Image inspection apparatus
When a normal inspection and an inspection through deep learning processing is applicable, high inspection accuracy is obtained while reducing a processing time. The normal inspection processing is applied to a newly acquired inspection target image, the non-defective product determination or the defective product determination is confirmed for the inspection target image having the characteristic amount with which the non-defective product determination or the defective product determination is executable based on the characteristic amount within the inspection target image and the threshold for confirming the non-defective product determination or the threshold for confirming the defective product determination. The deep learning processing is applied to the inspection target image having the characteristic amount with which the non-defective product determination or the defective product determination is not confirmable, and the non-defective product determination or the defective product determination is executed.
1. An image inspection apparatus that performs a quality determination of an inspection target based on an inspection target image acquired by capturing the inspection target, the apparatus comprising:
a normal inspection setting section that performs a setting of normal inspection processing by receiving a setting of a characteristic amount used for an inspection and a setting of a threshold for confirming a non-defective product determination or a threshold for confirming a defective product determination from a user, the threshold being compared with the characteristic amount;
a deep learning setting section that causes a neural network to learn by inputting a plurality of non-defective product images to which non-defective product attributes are given and/or a plurality of defective product images to which defective product attributes are given to an input layer of the neural network, and performs a setting of deep learning processing for classifying a newly input inspection target image into the non-defective product image and the defective product image; and
an inspection execution section that applies the normal inspection processing to a newly acquired inspection target image, confirms the non-defective product determination or the defective product determination for the newly acquired inspection target image having the characteristic amount with which the non-defective product determination or the defective product determination is executable based on the characteristic amount within the newly acquired inspection target image and the threshold for confirming the non-defective product determination or the threshold for confirming the defective product determination, applies the deep learning processing to the newly acquired inspection target image having the characteristic amount with which the non-defective product determination or the defective product determination is not confirmable, and executes the non-defective product determination or the defective product determination, wherein
the inspection execution section is configured to perform the quality determination by a difference inspection for detecting a blob of a difference between a registration image registered in advance and the newly acquired inspection target image, and confirm the defective product determination by the normal inspection processing when a blob area of the difference obtained by the normal inspection processing is equal to or larger than the threshold for confirming the defective product determination.
2. The image inspection apparatus according to claim 1 , wherein
the inspection execution section is configured to confirm the non-defective product determination for the newly acquired inspection target image having the characteristic amount with which the non-defective product determination is executable and confirm the defective product determination for the newly acquired inspection target image having the characteristic amount with which the defective product determination is executable, by the normal inspection processing, and apply the deep learning processing to only the newly acquired inspection target image having the characteristic amount with which the non-defective product determination is not confirmable and the newly acquired inspection target image having the characteristic amount with which the defective product determination is not confirmable.
3. An image inspection apparatus that performs a quality determination of an inspection target based on an inspection target image acquired by capturing the inspection target, the apparatus comprising:
a normal inspection setting section that performs a setting of normal inspection processing by receiving a setting of a characteristic amount used for an inspection and a setting of a threshold for confirming a non-defective product determination or a threshold for confirming a defective product determination from a user, the threshold being compared with the characteristic amount;
a deep learning setting section that causes a neural network to learn by inputting a plurality of non-defective product images to which non-defective product attributes are given and/or a plurality of defective product images to which defective product attributes are given to an input layer of the neural network, and performs a setting of deep learning processing for classifying a newly input inspection target image into the non-defective product image and the defective product image; and
an inspection execution section that applies the normal inspection processing to a newly acquired inspection target image, confirms the non-defective product determination or the defective product determination for the newly acquired inspection target image having the characteristic amount with which the non-defective product determination or the defective product determination is executable based on the characteristic amount within the newly acquired inspection target image and the threshold for confirming the non-defective product determination or the threshold for confirming the defective product determination, applies the deep learning processing to the newly acquired inspection target image having the characteristic amount with which the non-defective product determination or the defective product determination is not confirmable, and executes the non-defective product determination or the defective product determination, wherein
the inspection execution section is configured to perform the quality determination by a difference inspection for detecting a blob of a difference between a registration image registered in advance and the newly acquired inspection target image, and confirm the non-defective product determination by the normal inspection processing when a blob area of the difference obtained by the normal inspection processing is equal to or smaller than the threshold for confirming the non-defective product determination.
4. The image inspection apparatus according to claim 3 , wherein
the inspection execution section is configured to confirm the non-defective product determination for the newly acquired inspection target image having the characteristic amount with which the non-defective product determination is executable and confirm the defective product determination for the newly acquired inspection target image having the characteristic amount with which the defective product determination is executable, by the normal inspection processing, and apply the deep learning processing to only the newly acquired inspection target image having the characteristic amount with which the non-defective product determination is not confirmable and the newly acquired inspection target image having the characteristic amount with which the defective product determination is not confirmable.