Multi-angle vehicle defect measurement using surface-adaptive optical corrections

A system for estimating dimensions of a vehicle's exterior defects, comprising:

• multiple image sensors mounted on a scanning structure and arranged in a predefined configuration relative to expected vehicle positioning to capture multiple surfaces of a vehicle while the vehicle passes in front of the multiple image sensors through the scanning structure;
• at least one memory storing instructions;
• and
• at least one processor configured to execute the instructions to:
• receive images of the vehicle captured by the multiple image sensors;
• compute an angle to a detected defect as a function of pixel coordinates of the detected defect in an image, an angular field of view of a respective image sensor capturing the image, and a pixel resolution of the respective image sensor;
• calculate distances to defects from focal planes of respective image sensors, each of the distances is calculated as a respective one of different functions of a positioning of a respective image sensor relative to the vehicle and at least one vehicle part dimension, the respective function and the at least one vehicle part dimension are determined based on vehicle section location captured by the respective image sensor, the different functions comprising at least one function parametrized by the angle to the detected defect;
• and
• compute defect sizes by determining multiple defect dimensions as respective functions of the distances to defects and pixel coordinates of the detected defects, with at least one dimension being adjusted by an angular correction factor calculated according to an angular relationship between an optical axis of a respective image sensor and a respective vehicle surface.

A non-transitory computer-readable storage medium having stored thereon instructions that, when executed by at least one processor, cause the at least one processor to perform operations for estimating dimensions of a vehicle's exterior defects using multiple cameras, the operations comprising:

• receiving images of a vehicle captured by multiple image sensors mounted on a scanning structure and arranged in a predefined configuration relative to expected vehicle positioning to capture multiple surfaces of a vehicle while the vehicle passes in front of the multiple image sensors through the scanning structure;
• computing an angle to a detected defect as a function of pixel coordinates of the detected defect in an image, an angular field of view of a respective image sensor capturing the image, and a pixel resolution of the respective image sensor;
• calculating distances to defects from focal planes of respective image sensors, each of the distances is calculated as a respective one of different functions of a positioning of a respective image sensor relative to the vehicle and at least one vehicle part dimension, the respective function and the at least one vehicle part dimension are determined based on vehicle section location captured by the respective image sensor, the different functions comprising at least one function parametrized by the angle to the detected defect;
• computing defect sizes by determining multiple defect dimensions as respective functions of the distances to defects and pixel coordinates of the detected defects, with at least one dimension being adjusted by an angular correction factor calculated according to an angular relationship between an optical axis of a respective image sensor and a respective vehicle surface.