Algorithm application is a crucial component in the field of AOI (Automatic Optical Inspection) algorithms. AOI boasts over 20 different algorithms, each serving a specific purpose. Therefore, on the basis of familiarizing and understanding the various AOI algorithms, applying these algorithms to different inspection items is a prerequisite for AOI engineers to develop inspection programs.

Defective parts, primarily used for inspecting the integrity of components, check for material discrepancies within the component. This inspection is a standard item in AOI (Automated Optical Inspection) processes. Four detection algorithms can be applied to defective parts, which include TOC (Total Organic Carbon) algorithm, OCV (Optical Character Verification) algorithm, Match algorithm, and OCR (Optical Character Recognition) algorithm. Each detection algorithm for defective parts emphasizes different aspects of the inspection items.

Defect detection for TOC algorithms is primarily used for non-characteristic component defects, mainly focusing on capacitors. This detection method identifies component defects by extracting the intrinsic color of the components and determining if there has been a change in the intrinsic color. The intrinsic color parameters of the components have no default values and are determined based on the actual intrinsic color for color extraction.

OCV algorithm-based error detection is primarily used for clear character errors, mainly targeting resistors. This detection method assesses whether an error has occurred in the component by comparing the fitting degree of the contour of the character to be tested with the standard character contour. The default range for the judgment parameters of this detection is (0, 12). For instance, if the standard character is "123" and the character to be tested is "351", with a fitting return value of 28.3 and a judgment range of (0, 12), the component is deemed to have a "mismatch error."

The Match class detection algorithm is primarily used for detecting faulty components of fuzzy character types, such as diodes and transistors. This type of detection algorithm determines if a component has a "faulty part" by comparing the similarity between the area to be tested and the standard character area. The default range for faulty part detection is set to (0, 32).

OCR-based detection algorithms are primarily used for the inspection of components in critical parts, such as BGA, QFP, and other similar components. These algorithms detect and judge whether a component is defective by recognizing the characters to be tested and determining if they match the standard characters. For instance, if the standard character is "123" and the actual character is "122," the OCR algorithm will identify the component as having a "defective part."