bin based on the first spectral profile; and   in response to the quality of the first set of produce units deviating from a target quality range, generating a prompt to audit the first set of produce units in the first produce bin.   
     
     
         10 . The method of  claim 9 , further comprising:
 accessing the target quality range defining a target ripeness range;   estimating a time duration for the first set of produce units to reach a target ripeness range; and   at a visual display proximal the first produce bin:
 rendering a visual representation of the quality of the first set of produce units in the first produce bin; and 
 rendering the time duration for the first set of produce units to reach the target ripeness range. 
   
     
     
         11 . The method of  claim 9 , further comprising:
 accessing a second hyper-spectral image of a second produce bin, recorded at a second time by the imaging system;   detecting a second set of produce units of the first varietal in the second hyper-spectral image;   extracting a second spectral profile from the second hyper-spectral image depicting the second set of produce units;   characterizing a quality of the second set of produce units in the second produce bin based on the second spectral profile;   accessing the target quality range defining a target ripeness range;   estimating a time duration for the second set of produce units to reach a target ripeness range;   in response to detecting an outlier produce unit exhibiting an outlier characteristic in the second hyper-spectral image, generating a warning to remove the outlier produce unit from the produce bin; and   at a visual display proximal the second produce bin:
 rendering a visual representation of the quality of the second set of produce units in the second produce bin; 
 rendering the time duration for the second set of produce units to reach the target ripeness range; 
 rendering an order in which to move the second set of produce units from the second produce bin to the first produce bin; and 
 rendering the warning to remove the outlier produce unit from the second produce bin. 
   
     
     
         12 . The method of  claim 9 , further comprising:
 accessing a second hyper-spectral image captured by a fixed camera system, located within the store and defining a field of view excluding the first produce bin, responsive to an interaction with a patron within the store;   extracting a second spectral profile from the second hyper-spectral image depicting a second produce unit;   characterizing a second quality of the second produce unit based on the second spectral profile; and   rendering the second quality, for the patron, on a display proximal the fixed camera system.   
     
     
         13 . The method of  claim 9 :
 wherein accessing the first hyper-spectral image comprises accessing the first hyper-spectral image, of the first produce bin in the store, recorded at a first time by a mobile robotic system operating within the store; and   further comprising:
 accessing a second hyper-spectral image captured by the mobile robotic system located within the store, at a second time, and defining a field of view excluding the first produce bin, responsive to an interaction with a patron within the store; 
 extracting a second spectral profile from the second hyper-spectral image depicting a second produce unit; 
 characterizing a second quality of the second produce unit based on the second spectral profile; and 
 rendering the second quality, for the patron, on a display proximal the fixed camera system. 
   
     
     
         14 . The method of  claim 13 , further comprising, in response to the quality of the second produce unit deviating from a target quality range:
 generating a prompt to audit the first set of produce units in the first produce bin; and   generating a recommendation to the patron to discard the second produce unit.   
     
     
         15 . The method of  claim 9 , further comprising,
 accessing a second hyper-spectral image captured by a fixed camera system, located proximal a checkout apparatus, and defining a field of view excluding the first produce bin, responsive to an interaction with a patron within the store;   extracting a second spectral profile from the second hyper-spectral image depicting a second produce unit;   extracting a third spectral profile from the second hyper-spectral image depicting a third produce unit;   accessing a price of the second produce unit based on the second spectral profile;   accessing a price of the third produce unit based on the third spectral profile;   calculating a combined price based on the price of the second produce unit and the price of the third produce unit; and   rendering the combined price on a visual display located proximal the checkout apparatus.   
     
     
         16 . A method for tracking fresh produce in a store includes:
 accessing a first hyper-spectral image, of a first produce display in a store, recorded at a first time;   detecting a first produce unit of a first varietal in a first region of the first hyper-spectral image;   extracting a first spectral profile from the first region of the first hyper-spectral image;   characterizing a quality of the first produce unit in the first produce display based on the first spectral profile;   in response to the quality of the first produce unit deviating from a target quality range, calculating a time duration until the first produce unit is predicted to reach the target quality range based on the first hyper-spectral image; and   scheduling a future audit of the first produce display.   
     
     
         17 . The method of  claim 16 , wherein calculating the time duration until the first produce unit is predicted to reach the target quality range comprises calculating the time duration based on the difference between the first time that the first hyper-spectral image was recorded and a time that the first produce unit arrived at the store. 
     
     
         18 . The method of  claim 16 , further comprising:
 detecting a second produce unit of the first varietal in a first region of the first hyper-spectral image;   extracting a second spectral profile from the first region of the first hyper-spectral image;   characterizing a quality of the second produce unit in the first produce display based on the second spectral profile;   in response to the quality of the second produce unit deviating from a target quality range, calculating a time duration until the quality of the first produce unit is predicted to deviate from the target quality range based on the quality of the second produce unit and the distance of the second produce unit to the first produce unit.   
     
     
         19 . The method of  claim 16 , further comprising:
 accessing a second hyper-spectral image, of the produce display, recorded at a second time;   extracting a second spectral profile from a first region of the second hyper-spectral image of the first produce unit at the second time;   characterizing a second quality of the first produce unit in the produce display based on the second spectral profile;   estimating a rate of change of quality of the first produce unit; and   calculating a time duration until the first produce unit is predicted to reach the target quality range based on the estimated rate of change of the first produce unit.   
     
     
         20 . The method of  claim 16 , further comprising:
 accessing a second hyper-spectral image, of a second produce display in the store, recorded at a second time;   identifying a second produce unit of a second varietal in the second hyper-spectral image occupying a first region of the second hyper-spectral image;   extracting a second spectral profile from the first region of the second image depicting the second produce unit;   characterizing a quality of the second produce unit in the second produce display based on the second spectral profile; and   predicting a supply of ripe produce based on the ripeness level of the first produce unit and the second produce unit.