in your cloud recognition sample code there is no provision provided for image not recognize
That's because it works differently.
...is called when the target finder has new results, and if you delve deeper into the TargetFinder API you may find other status values that could be of use.
also it(camera scaning) continously runing till recognition,
That is the way the sample works. If you can fully understand the code then you can create your own sample that can work differently.
I just want if image can't be recognise then it go for text recognition(that thing we will try),
One thing you could try is setting up a Timer and then if there is no match within that period then you can go for TR, but exactly how you do it is entirely up to you.
Is it possible to capture image (instead of running cotineous recording for image recognition) & send it for text & image reco?
No - video only.