Automated vehicle condition grading

1. A method for automated vehicle condition grading, comprising:

obtaining, by a system, training data comprising:

sets of annotated images for a plurality of different vehicles, wherein each set of annotated images is associated with a different vehicle from the plurality of different vehicles, and

a plurality of user grades, wherein each user grade is associated with an image from the sets of annotated images and indicates a manual grading of the image by a user;

preprocessing, by the system, the training data to modify one or more image parameters from the sets of annotated images;

training, by the system, one or more machine learning models using the preprocessed training data;

obtaining, by the system and from a first client device or a data store, a set of images associated with a vehicle, the set of images being captured by at least one still camera or video camera and comprising each of the four corners of the vehicle;

processing, by the system, the set of images using the one or more trained machine learning models to generate a set of optical vehicle condition grades, wherein each optical vehicle condition grade, of the set of optical vehicle condition grades, is associated with an image from the set of images and indicates a discrete score from a range of discrete scores that are associated with damage to the vehicle that is shown in the set of images;

assigning, by the system, a plurality of weighted values to the set of images based on whether the vehicle includes a feature;

determining, by the system and based on applying the plurality of weighted values to the set of optical vehicle condition grades, an aggregated optical vehicle condition grade;

generating, by the system and based at least in part on the aggregated optical vehicle condition grade, a final vehicle condition grade for the vehicle, wherein the final vehicle condition grade indicates a degree of overall damage to the vehicle;

determining, by the system, vehicle panel information for one or more panels of the vehicle based on the set of optical vehicle condition grades, the aggregated optical vehicle condition grade, or the final vehicle condition grade; and

providing, by the system, the final vehicle condition grade and the vehicle panel information to a second client device.

2. The method of claim 1 , wherein assigning the plurality of weighted values to the set of images comprises:

assigning a first weighted value to a first image of the set of images based on a first section of the vehicle shown in the first image and whether the vehicle includes the feature, wherein the first image is assigned a first value for the first weighted value based on the vehicle including the feature and is assigned a second value for the first weighted value based on the vehicle not including the feature; and

assigning a second weighted value to a second image of the set of images based on a second section of the vehicle shown in the second image, wherein the second image is assigned a third value for the second weighted value regardless of whether the vehicle includes the feature.

3. The method of claim 1 , wherein the vehicle panel information indicates damage to the one or more panels of the vehicle.

4. The method of claim 1 , further comprising:

obtaining, from the first client device or the data store, information for a predetermined set of known parameters of the vehicle, wherein the predetermined set of known parameters comprise at least one of a vehicle body type, a vehicle door set type, a vehicle drive type, and a vehicle engine type; and

determining the plurality of weighted values based on the information for the predetermined set of known parameters of the vehicle.

5. The method of claim 1 , wherein the one or more trained machine learning models comprises a plurality of trained machine learning models, and wherein processing the set of images using the one or more trained machine learning models to generate a set of optical vehicle condition grades comprises:

processing each image of the set of images using a different trained machine learning model from the plurality of trained machine learning models.

6. The method of claim 5 , wherein each of the plurality of trained machine learning models is further associated with a set of subgrade models, and

wherein each subgrade model of the set of subgrade models is applied to generate an optical vehicle condition subgrade for the optical vehicle condition grade.

7. The method of claim 6 , wherein the final vehicle condition grade is generated based at least in part on a vehicle condition subgrade using the set of subgrade models.

8. The method of claim 5 , wherein each of the plurality of trained machine learning models is associated with a vehicle perspective of a plurality of vehicle perspectives, and

wherein each of the set of optical vehicle condition grades is associated with a vehicle perspective of the plurality of vehicle perspectives.

9. The method of claim 8 , wherein the plurality of vehicle perspectives comprises a driver front corner of the vehicle, a passenger front corner of the vehicle, a driver rear corner of the vehicle, and a passenger rear corner of the vehicle.

10. The method of claim 1 , further comprising:

generating an adjustment grade for the vehicle, wherein generating the final vehicle condition grade for the vehicle is based on the adjustment grade for the vehicle, wherein the adjustment grade indicates a grade associated with historical or current attributes of the vehicle.

11. The method of claim 10 , wherein generating the adjustment grade for the vehicle is based on one or more statistical models and/or one or more adjustment grade machine learning models.

12. The method of claim 10 , wherein generating the adjustment grade for the vehicle is based on additional information associated with the vehicle, wherein the additional information associated with the vehicle comprises one or more of:

vehicle mileage;

body type of the vehicle;

an engine status code;

a primary damage code;

a secondary damage code;

a vehicle age;

an odometer reading type code;

an indication of whether the vehicle is able to run;

an identification number of the vehicle; and

whether an airbag has deployed.

13. The method of claim 1 , wherein providing the final vehicle condition grade and the vehicle panel information to the second client device comprises:

providing the final vehicle condition grade to the second client device;

receiving a request for additional information from the second client device based on the second client device displaying the final vehicle condition grade and receiving a selection from a second user to display the additional information; and

in response to receiving the request, providing the vehicle panel information to the second client device.

14. The method of claim 1 , further comprising:

generating a predicted value for the vehicle based on the final vehicle condition grade; and

providing the predicted value to the second client device.

15. The method of claim 1 , further comprising:

obtaining, by the second client device, the final vehicle condition grade and the vehicle panel information;

generating, by the second client device and based on the final vehicle condition grade of the vehicle, a list of available vehicles according to a respective final vehicle condition grade; and

displaying, by the second client device, the list of available vehicles.

16. A system for automated vehicle condition grading, comprising:

one or more processors; and

a non-transitory computer-readable medium having processor-executable instructions stored thereon, wherein the processor-executable instructions, when executed by the one or more processors, facilitate:

obtaining training data comprising:

sets of annotated images for a plurality of different vehicles, wherein each set of annotated images is associated with a different vehicle from the plurality of different vehicles, and

a plurality of user grades, wherein each user grade is associated with an image from the sets of annotated images and indicates a manual grading of the image by a user;

preprocessing the training data to modify one or more image parameters from the sets of annotated images;

training one or more machine learning models using the preprocessed training data;

obtaining, from a first client device or a data store, a set of images associated with a vehicle, the set of images being captured by at least one still camera or video camera and comprising each of the four corners of the vehicle;

processing the set of images using the one or more trained machine learning models to generate a set of optical vehicle condition grades, wherein each optical vehicle condition grade, of the set of optical vehicle condition grades, is associated with an image from the set of images and indicates a discrete score from a range of discrete scores that are associated with damage to the vehicle that is shown in the set of images;

assigning a plurality of weighted values to the set of images based on whether the vehicle includes a feature, comprising:

assigning a first weighted value to a first image of the set of images based on a first section of the vehicle shown in the first image and whether the vehicle includes the feature, wherein the first image is assigned a first value for the first weighted value based on the vehicle including the feature and is assigned a second value for the first weighted value based on the vehicle not including the feature; and

assigning a second weighted value to a second image of the set of images based on a second section of the vehicle shown in the second image, wherein the second image is assigned a third value for the second weighted value regardless of whether the vehicle includes the feature;

determining, based on applying the plurality of weighted values to the set of optical vehicle condition grades, an aggregated optical vehicle condition grade;

generating, based at least in part on the aggregated optical vehicle condition grade, a final vehicle condition grade for the vehicle, wherein the final vehicle condition grade indicates a degree of overall damage to the vehicle;

determining vehicle panel information for one or more panels of the vehicle based on the set of optical vehicle condition grades, the aggregated optical vehicle condition grade, or the final vehicle condition grade; and

providing the final vehicle condition grade and the vehicle panel information to a second client device.