Such as topic, doing now move to the automated testing, product list sliding in the webview page can't find the target product through appium check, so want to use OpenCV by image recognition to judge, the first in the capture the whole screen saved as picture B, another intercept target image A beforehand, and then through the OpenCV SURF algorithm to obtain A, B of feature points, but there is A problem is whether the current B pictures contain A, there will be some feature points can matching success, even after filtered by Euclidean distance will still exist this kind of situation, is there A big can guide

CodePudding user response:

Matching to determine the distance threshold, if does not contain the target is A match, the distance it will be relatively large

CodePudding user response:

The

oh roar reference 1 floor? Response:

matching to determine the distance threshold, if does not contain the target is A match, the distance of larger

A judgment has been made, but still there will be A match to other places, there is no way to get A definite answer of whether contain B

CodePudding user response:

Now use another way to solve, not solved by image recognition, although is the webview page then enter the loading all the elements, even invisible to the naked eye can also be detected, but will change the coordinates of elements in the process of sliding, now is to find target to judge the first page if the webview is, whether the coordinates of the element in equipment Y resolution in the upper, if not have, until the now just effects can also