Ripeness detection system using hue color space and peak finding
A method for estimating ripeness of produce includes illuminating the produce with light, measuring intensities of the light reflected from the produce at different frequencies, and determining a degree of ripeness of the produce from the relative intensities of the light reflected from the produce at the different frequencies.
1. A method for estimating ripeness of produce, the method comprising:
illuminating the produce with light;
measuring intensities of the light reflected from the produce at different frequencies;
selecting, by a convolutional neural network, a cropped region of an image of the produce, wherein the cropped region of the image of the produce is selected such that an image of a target item of produce within the cropped region of the image of the produce occupies more than 50% of the cropped region by an area and such that the target item of produce is less than 50% occluded by an object within the cropped region of the image of the produce;
converting the cropped region of the image to a hue, saturation, and value (HSV) color representation;
constructing a hue histogram of hues in the cropped region of the image; and
determining a degree of ripeness of the target item of produce from both the intensities of the light reflected from the produce at the different frequencies and a dominant peak in the hue histogram.
2. The method of claim 1 , wherein the cropped region is a same size as the target item of produce for which ripeness is to be calculated.
3. The method of claim 1 , further comprising computing a histogram of pixel population of the cropped region of the image by hue and saturation.
4. The method of claim 1 , further comprising identifying the dominant peak in the hue histogram.
5. The method of claim 4 , further comprising determining the degree of ripeness of the target item of produce from a location of the dominant peak in the hue histogram.
6. The method of claim 5 , further comprising calibrating a hue histogram peak location versus produce ripeness scale for different varieties of produce.
7. The method of claim 1 , wherein the method is utilized to estimate ripeness of produce which exhibits a change in concentration of one or more of ß-carotene, lutein, lycopene, or other carotenoids as the produce ripens.
8. The method of claim 7 , wherein the method is utilized to estimate ripeness of produce selected from the group consisting of tomatoes, peppers, and cucumbers.
9. The method of claim 1 , further comprising providing an indication of whether the produce is sufficiently ripe for harvesting based on determination of the degree of ripeness of the produce.
10. The method of claim 9 , further comprising communicating the indication of whether the produce is sufficiently ripe for harvesting to a robotic system configured to harvest the produce.
11. The method of claim 10 , further comprising harvesting produce with the robotic system responsive to the produce being sufficiently ripe for harvesting.
12. A method for estimating ripeness of produce in an indoor agricultural environment having an ambient light power spectrum that differs from a power spectrum of natural outdoor light, the method comprising:
illuminating the produce with light in a bandwidth that is reflected from the produce to a degree that varies with ripeness of the produce;
measuring an intensity of the light reflected from the produce;
selecting, by a convolutional neural network, a cropped region of an image of the produce, wherein the cropped region of the image of the produce is selected such that an image of a target item of produce within the cropped region of the image of the produce occupies more than 50% of the cropped region by an area and such that the target item of produce is less than 50% occluded by an object within the cropped region of the image of the produce;
converting the cropped region of the image to a hue, saturation, and value (HSV) color representation;
constructing a hue histogram of hues in the cropped region of the image; and
determining a degree of ripeness of the target item of produce from both the intensity of the light reflected from the produce and a dominant peak in the hue histogram.
13. A system for estimating ripeness of produce, the system comprising:
an image sensor configured to measure intensities of light reflected from the produce at different frequencies; and
a processor configured to:
receive, from a convolutional neural network, a cropped region of an image of the produce, wherein the cropped region of the image of the produce is selected such that an image of a target item of produce within the cropped region of the image of the produce occupies more than at least 50 % of the cropped region by an area and such that the target item of produce is less than 50 % occluded by an object within the cropped region of the image of the produce;
convert the cropped region of the image to a hue, saturation, and value (HSV) color representation;
construct a hue histogram of hues in the cropped region of the image; and
determine a degree of ripeness of the target item of produce from both relative intensities of the light reflected from the produce at the different frequencies and a dominant peak in the hue histogram.
14. The system of claim 13 , wherein the cropped region is a same size as the target item of produce for which ripeness is to be calculated.
15. The system of claim 14 , wherein the processor is further configured to select the cropped region of the image such that greater than 50% of pixels in the cropped region of the image are included in an image of the target item of produce.
16. The system of claim 13 , wherein the processor is further configured to compute a histogram of pixel population of the cropped region of the image by hue and saturation.
17. The system of claim 13 , wherein the processor is further configured to identify the dominant peak in the hue histogram.
18. The system of claim 17 , wherein the processor is further configured to determine the degree of ripeness of the target item of produce from a location of the dominant peak in the hue histogram.
19. The system of claim 18 , further comprising a robotic harvester in communication with the processor and configured to harvest the target item of produce.
20. The system of claim 13 , wherein the processor is further configured to provide an indication of whether the produce is sufficiently ripe for harvesting responsive to determining the degree of ripeness of the produce.
21. The system of claim 20 , wherein the processor is further configured to communicate the indication of whether the produce is sufficiently ripe for harvesting to a robotic system configured to harvest the produce.
22. The system of claim 21 , further comprising the robotic system configured to harvest the produce.
23. A robotic harvester configured to harvest items of agricultural produce identified as sufficiently ripe for harvesting by the system of claim 17 .
24. The system of claim 13 , wherein the processor is further configured to calibrate a hue histogram peak location versus produce ripeness scale for different varieties of produce.
25. The system of claim 13 , wherein the produce exhibits a change in concentration of one or more of ß-carotene, lutein, lycopene, or other carotenoids as the produce ripens, wherein the produce is selected from the group consisting of tomatoes, peppers, and cucumbers.
26. A system for estimating ripeness of produce in an indoor agricultural environment having an ambient light power spectrum that differs from a power spectrum of natural outdoor light, the system comprising:
a projector configured to illuminate the produce with light in a bandwidth that is reflected from the produce to a degree that varies with ripeness of the produce;
a light intensity meter configured to measure an intensity of the light reflected from the produce; and
a processor configured to
receive, from a convolutional neural network, a cropped region of an image of the produce, wherein the cropped region of the image of the produce is selected such that an image of a target item of produce within the cropped region of the image of the produce occupies more than 50% of the cropped region by an area and such that the target item of produce is less than 50% occluded by an object within the cropped region of the image of the produce,
convert the cropped region of the image to a hue, saturation, and value (HSV) color representation,
construct a hue histogram of hues in the cropped region of the image, and
determine a degree of ripeness of the target item of produce from both the intensity of the light reflected from the produce and a dominant peak in the hue histogram.